Cardiac myofibroblast and fibrosis

: Fibroblasts are differentiated to myofibroblasts and produce collagen and other extracellular matrix when the heart is exposed to stresses. Myocardial infarction and pressure overload-induced hypertrophy are major stresses to induce differentiation of fibroblasts. Since collagen can compensate the missing tissue due to injury, appropriate production of collagen is beneficial for the injured heart against rupture. However, excessive deposition of collagen is called fibrosis and causes cardiac dysfunction. After fibroblasts are differentiated to myofibroblasts, myofibroblasts can further change their phenotypes. In addition, myofibroblasts are found to have a new function other than collagen production. Myofibroblasts have macrophage-like functions that engulf dead cells and secrete anti-inflammatory cytokines. So far, research on fibroblasts has been delayed due to the lack of available markers for selective isolation of fibroblasts. In recent years, it has become possible to genetically label fibroblasts, sequence the cells at single cell levels, and manipulate function or the number of cells. Based on new technologies, the origin of fibroblasts and myofibroblasts, time-dependent changes of fibroblast states after injury, and heterogeneity have been demonstrated. Here, I will introduce recent advances in fibroblasts and myofibroblasts.


Introduction
Cardiac fibrosis is defined as the state with excess extracellular deposition of collagens and extracellular matrix [1]. It occurs when the heart is exposed to stresses such as ischemic injury and chronic high blood pressure. Since fibrosis causes cardiac dysfunction, it is a target for treatment with drugs, medical devices or tissue transplantation. Collagen and extracellular matrix are produced by myofibroblasts that are differentiated mainly from resident fibroblasts. Manipulation of activity and number of myofibroblasts is proposed to be important for inhibition of progression to more severe fibrotic states or recovery from fibrotic state [2]. Since myofibroblasts are an important player in inflammation and fibrosis after cardiac injury [3], it is an urgent need to understand the origin, function and fate of myofibroblasts. Recent technological advances reveal some of these issues. progress of research topics for fibroblasts, myofibroblasts and fibrosis will be reviewed.

Classification of fibroblasts
There are several types of cells in the heart such as cardiomyocytes and immune cells.
They interact with each other to regulate homeostasis in healthy and diseased conditions [4]. Among several types of cells, fibroblasts are unique, since they produce extracellular matrix that supports morphological integrity at resting state. Recent histology-based and flow cytometric methods have demonstrated that fibroblasts account for about 13% of cells in the mouse heart [5,6]. When the heart is exposed to injury such as myocardial infarction and hypertrophy, fibroblasts differentiate into myofibroblasts and produce extracellular matrix. Myofibroblasts are only the cells that produce extracellular matrix.
Excess deposition of extracellular matrix causes fibrosis leading to tissue dysfunction. It has been recognized that management of the number or function of myofibroblasts is important for treatment of fibrosis.
Origin of myofibroblasts is analyzed by labeling the various types of cells with reporter genes under cell-specific promoter (lineage-tracing experiment) [7]. Genetic labeling has advantage over immunological detection of marker proteins, since maker proteins sometimes disappear during development or differentiation. In addition, marker proteins are expressed in not only cells that are analyzed but also functionally irrelevant cells.
Lineage tracking experiment is to express reporter gene under the control of cell-type specific promoter. It labels the cells permanently even after the promoter activity turns off and then allows to trace the cells that promoter is once activated.
Lineage-tracing experiments showed that about fibroblasts present in the left ventricle and ventricular septum are derived from endocardial cells via endothelial-mesenchymal transition (EndoMT) and epicardial cells through epithelial-mesenchymal transition (EMT) [8]. A small number of fibroblasts were generated by differentiation of neural crest cells. However, mature endothelial cells, epicardial cells, or bone marrow-derived cells 3 did not contribute to population of fibroblasts. After labeling fibroblasts using cell-type specific promoter-reporter gene, it was investigated whether fibroblasts exhibit different functions depending on their origin. After pressure overload, epicardial-derived fibroblasts (labeled with Tbx18 promoter-GFP) and endocardial-derived cells (labeled with Tie2 promoter-GFP) were isolated. Analysis of RNAs showed similar expression profiles between epicardial-derived and endocardial-derived cells [9][10][11]. In addition, these two groups of cells had similar proliferative activity [12]. Therefore, it was concluded that there is no significant relationship between function of fibroblasts and origin of fibroblasts.
Molkentin's group used genetic labeling technique to trace the changes in the characteristics of fibroblasts over time after myocardial infarction [13]. They found that fibroblasts change their properties four different states after myocardial infarction, that is, resting fibroblasts, active fibroblasts, myofibroblasts and matrifibrocytes ( Figure 1). Myofibroblasts produced collagen and α-smooth muscle actin (α-SMA). Matrifibrocytes are unique, since they localized at scar and express the genes detected in tendons, bones and cartilage. The physiological implications of these bone-related genes in the heart are not known. Since matrifibrocytes localize only at scar, it suggests that they play a special role in scar formation or maintenance. Matrifibroblasts have several unique properties.
When diphtheria toxin was administrated to mice with the toxin receptor expressing in myofibroblasts, matrifibrocytes were resistant to killing by the toxin [13]. Susceptibility to diphtheria toxin treatment was different from resting and active fibroblasts. In addition, matrifibrocytes did not show proliferative activity when mice were treated with angiotensin II and phenylephrine. In this context, Kim's group reported an interesting finding. In diabetes, inter-α-trypsin inhibitor heavy chain 1 (ITIH1) secreted from liver was found to be responsible for systemic glucose intolerance. It blocked insulin action on adipose tissue and skeletal muscle (14). ITIH1 works as a glue to tighten the binding between extracelluar matrix. Anti-ITIH1 neutralizing antibody releases the inhibition of ITIH1 and recovered insulin sensitivity. Inaccessibility of diphtheria toxin to periostinexpressing cells at the late stage of fibrosis may be caused by ITIH1 or ITIH1-like molecule. Future studies are waited to reveal the roles of matrifibrocytes in cardiac fibrosis.

Heterogeneity of fibroblasts
Single-cell RNA sequencing (scRNAseq) is a relatively new but rapidly developing technology [15]. It allows to comprehensively characterize gene expression and relationships between individual cells. Single-cell analysis of 11,492 cells revealed heterogeneity of fibroblasts and cardiomyocytes during pressure overload-induced cardiac hypertrophy [16]. In their report, fibroblasts were grouped into 6 clusters: FB1 to  [19]. They found that myofibroblasts expressing αSMA are not detected. Instead, they identified two fibroblast subpopulations expressing the matricellular proteins Cilp and thrombospondin 4.
These results demonstrate the heterogeneity of fibroblasts but the relationship and identity between the cells assigned by different groups are unknown. Furthermore, their function and contribution to cardiac fibrosis remain to be determined in future. It was reported that endothelial cells are not a major source of fibroblasts in adult mouse heart [10]. However, there is a controversy on conversion of endothelial cells to myofibroblasts. There is a report that fibroblasts acquire the endothelial cell-like phenotype during ischemia-reperfusion [21,22]. A series of experiments using mice with genetically labeled fibroblasts demonstrated that 20-40% of fibroblasts express various markers of endothelial cells, and the isolated cells can form a capillary network.
Expression of p53 was essential for the process of conversion from fibroblasts to endothelial cells [21]. In addition, stimulation of p53 signaling improved cardiac dysfunction during ischemia-reperfusion. However, an opposite result was reported by different group, in which resident fibroblasts did not contribute to neovascularization after cardiac injury [22]. In the report, pulse chase labeling of fibroblasts after ischemiareperfusion showed that resident fibroblasts do not express the genes involved in angiogenesis, which are characteristics of endothelial cells. Origin of almost all endothelial cells was resident endothelial cells. Different approaches resulted in distinct conclusion. Thus, it may be necessary to confirm the findings with other techniques such as scRNAseq and proteomic analysis of isolated cells.

Myofibroblasts as phagocytes
Myofibroblasts have been recognized only as the cells that produce extracellular matrix such as collagen and fibronectin. They also interact with the inflammatory cells through

Signaling that controls differentiation to myofibroblasts
TGF-β is a strong inducer of differentiation of fibroblasts to myofibroblasts [25].
Inflammatory cells recruiting to the injury sites release cytokines including TGF-β. Injury cells also release alarmins and damage-associated molecular patterns (DAMPs) [26].
These molecules cause inflammation leading to differentiation of fibroblasts to myofibroblasts. Thus, any inhibition of inflammatory responses will block appearance of myofibroblasts, which eventually suppresses fibrosis. At early stage of myocardial infarction, neutrophils are firstly recruited to the injury sites [27,28]. Leukotriene B4 is a powerful attractant of neutrophils by binding to Leukotriene B4 receptor (BLT1).
Inhibition of neutrophil recruitment by BLT1 gene knockout or BLT1 blocker decreased cardiac fibrosis by inhibition of inflammation [29,30].
Interleukin (IL) is an important group of inflammatory cytokines. Multiple IL receptors are expressed in cardiac fibroblasts, which regulates fibroblast states and function [27].
The effects of proinflammatory ILs on cardiac fibroblasts ae blocked by cardiac fibroblast-specific deletion of IL receptors. Knockout of IL11 receptor or IL17 receptor genes reduced injury-induced cardiac fibrosis and cardiac dysfunction [31,32]. These results reveal that cardiac fibroblast-specific deletion of IL receptor genes decreases the 9 infiltration to or activity of immune cells in the injury area. It also suggests that cardiac fibroblasts play an important role in the regulation of injury-induced inflammation that is mediated by IL signaling.
After myocardial infarction, monocytes are mobilized from bone marrow and differentiated to macrophages at the injury sites [33]. Macrophages can be depleted by the treatment of mice with chlodronate-liposomes [34]. Macrophage-depleted mice showed decreased fibrosis and improved cardiac functions. These results show that single step of inflammation subsequent to myocardial infarction blocks fibrosis possibly through inhibition of differentiation to myofibroblasts.
TGF-β stimulation activates both canonical Smad2/3 and non-canonical MAP kinase signaling pathways [25]. Fibrotic responses were inhibited by knockout of fibroblastspecific TGF-β receptor 1/2 or knockout of transcription factor Smad3 that is activated downstream of TGF-β receptor [35]. It demonstrated that TGF-β-Smad2/3 signaling of fibroblasts is a major factor of cardiac fibrosis induced by pressure overload. It was interesting that knockout of TGF-β receptor 1/2 in fibroblasts also inhibited cardiac hypertrophy by pressure overload [35]. It indicates that myofibroblasts interact with cardiac myocytes through direct cell-cell communication or indirect mediator-mediated interaction, which is secreted from myofibroblasts.
In addition to TGF-β, lysophosphatidic acid (LPA) stimulation induces differentiation of fibroblasts to myofibroblasts [36,37]. LPA binds their own GPCRs to activate cellular responses [38]. In vitro, LPA stimulation induced fibrosis by activation of myocardinrelated transcription factor-serum responsive factor (MRTF-SRF) pathway. Rho regulated actin oligomerization that is regulated by phosphorylation of actin by Rho kinase (ROCK).

ROCK-mediated phosphorylation of actin increased monomeric form of actin that results
in release of SRF inhibition. SRF together with MRTF activated transcription of various genes including profibrotic genes [39]. Compound CCG-203971 is a small-molecule inhibitor of the Rho-mediated MRTF-SRF pathway [40]. Administration of CCG-203971 inhibited bleomycin-induced lung fibrosis. Rho is signaling molecule that is activated downstream of various receptors including GPCRs and TGF-β receptors. Inhibitors of Rho-Rock pathway may suppress fibroblast activation and fibrosis more efficiently than receptor inhibition. G protein-coupled Receptor Kinase 2 (GRK2) is known as a regulator of G proteincoupled receptors (GPCRs) by phosphorylating agonist-bound GPCRs [41].
Cardiomyocyte-specific knockout of GRK2 demonstrated that GRK2 ablation protects the heart against cardiac dysfunction and fibrosis by myocardial infarction [42]. These mice also showed the reduction of the development of heart failure by myocardial infarction. Fibroblast-specific knockout of GRK2 using the collagen 1α2 promoter reduced the secretion of TNFα and suppressed gene expression of profibrotic factors after ischemia-reperfusion [43]. Inhibition of fibroblast function improved cardiac dysfunction.
These results show that inhibition of GRK2 protects the heart against cardiac stresses in cardiomyocytes as well as in fibroblasts.
Transient receptor potential channel canonical 6 (TRPC6) is a voltage-independent cation channel and mediates angiotensin II-stimulated hypertrophic responses [44]. The increased intracellular Ca 2+ also plays an important role in conversion of fibroblasts to myofibroblasts. It activates cellular signaling that is sufficient for promoting conversion to myofibroblast and resulting fibrosis [45]. TGF-β-induced upregulation of TRPC6 was inhibited by blockade of p38-MAPK-mediated signaling [46]. It is interesting that TRPC6 knockout fibroblasts did not show changes in Ca 2+ signaling and did not promote conversion of fibroblasts to myofibroblasts when the cells are treated with angiotensin II and TGF-β. These results demonstrate that TRPC6-Ca 2+ signaling is essential for induction of myofibroblasts in cardiac fibroblasts.
There are other signaling molecules involved in induction of fibrosis. Deletion of βcatenin gene in cardiac fibroblasts improves cardiac function and reduces fibrosis, which is due to decreased production of extracellular matrix proteins by cardiac fibroblasts [47].
It was recently found that functional primary cilia in cardiac fibroblasts is required for canonical TGFβ signaling-induced differentiation of cardiac fibroblasts to myofibroblasts [48]. Primary cilia express polycystin-1 that is known as regulator of cell proliferation, cell migration, and interactions with other cells. Polycystin-1 is required for maintaining cellular structures of primary cilia. When primary cilia of cardiac fibroblasts were disrupted specifically by deletion of polycystin-1 gene, TGFβ-Smad3 signaling-induced extracellular matrix protein production and fibroblast differentiation were impaired.
Deletion of polycystin-1 gene enhanced pathological cardiac remodeling after myocardial infarction [49], suggesting the important role of primary cilia in hypertrophy and fibrosis.
It also suggested that primary cilia are functional and participate in TGFβ-induced fibrosis and myofibroblast differentiation. Heat shock protein (Hsp) is chaperon molecule for conversion of fibroblasts to myofibroblasts and resulting fibrosis. Among various Hsp proteins, Hsp47 is known as collagen-specific chaperone. Cardiac fibroblast-specific deletion of Hsp47 significantly reduced cardiac fibrosis and improved cardiac diastolic dysfunction after pressure overload [50]. However, the reduced collagen production in these mice increased lethality after myocardial infarction due to the insufficient scar formation.
Collectively, these results suggest that manipulation of expression and activity of various signaling molecules involved in fibrotic pathway modulates fibroblast states leading to alteration of fibrosis.

Control of differentiation of fibroblasts by extracellular environment
Fibroblasts exist in interstitial spaces between cardiomyocytes in healthy condition.
When the heart is exposed to stresses such as myocardial infarction and hypertrophy, fibroblasts differentiate to myofibroblasts and produce extracellular matrix such as collagen. Fibroblasts generated at the injury site actively proliferate and form aggregate.
An in vitro three-dimensional culture system that mimics several states of fibroblasts was developed [51]. When fibroblasts were isolated and cultured in two or three dimensions, the properties of fibroblasts were changed. Standard polystyrene-coated culture plates (not coated with collagen or other extracellular matrix) were used for two-dimensional (2D) cultures, and ultra-low adhesion plates (coated with any extracellular matrix) for 3D cultures ( Figure 1). Fibroblasts cultured with 3D structure plates formed spheres (spherical masses) within 24 hours. The morphology of fibroblasts formed by 2D and 3D cultures was reversible but did not depend on the tension or rigidity of the extracellular environment. Interestingly, a correlation of the expressing genes was found between 3D culture fibroblasts and the remodeling heart (treatment with isoprenaline for 3 weeks or cryo-injury treatment). However, the expression of α-smooth muscle actin (α-SMA) was decreased in the aggregates. Since the expression of α-SMA, a marker of myofibroblasts, is decreased, it cannot be said that fibroblasts obtained by 3D culture is conventional myofibroblasts. However, expression pattern of mRNA of 3D culture fibroblasts is similar to that of matrifibrocytes reported by Fu et al. [52]. Analysis of fibroblasts with 3D culture may elucidate function and fate of matrifibrocytes. Interestingly, 3D cultured fibroblasts reversibly changed their morphology and expressing genes by transferring the cells to 2D culture. Thus, 3D culture but not normal 2D culture provides a sufficient signal to trigger remodeling. Since in vitro system is essential for analyzing mechanism of differentiation and function of fibroblasts and myofibroblasts, the exchange of culture conditions may be a promising technique to analyze complex behavior of fibroblasts.
Fibroblast fate is also regulated by stiffness of extracellular matrix [53]. The stiffness around fibroblasts increases during progression of fibrosis. Increased stiffness was sensed by integrin receptors and actin cytoskeleton that promote translocation of p38-MAPK to the nucleus and stimulates remodeling [54]. Integrin-actin cytoskeleton signaling complex also activates tyrosine kinases such as focal adhesion kinase (FAK), Src and Fyn.
These kinases stimulate GDP-GTP exchange of Rho through guanine nucleotide exchange factors (RhoGEFs). It leads to activation of Rho-ROCK-MRTF-A pathway that increases gene transcription in concerted action with SRF [55]. This pathway will be described in more detail in YAP-TAZ signaling.

YAP-TAZ signal in fibroblasts
Hippo pathway is known for its inhibitory activity on cardiomyocyte proliferation [56].
Yes-associated protein (YAP) is a transcriptional coactivator in Hippo pathway that is negatively regulated by large tumor suppressor kinase 1 (Lats1) and Lats 2 ( Figure 3). Deletion of YAP from transcription factor 21 (Tcf21) and Col1a1-expressing fibroblasts decreased their collagen deposition, proliferation, and activation after myocardial infarction [57]. Similar decreases in angiotensin II/phenylephrine-treated fibrosis were seen when YAP was knocked down from fibroblasts [58]. Myocardin-related transcription factor A (MRTF-A) levels were decreased in YAP knockout mice, suggesting that MRTF-A expression was regulated by YAP function. These results demonstrate the importance of YAP-MRTF-A signaling for determination of myofibroblast states in response to ischemic and chronic stresses. Interestingly, cardiomyocyte specific deletion of YAP decreased hypertrophy and significantly increased fibrosis [59]. Thus, YAP protects the heart against ischemic stress or pressure overload. 8. Development of treatment of fibrosis and heart failure GPCR stimulation dissociates heterotrimeric G protein in Gα and Gβγ. Gallein is a small molecule that inhibits Gβγ-dependent GRK2 activation [62,63]. When administered in vivo, gallein suppressed cardiac injury after ischemia-reperfusion, activation of fibroblasts, and the onset of heart failure possibly by GRK2 inhibition [62]. Gallein could attenuate the activation responses of fibroblasts isolated from heart failure patients.
However, the affinity of gallein for Gβγ is low (Ki is about 0.2μM), non-specific effects of gallein should be examined by the experiments that gallein-mediated protective responses disappear in GRK2 knockout mice. From high throughput screen, paroxetine was identified as a GRK2 inhibitor that attenuated the development of heart failure by myocardial infarction [64,65]. Although paroxetine is a selective serotonin reuptake inhibitor (SSRI), another SSRI fluvoxamine did not inhibit GRK2 and did not attenuate heart failure [65]. These results shows that new GRK2-selective inhibitor is a therapeutic option for treatment of heart failure.
Cyclic nucleotides (cAMP and cGMP) play important roles in cellular signaling. Cyclic AMP and cGMP are generated by adenylyl cyclase or guanylyl cyclase, respectively.
Various drugs that increase cAMP or cGMP are already used in clinic for treatment of cardiovascular diseases. These nucleotides are degraded by phosphodiesterases (PDEs).
Each PDE has different selectivity for cAMP or cGMP, although selectivity depends on cellular environment. It should be noted that cGMP selective PDE degrades cAMP when the concentration of cAMP is very high. There are PDEs that degrade both cAMP and cGMP. PDE2 is a PDE that is activated by cGMP and degrades cGMP and cAMP [67].
Inhibition of PDE2 suppressed the onset of heart failure by inhibiting cGMP degradation [68]. The elevated cGMP activated protein kinase G leading to activation of transcription factors (Nuclear Factor of Activated T cells, NFAT) and inhibition of TRPCs. PDE2 inhibitor such as BAY 60-7550 preferentially promoted NO-guanylyl cyclase-cGMP signaling and suppressed the onset of heart failure [69]. However, it should be careful to use PDE inhibitor for the treatment of heart failure, since cAMP causes harmful effects on heart failure.
Bcl2 family is involved in apoptosis and is categorized into three groups based on the homology of amino acid sequence, that is Bcl2-like, Bax-like and BH3-only. Bcl2-like protein has anti-apoptotic activity [70]. In contrast to Bcl2-like protein, Bax-like and BH3-only proteins promote apoptosis. BH3 domains of these apoptosis-promoting proteins bind hydrophobic BH3-binding pocket of anti-apoptotic Bcl2 to form heterodimer and inhibit anti-apoptotic activity by Bcl-2 [71]. Navitoclax (also known as ABT-263) is a BH3-mimetic antagonist of the Bcl2-anti-apoptotic protein [72]. It promoted apoptosis of myofibroblasts isolated from the patients of scleroderma [73].
Since enhanced apoptosis of myofibroblasts is effective in recovery from fibrosis, navitoclax will be a promising drug for treatment of scleroderma. Venetoclax that has excellent selectivity for BCL-2 is currently in clinical trials as an anticancer drug [74]. In the future, venetoclax may be also applied to treatment for cardiac fibrosis.
Many studies report that fibrosis is inhibited by knockout of gene in myofibroblasts or cardiomyocytes. However, inhibition of ongoing fibrosis by compounds does not always recover the heart from fibrotic states. Metformin activates AMP-activated protein kinase (AMPK). It was reported that metformin recovers the lung from bleomycin-induced fibrosis [75]. The effects of metformin were proposed to promote apoptosis of myofibroblasts or inactivate myofibroblasts. Activation of AMPK with metformin or compound A769662 also suppressed hypertrophic responses of neonatal ventricular myocytes by phenylephrine stimulation [76].  In the heart, aquaporin transfers hydrogen peroxide (H2O2) from extracellular space to cytosol [77]. H2O2 entering into the cells modifies proteins and changes protein function that cause detrimental effects. Several isoforms of aquaporins are expressed in the mouse heart. Since H2O2 are generated from superoxide anion (O2 -) at extracellular space produced by NADPH oxidases [78,79], aquaporin-mediated transfer of H2O2 to cytosol is critical for induction of hypertrophy and other responses. Treatment of mice with a clinically approved aquaporin 1 inhibitor, Bacopaside, attenuated cardiac hypertrophy [80]. These results suggest that aquaporin is a promising target to treat hypertrophy.
However, aquaporins are expressed in whole body, and inhibition of water transfer activity of aquaporins by drug or antibody may cause side effects in other tissues. It has been reported that anti-aquaporin-4 antibody is main pathogen of human neuromyelitis optica spectrum disorders and level of the antibody in the plasma links to poor visual prognosis in human [81].

Immunotherapy
Immunological strategy to target cardiac fibrosis was recently reported [82]. Chimeric antigen receptor (CAR) T cells were engineered to specifically recognize myofibroblasts and induced ablation of myofibroblasts leading to improved cardiac function. CAR consists of antigen-recognizing regions of single chain Fv fragment, transmembrane domain, intracellular domains of T cell activation receptor (CD3ζ) and co-stimulation receptor (CD28). After binding of CAR to myofibroblast-specific protein, in this case mouse fibroblast activation protein, CAR T cells causes cytotoxic killing to decrease the number of myofibroblasts. CAR T cells are already used for cancer therapy and are reported to have beneficial effects on the treatment. Strategy to eliminate myofibroblasts is supported by the following experiments. When diphtheria toxin receptor specifically expressing in myofibroblasts binds diphtheria toxin, the toxin decreases the number of myofibroblasts and reduces cardiac fibrosis by myocardial infarction [83]. However, several concerns are raised before application of CAR T cells to heart failure patients who are seriously suffered from fibrosis [84]. CAR T cells release cytokines for attacking myofibroblasts and heart failure patients are in advanced inflammatory state. Released cytokines will complicate the conditions of heart failure. It may cause detrimental effects on heart failure patients. Authenticity of elimination of myofibroblasts is also questioned.
Myofibroblasts produce extracellular matrix such as collagen and fibronectin that are in part protective against cardiac rupture. Another point is the lack of information of myofibroblast-specific marker proteins in human. Specific antigen is essential for was successively used for the treatment of tumor growth and metastasis [85]. Dual inhibitor was more effective than treatment with TGF-β inhibitor alone. It is expected to increase the specificity of TGF-β inhibitor. Thus, to combine TGF-β inhibitor and antibody that recognizes myofibroblast-specific protein into single molecule may be another strategy to treat cardiac fibrosis. Identification of myofibroblast-specific antigen helps increase specificity of TGF-β inhibitor against cardiac fibrosis and restricts the action of TGF-β inhibitor at the localized area.

Conclusion
Myofibroblasts are the cells that mainly differentiate from fibroblasts and are responsible for production of extracellular matrix in injury. Relationship between fibroblasts and myofibroblasts is more complex than previously thought. Single cell RNA sequencing and lineage tracing techniques demonstrate heterogeneity of fibroblasts, although functional differences are remained to be determined. Treatment of cardiac fibrosis is eagerly waited, signaling analysis of various fibroblast states will provide a therapeutic target that is suitable for drug development. Although direct reprogramming of cardiac fibroblasts to cardiomyocytes with chemical compounds is not mentioned in this review, technologies of direct reprogramming are currently in progress [86] and may be another option for the treatment of cardiac fibrosis in future.