MicroRNA: role in macrophage polarization and the pathogenesis of the liver fibrosis.

In recent years, the involvement of miRNAs in liver disease has received extensive attention. A large number of studies have shown that the expression level of miRNAs in the serum and liver tissue of patients with liver fibrosis is significantly changed. MiRNAs are implicated in the liver fibrosis process by affecting the proliferation, apoptosis and activation of HSCs, immune cells and hepatocytes ( 109 ). EVs such as exosomes represent an important mode of intercellular communication, serving as cargo carriers between cell membranes and cytoplasmic proteins, lipids and RNA. MiRNAs can be packaged into exosomes and secreted from macrophages to affect the process of liver fibrosis. The macrophage-derived exosomal miRNAs regulate the activation and apoptosis of HSCs involved in the pathology of liver fibrosis are summarized in Figure 1 . It was reported that the microRNA Csi-let-7a-5p delivered by EVs from Clonorchis sinensis can promote the activation of M1-like macrophages and contribute to the biliary injuries and fibrosis by targeting the Socs1- and Clec7a - modulated NF-κB signaling pathway ( 110 ). Chen et al. used a mouse model of CCl4-induced liver fibrosis to demonstrate that the expression of exosomal miR-500 was upregulated in LPS-induced macrophages, and exosomal miR-500 overexpression could promote the proliferation and activation of HSCs and accelerate liver fibrosis by inhibiting mitochondrial fusion protein 2 (MFN2) ( 111 ). MiR-103-3p in exosomes secreted by LPS-treated THP-1 macrophages can promote the activation and proliferation of HSCs by targeting KLF4, and is involved in the crosstalk between macrophages and HSCs during the progression of liver fibrosis ( 112 ). In patients with NAFLD, myeloid-specific IL-6 signaling enhanced the release of miR-223-enriched exosomes from macrophages, which transferred antifibrotic miR-223 to hepatocytes to reduce the expression of pro-fibrotic transcriptional activator with PDZ-binding motifs (TAZ) in hepatocytes to inhibit liver fibrosis ( 113 ). During the development of NASH, miR-690 expression was significantly lower in mouse and human NASH livers compared to controls; the KC-specific KO of miR-690 increased NASH development, whereas miR-690 therapy restored specific KC functions by targeting NADK and led to reduced fibrosis and steatosis ( 45 ). Similarly, it was found that serum exosomes from NASH patients contained decreased levels of miRNA-411-5p. Further investigation revealed that exosomal miR-411-5p from M2 macrophages could inhibit the activation of HSCs. Additionally, miR-411-5p was found to directly downregulate the expression of Calmodulin-Regulated Spectrin-Associated Protein 1 ( CAMSAP1 ) to inactivate HSCs ( 114 ). CCl4-induced liver fibrosis model, exosomes derived from relaxin-treated macrophages exhibited a potent antifibrogenic effect, which was primarily attributed to miR-30a-5p ( 115 ). MiR-30a-5p suppressed the activity of the ASK1, which is known to be involved in the activation of HSCs. This in turn led to the restoration of PPAR-γ activity in the activated HSCs ( 115 ). Furthermore, restorative Ly6C lo macrophages showed a higher expression of miR-30a-5p compared to Ly6C hi macrophages, and miR-30a-5p synergized with relaxin gene therapy to achieve an enhanced antifibrosis effect ( 115 ). Macrophage-derived exosomal miRNAs regulate the activation and apoptosis of HSCs involved in the pathology of liver fibrosis. In addition to miRNAs in macrophage-derived exosomes that are involved in liver fibrosis, some miRNAs may mediate the pathology of liver fibrosis by regulating macrophage polarization (summarized in Figure 2 ). MiR-155 was reported to be involved in high fat-high cholesterol-high sugar (HF-HC-HS) diet-induced steatosis and liver fibrosis, as miR-155 knockout mice showed significantly less liver injury, decreased steatosis, and attenuation in fibrosis under HF-HC-HS diet or CCl4 treatment, and KCs isolated from miR-155 KO mice displayed the M2 phenotype when exposed to even M1 priming conditions ( 116 , 117 ). In addition, serum exosomal miR-155 levels in patients with hepatic fibrosis and a hepatic fibrosis rat model were positively correlated with the severity of liver fibrosis, and miR-155 could be used as a biomarker for the diagnosis and progression of liver fibrosis ( 118 ). In a murine model of arsenic-induced liver fibrosis, the level of miR-21 and Arg-1 were increased; however, miR-21 deficiency in mice showed attenuated liver fibrosis and M2 polarization compared with WT mice exposed to arsenite ( 119 ). MiR-20a-5p was downregulated during liver fibrosis in human and CCl4-induced mouse model samples. Moreover, miR-20a-5p downregulation in liver fibrosis led to the activation of TGF-β signaling pathway by targeting TGFBR2, accompanied by the activation of hepatic macrophages and the production of ECM by HSCs. The reintroduction of miR-20a-5p may be a therapeutic regimen for clinical intervention in hepatic fibrosis ( 9 ). MSC-EXOs have been demonstrated to exhibit a protective effect against liver fibrosis. In the CCl4-induced liver fibrosis mouse model, miR-148a enriched MSC-EXOs have been shown to regulate intrahepatic macrophage through KLF6/STAT3 signaling ( 48 ). MiR-148a showed the ability to suppress pro-inflammatory macrophages and promote anti-inflammatory macrophages, ultimately helping to reduce the severity of liver fibrosis ( 48 ). MiR-130a-3p is an antifibrotic miRNA with decreased expression in the serum of patients with cirrhosis and the liver of mice with schistosomiasis. Overexpression of miR-130a-3p by the lentivirus vector (LV-miR-130a-3p) could alleviate liver granulomatous inflammation and liver fibrosis; moreover, LV-miR-130a-3p promoted the polarization of macrophages towards the restorative Ly6C lo phenotype, inhibited the activation and proliferation of HSCs and also induced the apoptosis of HSCs by inhibiting MAPK1 expression ( 49 ). MiR-130a-3p also cooperated with miR-142-5p to control macrophage polarization. The transduction of miR-130a-3p mimics and miR-142-5p anti-sense oligonucleotides (ASO) in IL-4-treated mouse macrophages synergistically inhibited M2 polarization and their profibrogenic activities in both humans and mice, and miR-142-5p and miR-130a-3p mediated M2 macrophages by targeting SOCS1 and PPARγ, respectively ( 120 ). During the spontaneous resolution of liver inflammation (SRLI), neutrophil-derived miR-223 downregulated Nlrp3 expression in hepatic proinflammatory macrophages and induced their alternative activation into a restorative phenotype, which released IL-10 thus mitigating fibrogenesis by reducing the activation of HSCs and collagen formation ( 121 ). Similarly, in fibrotic NASH induced by long-term administration of a high-fat, fructose and cholesterol (FFC) diet, treatment with synthetic miR-223 analog miR-223-3p significantly alleviated the fibrosis development and activation of HSCs by disrupting the activation of the NLRP3 inflammasome ( 122 ). miRNAs modulate the macrophage polarization and participate in the liver fibrosis through different signaling pathway.

Liver fibrosis is an abnormal wound-healing response that develops in response to liver injury caused by various factors. The activation of hepatic stellate cells (HSCs) is recognized as a central event in liver fibrosis, in which activated HSCs transdifferentiate into myofibroblasts and secrete large amounts of extracellular matrix (ECM) that is deposited among the cells, leading to liver fibrosis ( 1 ). Liver fibrosis is among the common sequelae of chronic damage induced by toxic agents, viral infections, autoimmune diseases, metabolic and genetic diseases ( 2 ). Without effective intervention and treatment, it can progress into cirrhosis, hepatocellular carcinoma (HCC), liver failure, and concurrent infection leading to death ( 3 ). Although HSCs are major contributors to the pathogenesis of liver fibrosis, certain immune cells such as T and B lymphocytes, NK cells, and macrophages also play important roles ( 4 ). Among them, macrophages are the most abundant liver immune cells and are critical in the process of liver injury and subsequent liver fibrosis ( 5 ). MicroRNAs (miRNAs) are about 22-26 nucleotides long endogenous non-coding RNAs expressed in animals, plants and some viruses. They participate in post-transcriptional gene regulation through a combination of translational repression and mRNA destabilization ( 6 ). Some studies have shown that miRNAs can regulate the activation of HSCs and are involved in various types of chronic liver diseases, such as viral hepatitis, nonalcoholic fatty liver disease and autoimmune liver disease, and play an indispensable role in the occurrence and development of liver fibrosis ( 7 ). Furthermore, in the pathological process of liver fibrosis, miRNAs may serve as key regulators of macrophage polarization, where macrophages can differentiate into the M1 phenotype with pro-inflammatory and anti-infective functions or the M2 phenotype with pro-fibrogenic and tissue remodeling roles ( 8 ). In this review, we summarize the characteristics of hepatic macrophages and their roles in liver fibrosis. Importantly, we focus on how miRNAs regulate the polarization of macrophages, thus affecting the eventual progression of liver fibrosis. Our study aims to provide new therapeutic ideas for improving liver fibrosis based on miRNAs and macrophages.

Macrophages are an important component of innate immunity and act as the host’s first line of defense against external infection or internal damage ( 16 ). According to their origin, intrahepatic macrophages are mainly divided into two types: resident Kupffer cells (KCs) and monocyte-derived macrophages (MoMϕs). KCs originate from yolk sac-derived colony-stimulating factor 1 receptor (CSF1R) + erythroid progenitors (EMPs), and develop further from EMPs into fetal liver mononuclear cells, which give rise to KCs ( 17 ). Kupffer cells, as the liver-resident macrophages, are located only in the intravascular compartment and are mainly located in the hepatic sinusoids. KCs function to remove cellular debris and metabolic waste ( 18 , 19 ), maintain liver homeostasis, promote tissue repair and regeneration, and initiate the innate and adaptive immune responses ( 20 ). During homeostasis, KC replenishment is independent of BM-derived progenitors, and occurs predominantly by the self-renewal of resident stem cells ( 21 , 22 ). Various pattern recognition receptors (PRRs) are highly expressed on the surface of KCs including Toll-like receptors (TLRs) and nucleotide binding oligomerization domain-like receptors (NLRs), which leads to the rapid response of KCs to various stimuli and activation signals during liver injury ( 23 ). The main stimuli recognized by KCs include reactive oxygen species (ROS); damage-associated molecular patterns (DAMPs) such as high mobility group box protein 1 (HMGB1), mitochondrial DNA and ATP; pathogen-associated molecular patterns (PAMPs) such as lipopolysaccharide (LPS), lipoteichoic acid (LTA) and β-glucan ( 24 ); hypoxia inducible factor 1α (HIF-1α); multiple metabolites; cell extracellular vesicles and microRNAs ( 25 ). KCs and MoMϕs in the liver can be distinguished from each other by their cell surface markers; however, no single marker is available to discriminate these populations. In mouse models, the main surface markers of KCs are CD11b low , F4/80 high , Clec4F + and CX3CR1 − ( 5 , 26 ). The surface markers of MoMϕs in mice are CD11b + , F4/80 int , Ly6C + , and CX3CR1 hi ( 5 ). MoMϕs develop from lineage-negative (LIN − ) hematopoietic stem cells in the bone marrow, can be mainly found at the portal triad in the healthy liver, and function to maintain the iron and cholesterol homeostasis ( 27 ). Under pathological conditions, KCs secrete cytokines and chemokines, including TNF-α, IL-1β and CCL2, to recruit circulating monocytes migrating and infiltrating into the liver ( 28 ). The liver-infiltrating monocytes then differentiate into MoMϕs. MoMϕs in the murine liver can be further divided into two subgroups according to the expression level of Ly6C: Ly6C hi and Ly6C lo monocyte/macrophages ( 25 ). CD11b hi F4/80 int Ly6C hi macrophages (Ly6C hi macrophages in short) are derived from recruited CCR2 + CX3CR1 lo Ly6C hi monocytes and exert proinflammatory and profibrotic functions, while CD11b hi F4/80 hi Ly6C lo macrophages (Ly6C lo macrophages in short) are converted from Ly6C hi macrophages induced by phagocytosis and are involved in anti-inflammatory and antifibrotic processes ( 29 , 30 ). It should be noted that the Ly6C hi and Ly6C lo phenotypes comprise a new system for macrophage classification based on cell origin and surface makers. Conventionally, macrophages with different functions are classified as M1 and M2 macrophage subsets. M1 macrophages are known as classically activated macrophages with pro-inflammatory properties and participate in tissue damage and inflammation, whereas M2 macrophages are known as alternatively activated macrophages with anti-inflammatory properties and function to promote tissue repair and regeneration. M1 macrophages are mainly stimulated by IFN-γ or LPS, characterized by high expression of CD80, CD86, major histocompatibility complex II (MHC II), Toll-like receptor 4 (TLR4), and inducible nitric oxide synthase (iNOS) ( 31 ). Meanwhile, M2 macrophages can be stimulated by T helper 2 (Th2) cytokines such as interleukin 4 (IL-4) and IL-13 ( 32 ), with high expression of mannose receptor 1 (MRC1/CD206), CD163, arginase-1 (Arg1), chitinase 3-like 3 (Chil3/Ym1), found in inflammatory zone 1 (FIZZ1) ( 33 ). Among them, Chil3 and Fizz1 are the markers only expressed by M2 macrophages in mouse. In addition, M2 macrophages can be further subdivided into M2a, M2b, M2c, and M2d subtypes by distinct stimuli. M2a is induced by IL-4 and IL-13, M2b is induced by immune complex (IC), the M2c type is stimulated by IL-10, transforming growth factor-β (TGF-β) and glucocorticoids, and the M2d type is activated by IL-6, TLR ligands and adenosine ( 34 ). Macrophages can be polarized into different subsets in response to different local microenvironments and play essential roles in the initiation, progression and resolution of tissue inflammation and injury in various liver diseases ( 35 ). In hepatic fibrosis, the activated macrophages secrete pro-inflammatory cytokines and chemokines and stimulate HSCs to transdifferentiate into myofibroblasts, which proliferate and produce ECM proteins ( 36 ). Although the activation of HSCs is thought to be a central driver of hepatic fibrogenesis ( 37 , 38 ), hepatic macrophages have emerged as essential in the pathogenesis of liver fibrosis. Moreover, due to their heterogeneity and plasticity, macrophages can exert both pro- or anti-fibrotic effects by regulating the activation or the cell death of HSCs and the formation and degradation of matrix collagen ( 39 , 40 ). In human and mouse models of diet-induced nonalcoholic steatohepatitis (NASH), the impaired macrophage-mediated clearance of necroptotic hepatocytes (necHCs) and increased activation of HSCs are responsible for liver fibrogenesis; hence, the reduced accumulation of necHCs in NASH liver could be a therapeutic strategy to treat hepatic fibrosis ( 41 ). Cai et al. further reported that c-mer tyrosine kinase (MerTK) signaling in macrophages activates HSCs to promote collagen synthesis and induces liver fibrosis through the ERK-TGFβ1 pathway ( 40 ). In bile duct ligation (BDL)-induced and carbon tetrachloride (CCl4)-induced liver fibrosis mouse models, the FGF12-mediated proinflammatory activation of hepatic macrophages could induce HSC activation mainly through the monocyte chemoattractant protein-1/chemokine (C-C motif) receptor 2 axis ( 42 ). The roles of MoMϕs in liver fibrosis were also investigated. For instance, the proportion of resident macrophages decreases during the process of inflammation and fibrogenesis, while that of the recruited MoMϕs (CD11b high F4/80 mid subsets) gradually increases during fibrogenesis ( 9 ), suggesting an important function of MoMϕs in liver fibrosis. De Souza et al. further demonstrated that the transplantation of bone marrow-derived CD11b + CD14 + monocytes caused the significant improvement of liver fibrosis by inhibiting oxidative stress and inflammation in a murine model of CCl4-induced chronic liver damage ( 43 ). In addition, liver fibrosis was attenuated by the transplantation of bone marrow-derived MSCs (BM-MSCs), and the therapeutic effect of BM-MSCs was attributed to promoting the Ly6C hi /Ly6C lo subset conversion and Ly6C lo macrophage restoration through activating the antifibrogenic cytokine and apoptotic pathways ( 44 ). Similarly, prepolarized BMDMs also exhibit a therapeutic effect on liver fibrosis. For example, M1 BMDMs significantly ameliorated liver fibrosis by modulating the hepatic microenvironment to recruit endogenous macrophages into fibrotic liver, which showed the phenotype of Ly6C lo restorative macrophages ( 39 ). Compared with Ly6C lo macrophages, Ly6C hi macrophages exerted a pro-fibrogenic effect by activating HSCs through secreting various cytokines including TGF-β, platelet-derived growth factor (PDGF), TNF-α, IL-1β, monocyte chemotactic protein 1 (MCP1), CCL3, and CCL5 ( 36 ). Taken together, hepatic macrophages play an important role in the initiation and progression of liver fibrosis. During this process, however, the function, metabolism and polarization of macrophages are regulated by various factors such as miRNAs, which ultimately affect the onset of liver disease. For instance, exosomal miR-690 derived from KCs inhibited inflammation in recruited hepatic macrophages in a mouse model of NASH ( 45 ). MiR-206 drove KCs toward M1 polarization, and promoted the recruitment of CD8+ T cells in HCC ( 46 ). In addition, miR-26a overexpression extensively inhibited the inflammation in both hepatocytes and KCs therefore attenuated HCC ( 47 ). MiR-155 knockdown in KCs could positively regulate the immunosuppressive function of KCs and prolong the survival of liver allografts. MiR-148a-enriched mesenchymal stem cell-derived exosomes (MSC-EXOs) modulated macrophages towards the anti-inflammatory phenotype and exerted ameliorative effects on liver fibrosis ( 48 ). In a mouse model of Schistosomiasis japonicum , miR-130a-3p promoted the differentiation of macrophages toward the Ly6Clo phenotype and alleviated liver granulomatous inflammation ( 49 ). The above studies demonstrate the diverse roles of miRNAs in hepatic macrophages, influencing the pathology of liver diseases. The regulatory effect of miRNAs on macrophage polarization in other models and tissues will be discussed in more detail below.

MicroRNAs (miRNAs) are endogenous, small non-coding RNA molecules widely expressed in all types of human cells. They predominantly function to negatively regulate gene expression at the post-transcriptional level and play important roles in various biological functions, such as immune response, cell proliferation and apoptosis ( 9 – 11 ). MicroRNAs are first transcribed in the nucleus by RNA polymerase II to generate primary miRNAs (pri-miRNAs), which are then cleaved by RNase III enzyme Drosha to generate precursor miRNAs (pre-miRNAs). These are translocated from the nucleus to the cytoplasm and then further processed by Dicer to produce double-stranded miRNAs containing mature miRNAs ( 11 , 12 ). Mature miRNAs are directed to the 3’ end of the untranslated region (UTR) of their specific target mRNAs by base-pairing, which represses protein expression by destabilizing the mRNA and translational silencing ( 10 , 13 ). However, in some cases, miRNAs can also upregulate gene expression by activating the translation of target mRNAs. Generally, a single miRNA can regulate multiple mRNAs simultaneously, and one mRNA can also be regulated by several miRNAs ( 13 ). MiRNA dysregulation has been implicated in the pathogenesis of a variety of human diseases, including cancer, cardiovascular disease, metabolic disease, diabetes, and virus-induced diseases ( 14 ). Due to their stable presence in body fluids such as blood, urine and saliva, miRNAs might be promising biomarkers for the early diagnosis and potential therapeutic targets of some diseases ( 15 ).

Searching the PubMed database using miRNAs and liver fibrosis as keywords yielded more than 1600 publications, while this number was nearly 3600 when using ‘macrophages’ and ‘liver fibrosis’. Therefore, both miRNAs and macrophages are research hotpots in the field of liver fibrosis. The pathogenesis of liver fibrosis is considered to be a complex, multifactorial process. For instance, activated HSCs are a major contributor to liver fibrosis because they produce excessive amounts of ECM as a result of long-term liver injury. In addition to HSCs, macrophages are also considered a ‘double-edged sword’ in the development of fibrosis. Hepatic macrophages are composed of several heterogeneous subpopulations, which can be classified as ‘pro-inflammatory’ M1 or ‘immunoregulatory’ M2 macrophages according to their function and phenotype. Given that miRNAs epigenetically fine-tune the expression of hundreds of target mRNA, there is growing interest in the regulatory role of miRNAs in macrophage activation, polarization, tissue infiltration, and the mitigation of inflammation. MiRNAs play different roles in the pathogenesis of multiple diseases; They have the potential as promising biomarkers and therapeutic targets in the treatment of various illnesses. However, the same miRNA may play different or even opposite roles in different pathological processes. For example, miR-192-5p-enriched hepatocyte exosomes promoted M1 phenotype polarization in NAFLD ( 51 ); however, miR-192-5p suppressed M1 macrophage polarization in a MSU crystal-induced mouse GA model ( 52 ). The exosomal miR-30d-5p of TNF-α-stimulated neutrophils promoted M1 macrophage polarization in a mouse model of sepsis-related acute lung injury ( 58 ), whereas miR-30d-5p-enriched exosome from the adipose-derived stem cell suppressed M1 microglial polarization in acute ischemic stroke-induced brain injury ( 59 ). These disease-specific functions of miR-192-5p and miR-30d-5p on macrophage polarization may be attributed to the difference in the origin of miRNA-enriched exosome and the disease microenvironment. Although numerous studies have shown that both miRNAs and macrophages are involved in the pathogenesis of liver diseases, the regulatory role of miRNAs in macrophage polarization has also been the focus of research. However, the mechanism of how miRNAs mediate the activation and polarization of macrophages and thus affect the progression of liver fibrosis remains unclear. Some miRNAs (i.e., miR-155, miR-21, miR-20a-5p, miR-148a, miR-130a-3p, and miR-223) can regulate macrophage polarization in liver fibrosis, while relevant studies are mainly limited to animal experiments, so further research is needed to test whether these miRNAs can be applied in clinical liver fibrosis-associated diseases. Due to the dual complexity of macrophage polarization and the pathogenesis of liver fibrosis, it is not feasible to study the pathology of liver fibrosis only based on miRNA or macrophages. A more comprehensive understanding of the cell-specific functions of miRNAs in liver fibrosis through the modulation of macrophage polarization is necessary, which can help identify novel diagnostic targets and design feasible miRNA-based therapies for liver fibrosis.

BZ and QG organized the article. WY and SW wrote the draft. YW, HC, HN, LL and XZ participated in conception and discussion of the article. XZ and BZ supervised the manuscript writing and edited the language, figure and table. WY and SW contributed equally to write the manuscript. All authors contributed to the article and approved the submitted version.