Cardiac fibrosis development is directly associated with the activation of the renin-angiotensin-aldosterone system (RAAS) [8]. The progression of cardiac dysfunction reduces renal perfusion, so the juxtaglomerular apparatus starts to release renin [9]. The consequence of this release is an increment of the angiotensin II level produced by the angiotensin-converting enzyme (ACE). For a long time, angiotensin II has been considered the principal promoter of cardiac fibrosis [10]. In fact, angiotensin II directly drives numerous myofibroblast functions such as, cell proliferation and migration, transforming growth factor beta (TGF-$\beta$) release, and ECM synthesis through stimulation of the angiotensin II type 1 receptor (AT1R) [11, 12, 13, 14]. The anti-fibrotic mechanisms triggered by RAAS inhibition have been widely described in the literature through animal studies, together with the direct role of angiotensin II infusion in enhancing cardiac fibrosis in rodents [15]. To date, ACE inhibitors and AT1R antagonists represent the mainstay therapy in heart failure patients. They act in multiple ways by reducing ECM production and cardiac hypertrophy, as demonstrated both in cell culture and animal models [16, 17, 18]. Nonetheless, it was demonstrated that the specific activation of AT1R in cardiomyocytes has only a minimal impact on cardiac hypertrophy, suggesting a main role of this receptor in cardiac non-myocytic cells [19].

Aldosterone is a steroid hormone produced by adrenal cortex under stimulation of angiotensin II [20]. It is a transcriptional regulator and is responsible for upregulating pro-fibrotic genes. Aldosterone and angiotensin II work together by activating the mitogen-activated protein kinase (MAPK) pathway in cardiomyocytes [21, 22], as well as in CFs [23, 24]. Interestingly, aldosterone can exert its profibrotic effect also when angiotensin II is inactivated, suggesting its independent role in promoting fibrosis [25]. There are two drugs—spironolactone and eplerenone—which have revealed potential antifibrotic effects clinically. They are receptor antagonists of the mineralocorticoid family, blocking aldosterone signaling. Zannad et al. [26] verified the serum level reduction of fibrosis markers and collagen synthesis after spironolactone treatment. More data on their specific effect on CFs is needed, though.

Pharmacological approaches for the treatment of cardiac fibrosis include targeting the adrenergic receptor system, as well. This neurohormonal pathway contributes to the maintenance of cardiovascular homeostasis through the release of norepinephrine, a neurotransmitter that binds to both type $\alpha$ adrenergic receptors ($\alpha$-ARs) on the peripheral vasculature, or type $\beta$1 adrenergic receptors ($\beta$1-ARs) mostly on cardiomyocytes, with positive inotropic and chronotropic effects [20]. However, under chronic pathological conditions, continuous activation of ARs in the heart leads to maladaptive effects, such as adverse remodelling of cardiac tissue. In these cases, one of the most effective therapies to prevent heart failure is the blockade of $\beta$1-ARs on cardiomyocytes through $\beta$-blockers drugs [27, 28, 29], which have a well-established and consolidated efficacy. Beta-1 adrenergic signaling, though, has been shown to exert a direct effect also on human cardiac mesenchymal stromal cells, increasing gene expression of fibrotic markers such as collagen type I alpha 1 (COL1A1) and thymus cell antigen 1 (THY1). Indeed, the cells isolated from patients undergoing treatment with $\beta$-blockers displayed reduced expression of markers of fibrotic activation, and a less fibrotic epigenetic profile, compared to cells derived from patients not following a $\beta$-blockers therapy regimen [30].

Other types of $\beta$-ARs are expressed in cardiac non-myocytes, which play different roles. In detail, CFs mainly express $\beta$2-ARs, but their role in cardiac fibrosis pathways is still debated. Studies performed over the last decade have reported conflicting results, showing both pro- and anti-fibrotic responses [31]. Several studies have demonstrated that the activation of the $\beta$2-AR signalling in CFs inhibits the production of collagens and the transition to myofibroblasts, even under pro-fibrotic stimuli [32, 33, 34]. Other anti-fibrotic interventions have focused on the enhancement of the $\beta$2-AR signalling pathway via inhibiting the G protein-coupled receptor kinase 2 (GRK2), which phosphorylates the receptor leading to its desensitization. Different works [35, 36, 37] have also described that the fibroblast-specific inhibition of GRK2 could prevent cardiac fibrosis and dysfunction in animal models of heart failure. Nonetheless, $\beta$2-ARs have been shown to increase DNA synthesis, proliferation, and production of pro-fibrotic interleukin 6 (IL-6) in isolated rat and human CFs [38, 39, 40]. Conversely, specific $\beta$2-blockade in human mesenchymal stromal cells could reduce THY1 and COL1A1 expression, and the release of several pro-inflammatory cytokines [41]. Further studies are required on this pathway to understand the reasons for these conflicting results and hopefully reach a consensus.

Besides $\beta$1- and $\beta$2-ARs, a third type of adrenergic receptor is expressed in cardiac tissue, namely the $\beta$3-AR, whose levels are low in healthy hearts, but become up-regulated under pathological conditions with anti-fibrotic effects [42]. The potential advantage of fibrosis targeting through $\beta$3-AR-based approaches is that these receptors cannot be phosphorylated by GRKs, thus being resistant to desensitization [43]. Several studies have demonstrated the beneficial effects of $\beta$3-AR agonists on preventing myocardial fibrosis in animal models of myocardial infarction (MI) and pressure overload [44, 45]. Interestingly, a $\beta$3-AR-selective agonist, Mirabegron, has already obtained food and drug administration (FDA) approval for the treatment of overactive bladder, and it is now under assessment for the treatment of hypertensive structural heart disease [46].

Some active molecules occur naturally in food and could be used in clinics due to the wide range of pharmacological effects.

Vinpocetine is derived from a natural plant alkaloid and is used as a dietary supplement. It appears to act through an IkappaB kinase (IKK)-dependent pathway, attenuating CF activation and ECM synthesis, independently from small mothers against decapentaplegic (SMAD) phosphorylation or activation [47, 63]. Moreover, both in vitro and in vivo, it targets cyclic nucleotide phosphodiesterase 1 (PDE1), whose expression is up-regulated in the heart under RAAS and beta-adrenergic stimulation [48, 64]. Vinpocetine could also act through IKK with anti-inflammatory effects because of its role in nuclear factor kappa B (NF-$\kappa$B) activation [47], suggesting the potential to contribute to anti-remodelling effects [65].

Concerning cyclic nucleotide PDEs as promising targets for the modulation of the fibrotic process, the PDE1A form has been identified as a key regulator of fibroblast activation and ECM remodelling in the heart [48]. This isoform was found to be specifically enhanced in activated myofibroblasts both in vitro upon angiotensin II and TGF-$\beta$ stimulation, and in vivo in mouse, rat, and human fibrotic hearts. Miller et al. [48] demonstrated that the PDE1-selective inhibitor IC86340 reduces activation, ECM deposition, and the expression of pro-fibrotic genes in rat CFs, as well as the levels of alpha smooth muscle actin ($\alpha$-SMA), collagen I, and plasminogen activator inhibitor I (PAI-I) levels, all markers of cardiac fibrosis. In vivo, isoproterenol-induced interstitial fibrosis was attenuated upon PDE1A inhibition in mouse hearts [48]. Another PDE family member involved in pathological cardiac remodelling is PDE10A, identified by Chen and colleagues [49] as abnormally expressed in human and mouse failing hearts, as well as in murine CFs stimulated with TGF-$\beta$. Cardiac remodelling and fibrosis in angiotensin-infused mice were attenuated by specific inhibition of PDE10A through the administration of the selective inhibitor TP-10 [49]. Interestingly, PDE10A is an already known safe druggable target, since several PDE10A inhibitors have already passed phase I clinical trials in humans for the treatment of schizophrenia [66, 67] and Huntington’s disease [68, 69].

Berberine is derived from a natural plant alkaloid able of targeting the insulin-like growth factor I receptor (IGF-1R). IGF-1 is an important mediator of cardiac hypertrophy, but its continuous expression in the myocardium contributes to pathological remodelling and interstitial fibrosis [70]. The IGF-1R is upregulated in CFs isolated from diabetic hearts, characterized by overexpression of matrix metalloproteinase matrix metallopeptidase-2 (MMP-2) and matrix metallopeptidase-9 (MMP-9), $\alpha$-SMA, and collagen I. The long-term treatment with berberine has demonstrated its capacity to downregulate IGF-1R and all its downstream effects in a diabetic rat model [50].

Naringenin, a flavonoid present in citrus fruits, exerts a cardioprotective effect by suppressing extracellular signal-regulated kinase (ERK), c-Jun N-terminal kinase (JNK), and phosphatidylinositol3-kinase/protein kinase B (PI3K/Akt) signalling pathways. In vitro, naringenin caused the inhibition of TGF-$\beta$-induced proliferation and collagen production in CFs, partly mediated by arresting DNA synthesis at G0/G1 [51]. A recent study proposed a further mechanism targeting cardiac fibrosis through naringenin. In this work, it suppresses the expression of profibrotic proteins such as collagen I, collagen III, and $\alpha$-SMA by inactivating the SMAD3 signalling pathway in angiotensin-stimulated mouse CFs. In addition, naringenin acts as a SMAD3-inhibitor, and was shown to reduce the proliferation and ECM production by CFs in rats with hypertension-induced atrial fibrosis [52].

Piperlongumine is another alkaloid that is naturally isolated from the long pepper piper, and it has been proposed as a potential natural compound for the clinical treatment of cardiac hypertrophy and fibrosis. Angiotensin-induced expression of profibrotic proteins in neonatal CFs was significantly decreased by piperlongumine treatment, reducing the expression of Krüppel-like factor 4 (KLF4), and its recruitment to the promoter regions of the pro-fibrotic cytokines TGF-$\beta$ and connective tissue growth factor (CTGF). Like many other therapeutic strategies, the effects of piperlongumine are not restricted to non-myocytes. In fact, treatment with piperlongumine exerted also an anti-hypertrophic effect in neonatal rat cardiomyocytes by reducing the phosphorylation of Akt, thereby preserving the level of Forkhead box transcription factor O1 (FoxO1) [53]. The FoxO protein family, in fact, plays an important role in various biological processes such as metabolism, apoptosis and oxidative stress [71], blunting pathological hypertrophy through the inhibition of the calcineurin/nuclear factor of activated T-cells (NFAT) pathway [72]. Thus, piperlongumine could be considered as a preventive strategy, as well.

Curcumin is a spice-derived pigment, widely studied as an anti-cancer, anti-oxidant, and anti-inflammatory agent [73, 74]. Recently, it was investigated also as a modulator of cardiac fibrosis. Hsu and colleagues [75] reported reduction of TGF-$\beta$ in mice with MI treated with curcumin, describing modulation of sirtuin1 (SIRT1) activity. SIRT1 is a conserved histone deacetylase involved in various biological processes [75]. Recently, several pieces of evidence support the role of SIRT1 in cardiac fibrosis, as well [76]. Treatment with curcumin led to the activation of SIRT1 and to the decrement of metalloproteases 2 and 9, and collagen I and III levels in angiotensin-treated CFs [54]. Moreover, curcumin affected macrophage-fibroblast crosstalk in a mouse model of MI, reducing pro-inflammatory signalling by macrophages. In turn, IL-18 expression and secretion by CFs were reduced, as well as the activation of the TGF-$\beta$-p-SMAD2/3 network [77]. Other studies have highlighted how curcumin also increases SMAD7, which is a direct inhibitor of SMAD2/3 phosphorylation and a disruptor of the SMAD-complex formation [78], further leading to the attenuation of pro-fibrotic TGF-$\beta$ signalling [55].

A recent study reported the ability of curcumin to also modulate autophagy. In this case, it inhibited mammalian target of rapamycin (mTOR) signalling thus enhancing the autophagic process. Consequently, all pro-fibrotic markers such as procollagen I and III in rat hearts treated with isoprenaline were reversed by curcumin administration [56]. However, whether this in vivo effect was due to direct or indirect mechanisms has not been defined yet.

The disaccharide trehalose—derived from myocytes—is known to be a natural activator of autophagy. It has shown protective effects on the heart [79], and can specifically enhance autophagy also in resident cardiac mesenchymal stromal cells [80] promoting anti-fibrotic responses. Autophagy boosting in these cells can reduce the expression of markers of activated fibroblasts, and preserve a less pro-fibrotic and pro-inflammatory paracrine function. When cells are exposed to metabolic stress, which is a main mediator of myocardial fibrosis, trehalose treatment also increased pro-angiogenic effects of cardiac mesenchymal stromal cells [57]. This effect has been also confirmed in human cells derived from diabetic patients, where the enhancement of autophagy by trehalose, combined with oxidative stress reduction by nitrogen oxides (NOX)-inhibitors, may contribute to reducing fibrotic activation and specification of primitive stromal cells [81].

A recent study by Schimmel et al. [58] has identified fifteen substances with antiproliferative effects in human CFs by high-throughput natural compound library screening. The validation by multiple in vitro fibrosis assays and selection by stringent algorithms have identified the steroid bufalin and the alkaloid lycorine as potentially effective antifibrotic molecules. Both were assayed for suppression of collagen I and matrix components production in activated fibroblasts, and the most significant effects were mediated by bufalin [58].

Dapagliflozin is a sodium-glucose Cotransporter-2 (SGLT2) inhibitor. SGLT2 is involved in the pathogenetic mechanisms of diabetes by increasing the risk of cardiovascular disorders, such as MI, stroke, and heart failure [82]. The development of diabetic cardiomyopathy is caused by functional and metabolic changes that occur in the myocardium at multiple levels [57]. In the last years, studies have shown how SGLT2 inhibitors exhibit cardiovascular safety and benefits [83]. Dapagliflozin blocks glucose resorption in the proximal tubule of the kidney and promotes glucosuria [84, 85, 86], but it also mediates antifibrotic effects in an AMPK$\alpha$-dependent manner. The AMPK$\alpha$ activity reduction has been observed in failing human and animal hearts, and it correlates to cardiac fibrosis [87, 88]. At the molecular level, it was shown that AMPK$\alpha$ activation mediated by dapagliflozin inhibits the up-regulation of the TGF-$\beta$/SMAD pathway detected in diabetic hearts, reducing proliferation, activation, and collagen production of CFs when stimulated with high glucose [59].

MCC950 is a selective inhibitor of the inflammasome complex NOD-like receptor family pyrin domain containing 3 (NLRP3), that has been proposed to reverse transverse aortic constriction (TAC)-induced cardiac remodelling. It is known that the prevention of myocardial remodelling is possible by suppressing the expression or activation of NLRP3 before TAC surgery [89]. Zhao et al. [60] have reported that NLRP3 activation promotes pathological cardiac remodelling due to pressure overload in a mouse model, whereas MCC950 treatment significantly attenuated fibrosis. Specifically, it reduced $\alpha$-SMA levels, blocked the TGF-$\beta$/SMAD4 pathway, and reduced cardiac inflammation, and macrophage infiltration. At the same time, MCC950 reduced calcineurin expression and MAPK activation, affecting and attenuating cardiac hypertrophy both around peripheral blood vessels and in the myocardial interstitium [60]. The action of MCC950 is exerted also on the microenvironment by inhibiting the release of inflammatory factors IL-18 and IL-1$\beta$. This evidence was studied in vivo on MCC950-treated post-MI mice, and the marked reduction of collagen I, collagen III, and $\alpha$-SMA levels was clear. The reduction of inflammatory cell infiltration was shown by the expression of cleaved IL-1$\beta$ and IL-18 decrement. These findings confirm the multi-pathway action of this drug in opposing cardiac fibrosis, partly mediated by suppressing the microenvironment inflammation [90].

A new strategy proposed to prevent the development of cardiac fibrosis is the intervention on ECM protein polymerization. Specifically, the small peptide pUR4, a fibronectin polymerization inhibitor, has been tested in mouse models of ischemia/reperfusion (I/R) injury to attenuate pathological remodelling, tissue fibrosis, and cardiac dysfunction. In vitro, pUR4 administration to murine and human myofibroblasts led to a reduction of cell proliferation and migration, as well as ECM protein deposition, thus ameliorating pathological features [61].

A recent study aimed to demonstrate how pharmacological interference with mechano-sensing cues could oppose stromal cell activation and myofibroblast differentiation. Indeed, mechano-sensing signalling due to myocardial stiffening plays a key role in remodelling progression. In particular, the shuttling of the main Hippo transcriptional component, the YAP (Yes-associated protein)/TAZ (transcriptional coactivator with PDZ-binding motif) complex, exerts homeostatic control of cardiac matrix, and a specific function in myocardial remodelling after injury [91, 92]. Garoffolo et al. [62] tested treatment with Verteporfin, a drug known to prevent the association of the YAP/TAZ complex with their cognate transcription factors TEA domain transcription factors (TEADs), on human fibroblasts activation and differentiation. The results revealed prevention of the fibrotic process by significant downregulation of more than half of the genes induced by TGF-$\beta$ exposure. Verteporfin was also tested in vivo in mice with permanent cardiac ischemia, significantly reducing indexes of fibrosis and morphometric remodelling [62]. In addition, Francisco and colleagues [63] have demonstrated that YAP activation in CFs triggers cell proliferation and expression of pro-fibrotic genes through the myocardin related transcription factor A (MRTF-A) transcriptional coactivator. Inducing YAP deletion in fibroblasts also prevented myofibroblast transition, and the development of cardiac fibrosis and dysfunction in mouse models. These findings further emphasize the potential of targeting YAP as a strategy for anti-fibrotic therapies [63].

It is worth mentioning that the expanding knowledge on the mechanism of action of commonly prescribed drugs may represent a starting point as repurposing strategy towards cardiac fibrosis. Statins are the most common drugs used to inhibit cholesterol biosynthesis. Recently, several studies have demonstrated their capability of exerting other beneficial effects on cardiovascular diseases [93, 94]. Specifically, atorvastatin was studied as a potential drug to ameliorate cardiac fibrosis. It was used to treat CFs in vitro and achieved reduction of Heat Shock Protein 47 (HSP47) levels. Indeed, HSP47 is a collagen specific molecular chaperon, normally induced by TGF-$\beta$. In vivo, treated Dahl salt-sensitive rats showed indexes of cardiac fibrosis reduction associated with improvement of diastolic function after atorvastatin administration [95].

In recent years, numerous studies have suggested that epigenetic regulation, including histone methylation, has extensive roles in fibrosis [124]. Wang et al. [125] studied how the lysine demethylase 5B (KDM5B), a histone H3K4me2/me3 demethylase, can affect the pathophysiological events in cardiac fibrosis. Initially, they demonstrated KDM5B upregulation in CFs in response to pathological stress. Secondly, it was proposed the molecular mechanism through which KDM5B exerts its action: it demethylates the activated H3K4me2/3 modification bounding the activating transcription factor 3 (Atf3) promoter, inhibiting its expression. In the literature [126] Atf3 is considered an antifibrotic regulator of cardiac fibrosis, and its suppression is correlated with the activation of TGF-$\beta$ signalling. Finally, the authors prevented the fibroblast transition into myofibroblast by inhibiting KDM5B in vitro and in vivo, thus proposing KDM5B as possible candidate target to ameliorate cardiac fibrosis and remodelling [125].

Another fundamental epigenetic modulation is histone acetylation/deacetylation, that enhances or opposes the accessibility of transcriptional factors, respectively. It is known that small molecules acting as histone deacetylase (HDAC) inhibitors play a role in the attenuation of cardiac remodelling and fibrosis [127]. Nural-Guvener et al. [128] proposed a strategy through selective class I-HDAC inhibition against CFs activation. In particular, they treated cluster of differentiation 90-positive (CD90+) CFs in vitro with mocetinostat (an HDAC inhibitor), showing reduction of collagen III and $\alpha$-SMA. Moreover, they demonstrated the efficacy of selective class I-HDACs inhibition in reducing fibrosis in a congestive heart failure animal model, as well as its effect in reversing interstitial fibrosis and improving cardiac function after MI [128].

In a previous study, it was demonstrated that a group of epigenetic reader molecules called bromodomain and extra-terminal (BET) acetyl-lysine binding proteins, play a crucial role in regulating pathological fibrosis [129]. In particular, Stratton and Bagchi et al. [130] identified by RNA-sequencing the bromodomain-containing protein 4 (BRD4) involvement in cardiac fibrosis promoting the activation of quiescent CFs into Periostin (Postn)-positive cells. The authors validated the pro-fibrotic role of BRD4 using the BET bromodomains inhibitor JQ1. Specifically, cultured primary adult rat ventricular fibroblasts stimulated with TGF-$\beta$ in the presence of JQ1 showed reduced expression of fibrotic markers, such as $\alpha$-SMA [130]. Moreover, mice were treated with the JQ1 inhibitor after transverse aortic constriction, and the transcriptomic profiling of the CFs isolated from the hearts suggested a significant downregulation of Postn mRNA expression.

The pivotal regulatory role of non-coding RNAs (ncRNAs) in every biological process is well established and consolidated, and so it is for the process of cardiac fibrosis. In the last years, many ncRNAs, mainly microRNAs (miRNAs), long non-coding RNAs (lncRNAs) and more recently also circular RNAs (circRNAs), have been discovered as fine tuners of cardiac fibrosis, paving the way to their therapeutic targeting. Indeed RNA-targeting therapeutics have a high potential for the treatment of several diseases.

LncRNAs contribute to cardiac diseases, including fibrosis. One of the first lncRNAs reported to be dysregulated during chronic cardiac remodelling is the Maternally Expressed Gene 3 (Meg3), which was found specifically expressed in CFs and downregulated during late cardiac remodelling in mouse hearts after TAC [131]. Meg3 was known to act as a transcriptional regulator of gene expression, recruiting and guiding chromatin remodelling complexes to target sites [132, 133]. In 2017, it was identified as a regulator of the expression of the fibrosis-related gene matrix metalloproteinase-2 (MMP2) in CFs through the recruitment of p53, this latter induced by TGF-$\beta$ signalling; moreover, Meg3 silencing in vivo after TAC decreased myocardial fibrosis and improved diastolic function, thus revealing a new potential target for the prevention of pathological remodelling in heart diseases [131].

In 2019 Hao et al. [134] identified the lncRNA AK137033, named Safe (Sfrp2 antisense as fibrosis enhancer), specifically enriched in the nuclei of CFs. Its expression resulted up-regulated upon both MI induction in mouse hearts and TGF-$\beta$ treatment of CFs in vitro. Safe silencing in CFs also prevented cell proliferation, myofibroblast differentiation, and ECM proteins secretion. In addition, in vivo knockdown of this lncRNA in a mouse model of MI reduced the infarcted area and cardiac fibrosis, leading to the improvement of cardiac function. Safe was found to exert its pro-fibrotic effect mainly via inducing the expression of its neighbouring protein-coding gene secreted freezled related protein 2 (Sfrp2), which was proven to synergically act with Safe and the RNA-binding protein HuR in the induction of fibrosis-related genes in CFs [134].

The Nuclear Enriched Abundant Transcript 1 (NEAT1) is a lncRNA involved in cancer and some fibrosis-related diseases. Its expression was found elevated in patients with heart failure. Neat1 acts as a pro-fibrotic lncRNA, by inducing Smad7 promoter methylation through recruitment of enhancer of zeste homologue 2 (EZH2). Silencing of NEAT11 significantly alleviated the progression of cardiac fibrosis and dysfunction in a mouse model of heart failure, while overexpression of Neat1 had the opposite effect. This suggests that NEAT11 could be a good candidate for targeted antisense therapy [135].

The scaffold attachment factor B interacting lncRNA (SAIL) was shown to be reduced in cardiac fibrotic tissue and activated CFs. In vitro overexpression of SAIL decreased proliferation and collagen production of CFs, and could alleviate cardiac fibrosis in a mouse model of cardiac fibrosis due to MI. SAIL blocks the access of the scaffold attachment factor B (SAFB) to RNA pol II and reduces the transcription of fibrosis-related genes. The human conserved fragment of SAIL (hSAIL) could also suppress the proliferation and collagen production of human CFs [136].

The lncRNA taurine upregulation gene 1 (TUG1) was found to be overexpressed after MI, and this correlated with reduced levels of miiR-133b in rat models of myocardial fibrosis. Also, angiotensin-treated cardiac myofibroblasts revealed increased expression of TUG1. In addition, TUG1 knockdown prevented myofibroblast activation, while its overexpression increased cell proliferation and collagen production. Authors proved TUG1 could act as a sponge against miR133b in CFs, thus promoting the expression of CTGF, a target of miR133b. On the basis of this evidence, the axis TUG1/miR133b/CTGF could be a new potential target for the treatment of cardiac fibrosis [137].

Several miRNAs play a central role in the regulation of cardiac fibrosis. MiR-21 is highly expressed in cardiac cells, preferentially in non-myocytes, and it is known to be up-regulated and involved in the development of cardiac fibrosis, mainly promoting the activation of the ERK-MAPK pathway in CFs [138]. MiR-21 inhibition has been already proven effective for the prevention of cardiac fibrosis and dysfunction in small animal models of heart injury [139, 140]. Recently, Hinkel et al. [141] have tested the feasibility and anti-fibrotic efficacy of miR-21 silencing also in large animal models of heart failure, laying the foundation for potential applications in humans. In detail, the local, intracoronary delivery of locked nucleic acid (LNA)-antimiR-21 in pig hearts after I/R injury led to a reduction of adverse myocardial remodelling, associated with decreased fibroblast proliferation and macrophage infiltration, as well as improvement of cardiac function [141].

A critical miRNA in cardiac fibrosis is miR-125b, which has been proven to be necessary and sufficient for fibroblast-to-myofibroblast transition. MiR-125b expression was found to be enhanced in failing human hearts and in various animal models of cardiac fibrosis, as well as in CFs stimulated with TGF-$\beta$. Mechanistically, miR-125b targets the 3′-UTR of Apelin, a repressor of the fibrogenic pathway acting on the angiotensin II/TGF-$\beta$ axis, thus promoting myofibroblast differentiation and activation. Indeed, in vivo miR-125b silencing through LNA delivery caused the attenuation of angiotensin-induced interstitial fibrosis in murine hearts [142]. In addition, also p53 was found to be a target of miR-125b, which induced in this way fibroblast proliferation. An in-depth analysis also revealed that miR-125b was at the centre stage of the fibrosis signalling, regulating many other related pathways, therefore representing a valid new potential target for anti-fibrotic therapies.

Another regulator of fibroblast-to-myofibroblast transition was identified in miR-146b-5p, which is overexpressed in the blood of patients with myocardial ischemia, and in the heart of mice with MI. Interestingly, miR-146b-5p is up-regulated in CFs, endothelial cells, and macrophages by hypoxia-induced NF-$\kappa$B, independently from the classical TGF-$\beta$ pathway, thus driving the switch toward a myofibroblast phenotype. In vivo treatment with miR-146b-5p antagomir markedly reduced cardiac fibrosis and improved myocardial function in both murine and porcine models of MI [143].

In contrast, a miRNA with anti-fibrotic properties with a potential therapeutic application is miR-101. This miRNA was found to be down-regulated in the peri-infarct area of rats undergoing coronary artery ligation, as well as in CFs stimulated with angiotensin II. On the contrary, in vitro overexpression of miR-101 suppressed fibroblast proliferation and ECM remodelling protein production, mainly targeting c-Fos, a downstream factor of the TGF-$\beta$ signalling. The induction of miR-101 overexpression in vivo in post-infarct rat hearts remarkably alleviated interstitial fibrosis and improved cardiac performance [144].

Circular RNAs have been involved in the progression of both cancer and cardiovascular diseases. Given their peculiar secondary structure, with a covalent link bridging each end of the linear RNA which is converted into its circular form, circular RNAs are the most stable RNAs. This feature makes them optimal targets in RNA therapeutics.

In a rat model of post-ischemic myocardial fibrosis, circHNRNPH1 expression was found increased specifically in CFs. The expression of circHNRNPH1 could restrict the differentiation of CFs into myofibroblasts by de-repressing TGF-$\beta$ signalling. The anti-fibrotic action of circHNRNPH1 is exerted through sponging miR-216-5p. This allows SMAD7 de-repression and the degradation of TGF-$\beta$ receptor 1. Given the role of circHNRNPH1 in the TGF-$\beta$-mediated CF activation, this circRNA could be used to design new therapies for cardiac fibrosis linked to all cardiovascular diseases [145].

Similarly, circNFIB (which is conserved un humans) was found to be decreased in mouse cardiac samples post-MI, and its overexpression could attenuate CF proliferation [146]. Its action is dependent on TGF-$\beta$ stimulation, and is mediated by the ability to sponge miR-433. circNFIB showed an anti-fibrotic effect both in vivo and in vitro on CFs, thus making it another good candidate for RNA-based anti-fibrotic therapies.

The circHIPK3 expression markedly increased in CFs and heart tissues in a mechanism mediated by TGF-$\beta$, and initiated by treatment with angiotensin II. Homeodomain-Interacting Protein Kinase 3 (CircHIPK3) was shown to promote CF proliferation, migration, and subsequent tissue fibrosis by sponging miR-29b-3p. This determined upregulating of its target genes actin alpha 2 (ACTA2), COL1A1 and COL3A1. The silencing of circHIPK3 attenuates CFs activation by inhibiting proliferation, migration, and $\alpha$-SMA expression. Interestingly, a synergistic action of miR-29 overexpression and circHIPK3 silencing showed anti-fibrotic effects in vivo [147].

A recent study on patients with cardiac hypertrophy showed a significant decrease in yet another circRNA, that is the circYap isoform [148]. In a mouse model of hypertrophy, the expression of circYap was negatively correlated with the expression of fibrosis genes. The restoration of circYap via ectopic expression improved heart function and cardiac fibrosis in a pressure-overload mouse model of hypertrophy. CircYap expression could directly affect both survival and cell morphology of fibroblasts in vitro [148].

Finally, the circ_LAS1L was found to be down-regulated in acute MI patients and CFs in another clinical study. The circ_LAS1L functional role is to sponge miR-125b, thereby de-repressing this miRNA target genes involved in activation, proliferation, and migration of CFs [149].

Several adult cell populations have been studied as therapeutic and/or regenerative products for the injured heart, including selected resident cardiac mesenchymal populations. Cardiosphere-Derived Cells (CDCs) have been shown to have beneficial effects, mainly through a protective paracrine action [150, 151], mediated in part by the release of extracellular membrane vesicles (EMVs), such as exosomes [152]. Human and rat dermal fibroblasts were treated with cardiosphere-derived EMVs (CSp-EMV) and then injected into rat hearts subjected to left anterior descending artery ligation [153]. In vitro treated fibroblasts displayed a less fibrotic phenotype, with significantly reduced levels of fibroblast specific protein 1 (FSP1) and discoidin domain receptor 2 (DDR2), and showed increased secretion of pro-angiogenic factors, such as stromal-cell derived factor 1 (SDF-1) and vascular endothelial growth factor (VEGF). Moreover, they exerted pro-angiogenic and anti-apoptotic effects in vitro. The in vivo administration of exosome-primed fibroblasts attenuated adverse remodelling, reduced the infarct area, and enhanced cardiac function. Authors speculated this therapeutic activity could be due to the specific profile of miRNAs carried by the EMV, since a similar effect on cardiomyocytes was previously proven for exosomes from CDCs [152]. This is an interesting example of a biological product directly influencing the functional phenotype of fibroblasts, with a relevant therapeutic outcome.

Recently, also the curative effect of exosomes derived from cortical bone stem cells (CBSCs) has been tested in murine models of I/R injury. CBSCs were already proven to exert beneficial effects, in terms of decreased scar size and improved cardiac function, when injected in mouse and pig models of MI [154]. The same effects were recapitulated in I/R mouse hearts where CBSC-derived exosomes (CBSC-dEXOs) were administered, highlighting their key role in mediating cardio-protection [155]. Mechanistically, in vitro treatment of CFs with CBSC-dEXOs decreased the expression of pro-fibrotic genes, such as collagen I, collagen III, and MMP2, even to a greater extent than whole CBSC-conditioned medium from which the exosomes were isolated. In addition, CBSC-dEXOs were also proven to significantly reduce CF activation induced by TGF-$\beta$. After analyzing the RNA content of CBSC-dEXOs, the authors suggested that this anti-fibrotic role could be exerted by small non-coding RNAs, mainly miRNAs: in fact, they caused the downregulation of ncRNAs implicated in ribosome stability and protein translation of factors necessary for the fibroblast-to-myofibroblast transition.

Cardiac regeneration aims at restoring functional cells that have been lost in the damaged heart. A therapeutic option to recover lost cardiomyocytes is the direct reprogramming of fibroblasts into cardiomyocytes through genetic tools or small molecules. This strategy could be also viewed as a way of simultaneously replacing potentially fibrotic cells with new parenchymal cells, thus representing also an anti-fibrotic approach. Several strategies have been proposed for the delivery of direct reprogramming factors in recent years, and we briefly summarize a selection based on artificial vectors/carriers.

Mónica P. A. Ferreira et al. [156], developed a functionalized spermine-modified dextran-based (AcDXSp) nanoparticle that allows the release of two molecules—CHIR99021 and SB431542—capable of inducing fibroblast reprogramming. Further functionalization of the nanoparticles with atrial natriuretic peptide (ANP) allowed tropism towards cardiac cells [157]. Specifically, AcDXSP nanoparticles were incubated with non-myocytes, and the effect of the release of the active molecules was confirmed. In fact, the administration of CHIR99021 significantly enhanced $\beta$-catenin stabilization, increasing its levels in both cytoplasm and nucleus. Instead, the release of SB431542 (a TGF-$\beta$ inhibitor) prevented the intracellular translocation of SMAD3 to the nucleus, and inhibited its phosphorylation by anaplastic lymphoma kinase (ALK) [158]. This evidence thus supported the efficacy of this nanoparticle-based system for the direct reprogramming of fibroblasts into cardiomyocytes as a potential cardiac therapy [156].

Poly (lactic-co-glycolic acid) (PLGA)-polyethyleneimine (PEI) nanocarriers encapsulated with microRNAs represents a low-toxicity system to induce the direct reprogramming of cardiac human fibroblasts into cardiomyocyte-like cells. These nanoparticles were loaded with miR-1 and miR-133a, that initiate and drive muscle development and differentiation [159, 160]. Direct fibroblasts reprogramming was evaluated by the expression of markers such as cardiac troponins and $\alpha$-actinin, confirming this strategy as a valid method for in vivo delivery and fibroblast targeting [161].

Qiaozi Wang et al. [162] have developed another non-viral system to reprogram CFs in situ, by coating neutrophil-mimicking membranes on mesoporous silicon nanoparticles (MSNs) with the FH peptide, and loading them with miR-1, -133, -208, and -499 (miR-Combo). Homing of the nanoparticle to the injured heart was achieved by the natural inflammation-homing capacity brought by neutrophil membranes, while the FH peptide’s high affinity to tenascin-C (TN-C), specifically expressed by CFs in the injured hearts, granted cell tropism. This study reported the intravenous injection of the nanoparticles loaded with the miR-combo in a mouse model of myocardial I/R, and evaluated the efficient in situ reprogramming of CFs into cardiomyocyte-like cells. Despite low rates of reprogramming, this strategy of delivering showed nonetheless an improvement in cardiac function, and attenuation of fibrosis in vivo [162].

Recently an innovative strategy based on chimeric antigen receptor T-cells (CAR-T) was proposed as an additional method to reduce cardiac fibrosis. The gene signature of CFs derived from healthy versus diseased human hearts were analysed, and fibroblast activation protein (FAP) was found as a specific marker of activated fibroblasts which is not expressed in cardiomyocytes. The study suggested that the adoptive transfer of antigen specific CD8+ T cells targeting the FAP protein can lead to partial removal of activated CFs in vivo, with significant reduction of cardiac fibrosis and restoration of function after injury in mice [163]. Joel G Rurik et al. [164] developed a therapeutic approach to generate transient CAR-T cells directly in vivo by delivering modified messenger RNA (mRNA) in T cell–targeted lipid nanoparticles (LNPs). Specifically, they injected CD5-targeted LNPs in a mouse model of heart failure, favouring the delivery of modified mRNA encoding for CAR in T lymphocytes, thus producing CAR-T cells in vivo. The transient CAR-T cells generated were specific for the FAP protein, so their action reduced the abundance of activated fibroblasts and consequently fibrosis, also restoring cardiac function after injury [164].