Epigenetics as a versatile regulator of fibrosis

HAT consists of three major families: general control nonderepressible 5 (Gcn5)-related N-acetyltransferases (GNATs), p300/CBP, and MYST proteins [50‐52], among which p300/CBP is most important in fibrosis; its mechanism has been elaborated, and corresponding inhibitors have been discovered.

The role of p300 in fibrosis has been verified in multiple studies (Fig. 2A, Table 3). P300, as a histone acetylase, can induce histone acetylation of MCP-1 [53], NOX4 [54], and other gene promoters to promote the fibrogenesis process in IPF. In the process of p300 regulation, in addition to the cis mechanism (charge effects), a type of protein plays a vital role as a “reader” in the trans mechanism [55]. Bromodomain-containing protein 4 (Brd4) is a member of the bromodomain and extraterminal (BET) family of proteins, which function as epigenetic “readers” of acetylated lysine groups on histones. The vital roles of Brd4 and its inhibitor JQ1 have been proven in the epigenetic regulation of p300 in various profibrotic genes, such as NOX4, snail family transcriptional repressor 1 (SNAI1), and CTGF in IPF and myocardial infarction (MI) induced cardiac fibrosis [54, 56‐59]. P300 functions in the TGF-β signaling pathway as a coactivator with Smad3. First, TGF-β regulates p300 expression by Early Growth Response 1 (EGR1) [60, 61] and can regulate the translocation of p300 through posttranslational modification; for example, the phosphorylation of p300 by AKT signaling has been reported to induce its translocation to the nucleus in liver fibrosis [62]. Nucleic p300 then increases the synthesis of collagens by interacting with TGF-β-activated Smad3 on the collagen gene promoter.

Previous study has identified strategies that work against this profibrotic effect, a small molecule inhibitor, L002, has been found to mediate the suppression of the acetylase activity of p300 in fibroblasts, resulting in the repression of TGF-β-induced H3K9 acetylation, thus inhibiting myofibroblast differentiation and collagen synthesis in hypertrophic nephropathy [63]. p300 may interact with other epigenetic regulators. Sirtuins and other microRNAs, such as miR-200b, miR-132, and miR-133a, have been identified in fibrosis by regulating the expression of p300 [64‐69].

The acetyl of histone could be removed by deacetylation through a series of histone deacetylases, triggering a compact nucleosome structure and preventing active transcription. HDAC can be classified into four distinct groups based on its function, DNA sequence, and domain organization. Class I HDACs include HDAC1, HDAC2, HDAC3, and HDAC8, which are widely expressed and found mainly in the nucleus. Class IIa HDACs include HDAC4, 5, 7, and 9, while class IIb HDACs include HDAC6 and 10. These two classes are subdivided based on the number of catalytic domains the proteins possess. Class III HDACs include sirtuins (SIRTs) and nicotinamide adenosine-dependent (NAD) enzymes. Class IV HDACs contain only one member, HDAC11, which shares sequence domains with class I and class III HDACs. HDACs epigenetically alter the gene transcription process via the deacetylation of core histones. They increase the positive charges on histones and possibly strengthen histone-DNA interactions and repress transcription. However, whether HDACs can directly activate transcription and the exact detailed mechanisms by which they regulate transcription remain to be clarified [70].

HDACs can regulate fibrosis via fibroblast proliferation, senescence, and ECM production [71‐74]. Several signaling pathways participate in fibrosis, including the TGF-β pathway, the Wnt pathway, and apoptosis signaling pathways [75]. In the following chapter, we will discuss the function of HDACs in regulating fibrosis in all these processes (Fig. 2B, Table 3).

HDAC1 and HDAC2 coexist to form Sin3, NuRD, and CoREST complexes [76]. HDAC3 can also interact with SMRT/NCoR to stimulate the enzymatic activity of HDAC3 [77]. These complexes have been found to be significant in fibrogenesis through the regulation of COX-2 and IP-10. The binding of the CoREST and mSin3a transcriptional corepressor complexes, as well as the NCoR complex with the COX-2 promoter, is markedly strengthened, resulting in the insufficient acetylation of histone H3 and H4 and weakening the binding of the transcription factors NF-κB, C/EBPβ, and CREB-1 to the COX-2 promoter, eventually leading to diminished COX-2 transcription in IPF [78]. The epigenetic regulation of IP-10 is almost the same [79]. COL1A1 and SMAD7 are also inhibited via the recruitment of repressor complexes comprising SP1, SIN3A, CoREST, LSD1, and HDAC1 to the promoter in systematic sclerosis (SSc) [80]. In addition, HDACs have been found to be recruited by transcription factors to the promoter. For example, HDAC3 is recruited by activating transcription factor 3 (AFT3) to the Wnt inhibitor factor 1 (WIF-1) promoter and inhibits WIF-1 expression in SSc, which induces COL1A1 expression [81]. In contrast to the findings above, some class I HDACs function independently in fibrotic pathways. HDAC8 inhibition at least partially represses TGF-β-induced fibrosis by increasing PPARγ gene transcription via restoration of H3K27 acetylation at the enhanced region and finally regulates CTGF, plasminogen activator inhibitor type 1 (PAI-1), and α-smooth muscle actin (α-SMA) expression in IPF [82]. However, some class I HDACs were also found to play an antifibrotic role in fibrosis. HDAC1 was recruited to the mitogen-activated protein kinase kinase 3 (MAP2K3) promoter by AFT3, resulting in MAP2K3 gene-associated histone deacetylation, thereby inhibiting MAP2K3 expression. MAPK signaling is then activated to inhibit profibrotic effects in SSc [83].

Recruitment of HDAC4, HDAC5, and HDAC7 promotes the deacetylation of H3 and H4 histones. They can enhance fibroblast proliferation by inhibiting thy-1 and promoting fibroblast activation through downregulation of peroxisome proliferator-activated receptorγ coactivator-1 (PGC-1) in IPF [84‐86]. They can also inhibit fibroblast apoptosis by downregulating Fas signaling [87] and ECM degradation by inhibiting the expression of MMP9 in liver fibrosis [88]. Nucleocytoplasmic shuttling has been found in fibrosis. Nucleic HDAC7 can bind to the promoter of hepatocyte growth factor (HGF), which leads to the repression of HGF and induces liver fibrosis. Cylindromatosis (CYLD) was discovered to stimulate the export of HDAC7 to cytoplasm independently of the classic mechanism to ameliorate organ fibrosis [89]. HDAC7 has also been shown to regulate collagen and other ECM in systemic sclerosis fibroblasts and siRNA mediated depletion of HDAC7 reduced ECM in these cells [90]. Interestingly, HDAC4 was discovered to remove acetylated histones H3 and H4 from the SIRT1 promoter, and SIRT1 can thus deacetylate PPARγ to block fibroblast activation [91].

Class III HDAC and SIRT has been found to involve in nonhistone protein modifications in fibrosis and only a few studies have revealed its role in histone modifications [92]. SIRT6 can not only inhibit insulin-like growth factor (IGF) and resist fibroblast apoptosis through the Akt pathway but can also regulate the expression of the transcription factor FOS, regulating the transcription of fibrosis-related genes upstream [93, 94]. SIRT1 was found to induce ECM degradation by deacetylating histones on the MMP9 promoter, thereby suppressing its transcription in chronic obstructive pulmonary disease (COPD) [95].

The studies described above have illustrated the multifunctional and diverse mechanisms of HDACs by targeting various genes in the processes of fibroblast proliferation, apoptosis, ECM deposition, and degradation. Notably, HDACs act as fundamental regulators participating in epigenetic modifications, thus making them a generally recognized potential treatment target for fibrosis.

Compared to histone acetylation, methylation has been characterized as a more complex entity since distinct histone lysine residues may have various functions when methylated. Additionally, different degrees of methylation on the same residue may vastly differ in their functions [96, 97]. Unlike acetylation, methylation does not change the histone charge, nor does it directly affect the histone-DNA interaction. Rather, methylation regulates gene expression through gene transcription, chromatin remodeling, and other epigenetic modifications [98, 99].

Enhancer of Zeste Homolog 2 (EZH2) is a histone methylase, together with embryonic ectoderm development (EED) and suppressor of zeste 12 (SUZ12), forming polycomb repressive complex 2 (PRC2) to mediate H3K27me3, which is essential for stable silencing (Fig. 3A, Table 3) [100]. EZH2 has been mostly found to regulate fibroblast proliferation, cell transdifferentiation, and ECM production and was found to be involved in various signaling pathways. Increased EZH2 caused downregulation of Smad7 and phosphatase and tensin homolog deleted on chromosome 10 (PTEN). As a result, the TGF-β/Smad3 signaling, EGFR, and PDGFR signaling pathways are activated, leading to the activation of fibroblasts and ECM deposition in diabetic nephropathy [101]. Epigenetic silencing of Dkk1 by EZH2 is a critical mechanism mediating HSC activation and fibrogenesis in liver fibrosis [102]. EZH2 has also been proven to mediate the transcriptional repression of the antifibrotic gene PPARγ [103]. Although in SSc. EZH2 inhibition actually enhanced fibrosis [104].

The demethylation of H3K27 was associated with Jumonji domain-containing protein D3 (JMJD3, encoded by KDM6B) and ubiquitously transcribed tetratricopeptide repeat on chromosome X (UTX, encoded by KDM6A). In recent studies, repression of EZH2 and activation of JMJD3 were demonstrated to work together in TGF-β-induced fibrosis, repressing H3K27me3 and thus increasing profibrotic gene expression [105]. The direct interaction of JMJD3 with promoter regions of X-linked inhibitor of apoptosis protein (XIAP) and survivin, which are members of the inhibitor of apoptosis protein (IAP) family, attends the resistance of fibroblasts to apoptosis [106]. Similar to EZH2, the paradoxical role of JMJD3 was discovered in skin fibrosis. Inhibition of JMJD3 ameliorated bleomycin-induced and topoI-induced fibrosis in well-tolerated doses, mechanically, inactivation of JMJD3 reduced the expression of fos-related antigen-2 (fra-2), a member of the AP1 family of transcription factors that has previously been shown to play a central role in the pathogenesis of SSc [107].

Significantly, different methylases can work together or interfere with other epigenetic factors, establishing a crosstalk network. A set of studies found that pulmonary fibrosis was regulated by G9a/EZH2-mediated H3K9me3/H3K27me3, interacting with DNA methylation in a bidirectional and mutually dependent manner to reinforce COX-2 and CXCL10 epigenetic silencing [108, 109]. DNA methylation has also been found to recruit histone methylase. It has been reported that DNA methylation plays a role in H3K9me since DNMT3a can recruit the histone methylase SUV39H1 using its PHD-like motif [110]. MeCP2 also recruits EZH2 to induce H3K27me and ultimately induce the transcriptional repression of PPARγ in organ fibrosis [103].

In addition to EZH2, another methylase, G9a, catalyzes the methylation of H3K9, and this modification serves as a binding site for chromodomain protein heterochromatin protein 1 (HP1), thus generating local heterochromatin (Fig. 3B, Table 3) [111]. Either G9a or HP1 can inhibit the expression of gene PPARGC1A and promote fibrosis. PGC1a has been proved to inhibit fibrosis in multiple animal models, and is possibly related to mitochondrial metabolism [112]. A study found that G9a-induced H3K9me1 had a pivotal role in reducing Klotho expression, and Klotho appeared to be the primary mediator of antifibrotic effects through inhibition of the TGF-β1, Wnt, and other fibrosis-related signaling pathways [113, 114]. H3K9me3 is responsible for the decreased expression of the death receptor Fas and resistance to Fas-mediated apoptosis in fibroblasts [87]. Moreover, in radiation-induced pulmonary fibrosis, enrichment of H3K9me2/3 has been found in E-cadherin promoter in epithelial cells, and its positive regulation of EMT can be inhibited by the G9a inhibitor BIX01294 [115, 116].

Existing studies have already identified some H3K9 demethylases, such as JMJD1A/KDM3A, JMJD2A/KDM4A, JMJD2B/KDM4B, JMJD2C/KDM4C, JMJD2D/KDM4D, and the PHD finger proteins 2 and 8 (PHF2 and PHF8) [99, 117]. JMJD1A may regulate the expression of Yes-associated protein 1 (YAP1) and TGF-β2 to increase ECM proteins [118]. Furthermore, JMJD1A reduced H3K9me2 on the CTGF promoter, thereby activating CTGF transcription and promoting myofibroblast activation [119]. Meanwhile, it can also maintain the homeostatic balance of the ECM by binding to the tissue inhibitor matrix metalloproteinase 1 (TIMP1) promoter [120]. KDM4D, another H3K9 demethylase, can inhibit H3K9me2 and H3K9me3, thereby activating the TLR4/MyD88/NF-kB signaling pathway to activate fibroblasts [121]. However, JMJD1A was also discovered to inhibit fibrosis by increasing PPARγ expression [122]. In another study, the sequence-specific transcription factor SREBP2 interacted with KDM4A, B, and C to activate miR-29 transcription, which also plays an anti-fibrotic role [123].

MLL family proteins usually act in the histone methyltransferase complex COMPASS, which consists of ASH2, RBBP5, WDR5, and hDPY30 (Fig. 3C, Table 3) [124, 125]. A study found that the components of COMPASS, including WDR5, ASH2, and MLL1, were recruited to the promoters of fibrogenic genes to activate the transcription of collagens in both diabetic nephropathy and CCl4 induced hepatic fibrosis [126, 127]. ASH1 and SET7/9 are two histone methyltransferases that have been found to bind the regulatory regions of ECM genes, TIMP1, CTGF, and TGF-β1, with increased levels of H3K4me1, H3K4me2, and H3K4me3 [128‐130].