Types of cardiac fibrosis

Cardiac fibrosis is classified into two types: reactive fibrosis and reparative (replacement) fibrosis. Reactive fibrosis, which is characterized by the excessive deposition of ECM in the interstitial or perivascular spaces, is triggered by hemodynamic stress, such as pressure overload, and it is not directly associated with CM death [6, 7]. Reactive fibrosis is considered an adaptive response aimed at normalizing the increased wall stress and preserving the cardiac output. However, excessive fibrosis in interstitial spaces may cause mechanical stiffness, resulting in cardiac diastolic dysfunction, and impairment in electric conduction by forming a barrier between CMs, leading to cardiac systolic dysfunction. In addition, excessive fibrosis in perivascular areas decreases the flow of oxygen and nutrients, leading to an energy-starved condition in the myocardium [3]. Therefore, reactive cardiac fibrosis is closely associated with physiological and pathological cardiac conditions. Reparative fibrosis, which occurs in response to the loss of viable myocardium and forms a scar, maintains the structural integrity of the ventricles. A balance between reactive and reparative fibrosis is important for the prevention of excessive and inappropriate cardiac dysfunction, particularly after CM death due to cardiac injury, such as MI [8].

Cardiac fibroblasts

Cardiac fibroblasts are flat, spindle-shaped cells located in the myocardium, with multiple processes originating from the cell body, and lack a basement membrane [3]. They play a major role in cardiac fibrosis by producing the ECM [3], and recent studies have demonstrated that mouse cardiac resident fibroblasts derived from the cells of the embryonic proepicardial organ (PEO) [9, 10] are major cell type producing the fibrotic ECM in a pressure overload model [11, 12]. However, other cell types have also been reported as origins of cardiac fibroblasts such as embryonic endothelium, which undergo endothelial-to-mesenchymal transition (EndMT) [13], circulating bone marrow cells [14], pericytes, and endothelial cells [15]. Because these cardiac fibroblasts lack a specific marker [11, 16, 17], investigating their regulation remains a challenging task.

Paracrine factors associated with cardiac fibrosis

Transforming growth factor-beta (TGF-β) and angiotensin II (Ang II) are major factors that regulate cardiac fibrosis (Fig. 1). The expression of the Ang II type 1 (AT1) receptor is greater in fibroblasts than in CMs [18]. The activation of the AT1 receptor in fibroblasts by Ang II leads to the secretion of TGF-β, which stimulates fibroblast proliferation and ECM protein synthesis in an autocrine manner [19, 20] and induces CM hypertrophy in a paracrine manner [18]. The infusion of a subpressor dose of Ang II into mice induces both cardiac hypertrophy and fibrosis [21]. Clinical studies have demonstrated that the blockade of RAS signaling by an ACE inhibitor or ARB effectively reduces cardiac fibrosis and remodeling and that this is independent of the blood pressure-lowering effect [22]. However, the concomitant use of aliskiren, the direct renin inhibitor, with an ACE inhibitor or ARB in post-MI patients with reduced left ventricular (LV) ejection fraction does not further attenuate LV remodeling but is instead associated with more adverse effects [23]. The effect of blocking RAS signaling for cardiac fibrosis may eventually reach a plateau, with an excessive RAS blockade increasing adverse effects. Therefore, the appropriate regulation of RAS signaling is important for the prevention of cardiac fibrosis without any adverse effects.

TGF-β plays an essential role in cardiac fibrosis. Treatment with a subpressor dose of Ang II does not induce cardiac hypertrophy or fibrosis in Tgfb1-deficient mice [24]. Therefore, Ang II-induced cardiac fibrosis is believed to be mediated, at least in part, by TGF-β. Although cardiac hypertrophy and fibrosis induced by TGF-β signaling are adaptive responses to acute stress [3], the inhibition of TGF-β signaling may be useful for treating cardiac fibrosis. Therapies targeting TGF-β signaling have already been investigated in various mammalian models. An intraperitoneal injection of a TGF-β neutralizing antibody into rats subjected to pressure overload not only inhibits fibroblast activation and cardiac fibrosis but also prevents diastolic dysfunction [25]. In contrast, in a mouse aortic banding-induced pressure overload model, an orally active, small molecule inhibitor of the TGF-β type I receptor (TGFBR1, also known as activin receptor-like kinase 5), SM16, attenuates the development of cardiac fibrosis but causes death because of rupture at the site of aortic banding [26]. Further studies using other models of hypertension-induced cardiac fibrosis, which are independent of aortic banding, should be conducted. An MI model has been used to evaluate the effects on cardiac fibrosis and function. The treatment of rats with GW788388, another orally active TGFBR1 inhibitor, 1 week after MI, significantly reduces TGF-β signaling and attenuates LV remodeling and systolic dysfunction [27]. However, an intraperitoneal injection of a TGF-β neutralizing antibody started either 1 week before or 5 days after MI increases mortality and exacerbates LV dilatation and contractile dysfunction in mice [28]. These results indicate that the consequences of inhibiting TGF-β are variable, depending on the disease model and the timing of inhibition, presumably because TGF-β signaling in the heart during stress plays different roles during the early and late phases of cardiovascular disease.

Cardiac hypertrophy induced by cardiac fibroblasts

Some paracrine factors from cardiac fibroblasts induce CM proliferation and/or hypertrophy. Embryonic, but not adult, cardiac fibroblasts secrete high levels of fibronectin, collagen III, and heparin-binding EGF-like growth factor in mice. These embryonic cardiac fibroblast-specific factors collaboratively interact and promote embryonic CM proliferation (Fig. 2) [29]. On the other hand, in adult mice, various paracrine factors secreted by cardiac fibroblasts, including TGF-β, induce CM hypertrophy but not proliferation; the Krüppel-like factor 5 (KLF5) transcription factor expressed in adult cardiac fibroblasts promotes CM hypertrophy and cardiac protection (Fig. 1). KLF5 transactivates the expression of platelet-derived growth factor A (PDGF-A), which leads to the migration and proliferation of fibroblasts in an autocrine manner. Further, KLF5 transactivates insulin-like growth factor-1 (IGF-1) to promote CM hypertrophy in a paracrine manner. The cardiac fibroblast-specific deletion of Klf5 ameliorates cardiac hypertrophy and fibrosis elicited by a moderate-intensity pressure overload [30]. On the other hand, a high-intensity pressure overload causes severe heart failure and early death in these mice. Furthermore, in wild-type mice, the administration of a peptide inhibitor of IGF-1 severely exacerbates heart failure induced by a high-intensity pressure overload. These results demonstrate that cardiac fibroblasts play pivotal roles in cardiac adaptive responses to pressure overload, which are, at least in part, regulated by IGF-1.

Cellular sources of cardiac regeneration

The regenerative capacity greatly differs in adult mammalian organs, and organ-specific stem cells have been shown to contribute to regeneration in certain organs, such as the intestines, lungs, taste buds, and hair follicles [31–34]. In the mammalian heart, CMs rapidly proliferate during embryonic development; however, CMs exit the cell cycle, with the number of binucleated CMs increasing soon after birth [35]. One-day-old mice retain an adequate CM proliferative capacity and can completely regenerate CMs after cardiac injury, such as apical resection and MI. In response to cardiac injury, inflammation causes the proliferation of myofibroblasts and increases fibrosis in the regenerative area, but the myocardium is finally regenerated without fibrosis. Therefore, cardiac fibroblast-rich scar tissue may be an important component of cardiac repair in neonatal mice [17]. However, this efficient regenerative potential is lost within the first week of postnatal life [36, 37]; adult mice do not regenerate CMs adequately to compensate for the impaired cardiac function, inducing reparative fibrosis after injury instead. On the other hand, in certain lower vertebrates, such as teleost fish and urodele amphibians, adult CMs have sufficient regenerative capacity, and the myocardium can completely regenerate after injury without forming scar tissue [38, 39]. It is not known what causes these differences in the regenerative capacity of CMs between adult mammals and lower vertebrates. The fact that CMs in lower vertebrates are mononucleated and smaller in size with fewer myofibrils than those in adult mammals may be responsible for the differences observed in the CM regenerative capacity between these groups [40].

Impact of reactive oxygen species on CM regeneration

Recently, considerable attention has been given to the impact of reactive oxygen species (ROS) on cardiovascular diseases. Cardiac injury has been shown to increase the amount of ROS in the heart, which induces CM cell cycle arrest via the activation of responses to DNA damage (Fig. 2) [68, 69]. The inhibition of ROS by pretreatment with N-acetyl-L-cysteine has been shown to promote CM regeneration after ischemia–reperfusion injury even in 21-day-old mice [69]. In addition, the presence of oxygen in the environment has been reported to influence the production or scavenging of ROS and regeneration of CMs. Hyperoxic (100% oxygen) and hypoxic (15% oxygen) environments have been found to diminish and enhance CM proliferation, respectively, in neonatal mice with adequate CM regenerative capacities (Fig. 2) [69]. Furthermore, in adult mice, gradual exposure to severe hypoxia after MI, in which inspired oxygen is gradually decreased by 1% beginning 1 week after MI for 2 weeks, and then maintained at 7% for another 2 weeks, has been found to induce CM regeneration and coronary angiogenesis, resulting in improvements in the LV systolic function [70]. To evaluate the proliferation of hypoxic CMs in the adult heart, hypoxic CMs in αMHC-creERT2-ODD; R26R/tdTomato mice were genetically labeled at 2 months of age and fate mapped for 1 month under normal conditions; the results of this study demonstrated that labeled hypoxic CMs have a higher proliferative capacity than unlabeled CMs and can be a source of newly generated CMs [71].

Role of macrophages in cardiac regeneration

One-day-old mice can completely regenerate their hearts after MI injury. However, 14-day-old mice do not retain sufficient capacity for cardiac regeneration and cause fibrosis in response to cardiac injury. Clodronate liposome-mediated depletion of monocytes/macrophages in 1-day-old mice after MI reduces the angiogenic response, blocks the cardiac regenerative capacity, and induces cardiac fibrosis and dysfunction [72]. To identify the role of cardiac monocytes/macrophages in cardiac regeneration, immunophenotyping and gene expression profiling of cardiac monocytes/macrophages from 1-day-old and 14-day-old mice were isolated and compared after MI [72]. Regenerative macrophages from 1-day-old mice displayed both M1- and M2-associated gene transcription patterns and expressed more chemokines, proangiogenic factors, and oxidative stress responders, which may facilitate the formation of new myocardium than macrophages from 14-day-old mice.

Embryonic-derived resident cardiac macrophages (MHC-II^lowCCR2⁻) and two types of resident cardiac macrophages (MHC-II^lowCCR2⁻ and MHC-II^highCCR2⁻) are the major populations of monocytes/macrophages in neonatal and adult mouse hearts, respectively; monocytes (MHC-II^lowCCR2⁺) and monocyte-derived macrophages (MHC-II^highCCR2⁺) are not abundant in either neonatal or adult hearts under normal physiological conditions [73]. To elucidate essential cardiac monocyte/macrophage subsets involved in cardiac regeneration, Lavine et al. used a diphtheria toxin receptor-mediated CM ablation mouse model [73], in which cardiac injury was induced without concomitant systemic inflammation. In response to diphtheria toxin receptor-mediated cardiac injury, the neonatal heart selectively expanded the population of embryonic-derived resident cardiac macrophages and cardiac dysfunction recovered to baseline. In contrast, in adult mice, the heart recruits CCR2⁺ pro-inflammatory monocytes and monocyte-derived macrophages and loses CCR2⁻ resident cardiac macrophages after cardiac injury; cardiac function recovery was not observed. However, the administration of selective CCR2 inhibitors in adult mice after cardiac injury inhibited CCR2⁺ monocyte recruitment to the heart and preserved CCR2⁻ resident cardiac macrophages, resulting in reduced inflammation and enhanced angiogenesis. Collectively, embryonic-derived resident cardiac macrophages are key mediators of angiogenesis, leading to cardiac regeneration in response to cardiac injury (Fig. 3).

Interactions between endothelial cells and fibroblasts

EndMT is a fundamental cellular mechanism that regulates embryonic development and fibrotic diseases. During the embryonic development of the heart, the endocardium undergoes EndMT and forms an atrioventricular cushion: the primordial valves and septa of the adult heart [74]. Zeisberg et al. demonstrated that Tie1-expressing endothelial cells in the adult heart underwent EndMT and differentiated into fibroblasts during cardiac fibrosis in response to pressure overload [15]. Endothelial cells undergoing EndMT lost tight junctions that hold neighboring cells, gained the ability to move, and contributed to the total pool of cardiac fibroblasts. Although endothelial cells are not major origins of cardiac fibroblasts under normal conditions, inflammation induces EndMT of endothelial cells. As a result, approximately one-third of all cardiac fibroblasts originated from endothelial cells in the fibrotic heart in response to pressure overload. During this process, TGF-β1 induces EndMT, whereas bone morphogenic protein 7 (BMP-7) prevents EndMT and preserves the endothelial phenotype [15]. Therefore, the anti-fibrotic effects of recombinant human BMP-7 have been investigated. An intraperitoneal injection of recombinant human BMP-7 inhibited EndMT and the progression of cardiac fibrosis and improved diastolic cardiac function in a moderate-intensity pressure overload model. Furthermore, the inhibition of EndMT and cardiac fibrosis by recombinant human BMP-7 has been observed in a mouse model of chronic heart rejection caused by heterotopic heart transplantation with a class II major histocompatibility mismatch between donor and recipient [15].

Cardiac fibroblasts can undergo mesenchymal-to-endothelial transition (MEndT) immediately after ischemic cardiac injury [75]. Approximately 30% of fibroblasts in the injury zone undergo MEndT, and fibroblast-derived endothelial cells exhibit anatomical and functional characteristics of native endothelial cells and contribute to angiogenesis of the injured heart. p53, a transcription factor, regulates MEndT in cardiac fibroblasts [75]. The loss of p53 in Col1a2-expressing fibroblasts severely decreases the formation of fibroblast-derived endothelial cells, reduces the post-MI vascular area, and worsens cardiac function. Conversely, the stimulation of the p53 pathway after ischemic cardiac injury by an intraperitoneal injection of the small molecule: reactivation of p53 and induction of tumor cell apoptosis (RITA), which inhibits ubiquitin-mediated p53 degradation, augments MEndT, enhances angiogenesis, and improves cardiac function. However, although cardiac fibroblasts cultured in vitro under serum-free conditions have been found to form tubular structures resembling the endothelial cell architecture and express endothelial markers, cardiac fibroblasts cultured under serum-fed conditions fail to generate tubular structures, even when p53 is artificially overexpressed. This result suggests that p53 expression alone is insufficient to induce MEndT and that the microenvironment, growth factors, and other signals are involved in this process. Collectively, these close interactions between endothelial cells and fibroblasts regulate cardiac fibrosis and angiogenesis (Fig. 3), and the regulation of both EndMT and MEndT is a potential therapeutic target for enhancing cardiac repair.

Direct reprogramming of cardiac fibroblasts into CMs

In 2006, Takahashi and Yamanaka generated induced pluripotent stem (iPS) cells from mouse fibroblasts by introducing four factors: Oct3/4, Sox2, c-Myc, and Klf4 [76]. Subsequently, the direct reprogramming of fibroblasts by lineage-specific transcription factors into the primary functional cells of each organ, such as neurons, hepatocytes, and renal tubular epithelial cells, was accomplished [77–80]. Further, the direct reprogramming of mouse cardiac fibroblasts into CMs is induced by a combination of cardiac-specific transcription factors (Gata4, Mef2c, and Tbx5) in vitro [81]. Furthermore, endogenous cardiac fibroblasts were directly reprogrammed into CMs by the retrovirus-mediated delivery of cardiac-specific transcription factors in vivo, with such newly generated CMs reducing scar formation and cardiac dysfunction after MI [49, 50]. Several laboratories demonstrated that in vivo reprogramming yields higher quality of CMs than in vitro reprogramming. These results suggest that factors within the native microenvironment, such as the ECM, growth factors, local signals, and mechanical forces, enhance the maturity of CMs in the heart.

Although the direct reprogramming of cardiac fibroblasts into CMs in vivo can be a new cardiac regenerative therapy (Fig. 2), the efficiency of reprogramming is currently low to adequately improve cardiac function, and the mechanisms of reprogramming and properties of newly generated CMs have not yet been fully defined [82]. Therefore, the modification of transcription factors and induction of microRNAs have been studied, with the goal of improving the quality of cardiac reprogramming [50, 83]; the addition of factors that regulate the native microenvironment may enhance the efficacy of cardiac direct reprogramming.