Role of extracellular vesicles in liver diseases and their therapeutic potential

Abstract

More than eight hundred million people worldwide have chronic liver disease, with two million deaths per year. Recurring liver injury results in fibrogenesis, progressing towards cirrhosis, for which there doesn’t exists any cure except liver transplantation. Better understanding of the mechanisms leading to cirrhosis and its complications is needed to develop effective therapies. Extracellular vesicles (EVs) are released by cells and are important for cell-to-cell communication. EVs have been reported to be involved in homeostasis maintenance, as well as in liver diseases. In this review, we present current knowledge on the role of EVs in non-alcoholic fatty liver disease and non-alcoholic steatohepatitis, alcohol-associated liver disease, chronic viral hepatitis, primary liver cancers, acute liver injury and liver regeneration. Moreover, therapeutic strategies involving EVs as targets or as tools to treat liver diseases are summarized.

1. Introduction

More than eight hundred million people worldwide have chronic liver disease, with two million deaths per year [1,2]. Chronic liver injury due to alcohol consumption, diet or viral infection results in fibrogenesis progression leading to liver cirrhosis and its complications including ascites, gastro-intestinal bleeding and liver cancer. In addition to chronic liver diseases, acute liver failure, which can be caused by viral infection or drugs, is also a life-threatening condition with high mortality and resource cost [4]. Besides liver transplantation, effective therapies against chronic and acute liver injuries are lacking [3,4].

Extracellular vesicles (EVs) are nano-sized particles delimited by a lipid bilayer, naturally released from presumably all types of cells in the extracellular space. EVs cannot replicate, are continuously released by cells in healthy and pathological conditions, and can be found in all biological fluids [5]. EVs from a living cell include microvesicles and exosomes. Due to a lack of specific markers, it remains difficult to assign isolated EVs to a particular biogenesis pathway. Therefore, the most accepted classification of EVs currently is based on their physical properties such as size, i.e. large versus small EVs [5,6]. Large EVs range from 100 to 1000 nm, with the majority of them not exceeding 400 nm, and are mostly enriched with microvesicle markers such as ADP ribosylation factor 6 (Arf6) and Annexin V [7]. Small EVs range from 40 to 200 nm, and are enriched with exosome markers including cluster of differentiation 63 (CD63), CD81 and CD9 [8]. EVs composition varies according to pathophysiological state of the cell of origin [9,10]. In liver diseases, [9,10] the release and the cargo of EVs change in response to liver injuries, such as non-alcoholic steatohepatitis, alcohol-associated liver disease, viral hepatitis, drug-induced liver injury, ischemia–reperfusion injury and hepatobiliary malignancies [6,10–14]. Circulating EV concentrations are thus attractive biomarkers in liver diseases [6]. But, EVs are also significantly involved in cell-to-cell communication through delivery of diverse cargo such as protein, micro RNAs and DNA, as demonstrated using ex vivo and in vitro experiments which will be discussed in this review. Most of the in vivo studies have shown an increase of circulating EVs following liver injury without identifying the cell of origin, certainly due to a lack of cell of origin specific markers or to low resolution technologies. However, with recent technological advances in the field of intravital imaging [15] and reporter mouse models [16], the study of circulating liver-specific EVs is emerging. Moreover, there is an increasing interest in considering the therapeutic potential of EVs, which arises as a potential tool to meet patient’s needs.

In this review, we will discuss the role of EVs in the pathogenesis of different types of liver injury including acute and chronic injury, in the development and complications of liver cirrhosis, and will end by current therapeutic studies utilizing EVs to fight liver diseases.

2. EVs and causes of liver diseases

2.1. Non-alcoholic fatty liver disease and non-alcoholic steatohepatitis

Accumulation of fat in the liver not due to alcohol consumption is present in 25% of the world population and is referred to as nonalcoholic fatty liver disease (NAFLD) [17,18]. The etiopathogenesis has been assigned to the composition of nutrients (e.g. high fructose corn syrup), obesity, insulin resistance, alterations of the microbiome, genetic predispositions and changes in bile acid pool [19–21]. Nearly 30% of patients with NAFLD develop non-alcoholic steatohepatitis (NASH) characterized by hepatocellular injury, hepatic inflammation with or without liver fibrosis and are at risk for developing end-stage liver disease [17,18,22]. Accumulation of lipid intermediates causes hepatocellular lipotoxicity, leading to cellular stress, dysfunction and apoptosis [23]. Toxic lipids initiate signaling processes to recruit monocytes into the liver with differentiation and polarization of these monocytes into inflammatory macrophages [23].

EVs have recently been emphasized to play important roles in the pathogenesis and progression of NAFLD. Circulating EV number is increased in animal models and patients with NASH [6,24,25]. EVs bear inflammatory cargo which leads to the progression of NAFLD. Indeed, transfer of EVs isolated from the serum of high-fat diet-fed mice into chow diet-fed mice results in immature myeloid cell activation and homing to the liver, as well as increased levels of pro-inflammatory markers in the liver and serum aminotransferases [26]. In vitro treatment of hepatocytes with toxic lipids such as palmitate and its active metabolite lysophosphatidylcholine (LPC) evokes an increased EV release, which is mediated by the stress kinase mixed lineage kinase 3 (MLK3) [27]. MLK3 also regulates the cargo of lysophosphatidylcholine (LPC)-mediated EVs as it drives the enrichment of hepatocyte-derived EVs with C-X-C motif chemokine ligand 10 (CXCL10), inducing monocyte/macrophage accumulation in the liver [27]. This monocyte/macrophage accumulation is mediated by an integrin β1-dependent mechanism [28]. Consequently, Mlk3 knockout mice fed a NASH-inducing diet contain a lower number of EVs in the plasma and decreased CXCL10 expression in plasma and circulating EVs [27,29].

EVs derived from hepatocytes treated with non-esterified fatty acids bear other pro-inflammatory molecules, such as sphingosine-1-phosphate (S1P), ceramides, mitochondrial DNA (mtDNA), micro-RNAs and tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) [10]. The release of EVs from LPC-treated hepatocytes is dependent on TRAIL receptor 2 (TRAIL-R2) signaling cascade involving downstream pro-apoptotic caspase 3, which subsequently induces proteolytic activation of Rho-associated kinase 1 (ROCK1). ROCK1 in turn mediates the release of large EVs from the plasma membrane [12]. Furthermore, TRAIL knockout mice show reduced hepatic steatosis in a NASH mouse model [30], suggesting a need for developing TRAIL inhibitors (see Fig. 1).

Fig. 1. EVs in progression of nonalcoholic fatty liver disease (NAFLD).

Various lipotoxic stimuli enhance steatotic hepatocyte EV release. Released EVs carry biological cargo composed of proteins, DNA and lipids favoring infiltration of immune cells in the liver, underlying progression of liver inflammation. Other organs involved in NAFLD progression release EVs in the blood. They can either be taken up by the liver and contribute to development of liver inflammation and tissue remodeling or participate in systemic insulin resistance progression, which is a major driver of NAFLD progression. Prepared with Biorender.com.

It has been demonstrated that endoplasmic reticulum stress is involved in hepatic inflammation in steatohepatitis mouse model [16]. Palmitate-treated hepatocytes release lipotoxic and proinflammatory EVs enriched with ceramides and S1P through the unfolded protein response sensor inositol-requiring protein 1α (IRE1α) pathway [24]. Interestingly, concentrations of C16:0 ceramide are increased in plasma EVs from mice and patients with NASH [31]. Moreover, S1P-enriched lipotoxic EVs mediate macrophage accumulation in the liver in a mouse model of NASH promoting macrophage-dependent inflammation [31]. Proinflammatory macrophages can be bone-marrow derived, and thus recruited to the liver from the blood stream, or liver resident called Kupffer cells (KCs). The pro-inflammatory phenotype of KCs can be induced by fat-laden hepatocyte-derived EVs in a hypoxia-dependent manner [32]. As most of the results have been obtained using cell culture experiments, it would be interesting to examine whether toxic lipids and hypoxia also induce EV release and contribute to NASH in mouse models, although still technically challenging.

Mitochondrial DNA (mtDNA) is a newly recognized cargo in lipotoxic EVs in the context of NASH. MtDNA in hepatocytederived circulating EVs is increased in both mice and patients with NASH [33]. Moreover, mtDNA-enriched EVs activate TLR9 in KCs, triggering the secretion of pro-inflammatory cytokines, such as tumor necrosis factor α (TNFα) and interleukin 1β (IL-1β), and further increase liver damage [33]. In a recent study, the inflammatory profile of hepatocyte-derived mtDNA-enriched EVs was attenuated by the administration of IL-22 [34], suggesting a strong therapeutic potential for IL-22. Understanding the mechanism of how hepatocyte-derived EVs are enriched with mtDNA might be an interesting therapeutic target.

MiRNAs in liver diseases have been the focus of several studies and circulating miRNAs can be used as biomarkers of chronic liver disease [6,9]. MiR-122, a liver-specific miRNA, is altered in liver diseases of various causes such as viral hepatitis, alcohol-associated liver disease or NASH [9]. It has been reported that in a NASH mouse model, miR-122 shifts from argonaute-2 (Ago2)-bound form to being encapsulated into circulating EVs [35]. However, the degree to which the global circulating pool of miRNAs reflects the contribution of liver or other tissues remains unclear [36]. In this regard, the adipocyte-specific knockout of Dicer – a crucial enzyme in miRNA maturation and sorting – reduced by 80% the circulating EV-contained miRNAs without any decrease of the circulating EV concentration [37]. This suggests that adipose tissue is the major source of circulating miRNAs. The contribution of other highly metabolic organs, such as liver and skeletal muscle, in circulating miRNA pool in the context of NASH remains to be explored. While most of the studies report a biomarker potential of circulating miRNAs, a recent study shows that lipotoxic hepatocytes release EVs enriched in miR-192-5p that are proinflammatory through modulating RPTOR independent companion of mammalian target of rapamycin Complex 2 (Rictor)/Akt/ Forkhead box protein O1 (FoxO1) signaling pathway [38]. Moreover, steatotic hepatocyte-derived EVs promote endothelial inflammation and facilitate atherogenesis by miR-1 delivery, Krüppel-like factor 4 (KLF4) suppression and nuclear factor kappa-light-chainenhancer of activated B cells (NF-κB) activation [39]. On the other hand, inflammatory cells also release EVs to communicate with hepatocytes. Indeed, neutrophil-specific miR-223 is transferred through EVs from neutrophils to hepatocytes via low-density lipoprotein receptor (LDLR) and apolipoprotein E (APOE) to ameliorate NASH in mice [40]. The mechanism by which miRNAs are targeted to EVs and participate to the pathogenesis of NASH needs further attention and might present an interesting therapeutic potential (see Fig. 1).

Recent studies report that EVs derived from outside the liver, such as microbes or adipose tissue, play important roles during NASH progression. Diet-induced changes in microbiota alter the composition of microbe-derived EVs. In vivo administration of microbe-derived EVs from high-fat diet promotes insulin resistance in both skeletal muscle and adipose tissue in mice fed a regular diet [41]. This suggests a need to explore in more details the role of microbe-derived EVs in NASH pathogenesis and in liver disease more generally. Adipose tissue – liver inter-organ crosstalk seems to have a prominent role in NASH pathobiology. It has been shown that adipocyte-derived EVs promote hepatic inflammation via transporting IL-6 and monocyte chemoattractant protein-1, increase insulin resistance and impair gluconeogenesis [42]. These adipocyte-derived EVs can also contribute to tissue remodeling via alteration of tissue inhibitor of metalloproteinases 1 expression in hepatocytes and HSCs [43]. However, EVs derived from brown adipose tissue seem to have a beneficial role in NASH mouse models [44], suggesting a therapeutic potential of the brown adipose tissue-derived EVs in metabolic diseases. Adipose tissue – liver crosstalk also occurs in the other direction. Indeed, a recent study demonstrated that liver is the first organ to respond to lipid overload and sends hepatocyte-derived EVs targeting adipocytes to regulate adipogenesis and lipogenesis [45]. More precisely, geranylgeranyl diphosphate synthase (Ggpps) expression in liver is enhanced by lipid overload and regulates EV secretion through Ras-related protein Rab27A geranylgeranylation to remodel adipose tissue to adapt to metabolic changes in response to lipid overload. Consistently, liver-specific Ggpps-deficient mice have reduced fat adipose deposition [45]. Utilizing geranylgeranylation inhibitors, such as GGTI-2133, might be a treatment option for patients with NASH.

2.2. Alcohol-associated liver disease

Alcohol-associated liver disease (ALD) is the leading cause of liver-related morbidity and mortality worldwide, contributing to more than half liver cirrhosis cases in Western countries [46]. It has been reported that patients with alcohol-associated liver disease (ALD) and alcohol-fed mice have increased EV concentration in their circulation as compared to healthy individuals [47,48]. Patients with alcoholic hepatitis (AH), the most severe form of the ALD spectrum, have higher plasma concentrations of several subpopulations of EVs than heavy drinkers without AH [49,50]. In addition to increased circulating EV release, EV content is also altered in AH. In this regard, in a murine model of AH, hepatocyte-derived EVs are enriched with miR-27a and miR-181. When transfected in HSCs, miR-27a and miR-181 repress nuclear receptor subfamily 1 group D member 2 (Nr1d2), an HSC quiescent marker leading to HSC activation [51]. Furthermore, in the same AH murine model, hepatocytederived EVs upregulate pro-fibrotic IL-1β and IL-17 expression in macrophages in a TLR9-dependent manner [51]. In ALD, miR-27a is also enriched in EVs derived from alcohol-exposed monocytes and promotes M2-like functional polarization of naïve monocytes [52]. In line with these studies, alcohol-induced liver injury promotes enrichment of miR-122 in EVs derived from hepatocytes, which can be taken up by monocytes. Then, transferred miR-122 inhibited heme oxygenase 1 pathway leading to increased sensitization to lipopolysaccharide (LPS) and pro-inflammatory cytokine expression [53]. This group also demonstrated that, in hepatocytes and macrophages, alcohol-mediated miR-155 targets impairs autophagic flux at multiple steps by targeting mammalian target of rapamycin (mTOR), Ras homolog enriched in brain (Rheb), lysosomal-associated membrane protein 1 (LAMP1) and LAMP2. LAMP1 and LAMP2 downregulation leads to decreased lysosomal function and increased EV release [54].

Beside miRs, alcohol-induced EVs contain other types of molecules which are shown to be pro-inflammatory. It has been shown that CD40 ligand (CD40L)-containing EVs derived from alcoholinjured hepatocytes activate macrophages [14]. In line with this, alcohol consumption promotes the release of mtDNA-enriched EVs from injured hepatocytes in an apoptosis signal-regulating kinase 1 (ASK1) and p38-dependent manner. In turn, these mtDNA-enriched EVs induce hepatotoxicity and neutrophil accumulation in the liver, leading to alcoholic steatohepatitis [55,56]. Alcohol also mediates the release of hepatocyte-derived EVs enriched with cytochrome P450 family 2 subfamily E member 1 (CYP2E1), which is an alcohol metabolizing enzyme. These CYP2E1-enriched EVs participate in the toxicity of alcohol towards monocytes [57]. In a recent study utilizing a model of acute alcohol-induced liver injury, lipid accumulation in hepatocytes can also be promoted by EVs released by intestinal epithelial cells, strengthening the importance of the gut-liver axis in liver disease [58]. It might be interesting to explore new therapeutic strategies against the release of pathogenic EVs in patients with AH (see Fig. 2).

Fig. 2. EVs in progression of alcohol-associated liver disease (ALD).

Alcohol injury provokes an increased EV release by inhibition of autophagy pathway. Released EVs are enriched with various molecules (miRNAs, proteins, DNA) and exert effects on several liver cell types in the liver. They can activate hepatic stellate cells and contribute to fibrosis deposition. They are taken up by monocytes, macrophages and neutrophils, overall contributing to increased liver inflammation. Importantly, liver lipid accumulation can also be provoked by gut derived EVs. Prepared with Biorender.com.

2.3. Viral hepatitis

Viral hepatitis infection is a risk factor for chronic liver disease affecting close to 397 million individuals worldwide with both hepatitis B virus (HBV) and HCV infection cases combined. Longterm infection can lead to cirrhosis and cancer. Due to the similarities with EVs, such as size and tethering mechanisms, viruses may hijack EV biogenesis mechanism to increase infectivity and transmissibility. Moreover, EVs released from infected cells may modulate the host immune system against the virus [59].

Several studies have reported that EVs from HCV-transfected or HBV-infected hepatocytes contain viral RNA [60,61]. EVs with HBV-RNA were able to induce expression of natural killer group 2D (NKG2D) ligand on macrophages, which stimulates interferon γ (IFN-γ) from NK cells [61]. EVs with HCV-RNA were shown to activate plasmacytoid dendritic cells which produce type I IFN [60]. In addition to viral RNA, infected hepatocyte-derived EVs contain viral DNA. HBV-DNA-enriched EVs were taken up by uninfected hepatocytes leading to detectable HBV DNA levels, as well as hepatitis B s antigen and hepatitis B core antigen expression in hepatocytes [62]. Moreover, HBV RNA and DNA-enriched EVs from chronic hepatitis B patients were able to be taken up by NK cells isolated from healthy donors leading to decreased NK cell cytotoxicity, TNFα and IFNγ expression [62]. Similarly, EVs from HCV-infected hepatocytes induced galectin-9 production in monocytes [63], where galectin-9 promotes the expansion of regulatory T cells and apoptosis of HCV-specific T cells [64]. Moreover, EVs from HBV-infected hepatocytes were endocytosed by macrophages where they upregulated programmed death-ligand 1 (PD-L1) expression, which is known to bind to programmed cell death protein 1 (PD-1) on T cells and suppress T cell activation [65]. The same group recently demonstrated the immunosuppressive function of EVs from HBV-infected hepatocytes in vivo [66].

The intracellular adaptor protein syntenin participates in exosome biogenesis and is shown to increase the enrichment in EVs of the HCV envelope glycoprotein E2. These E2-coated exosomes renders HCV infectivity less susceptible to antibody neutralization in hepatoma cells and primary human hepatocytes [67]. Nonenveloped viruses, such as hepatitis A virus (HAV), can also hijack the exosome biogenesis machinery to form a quasi-enveloped virus. In the case of hepatitis A virus (HAV), this happens through interaction of the HAV structural protein pX with one of the exosome biogenesis proteins, ALG-2-interacting protein X (ALIX), leading to the secretion of virions and foreign proteins through exosome-like vesicles [68]. In line with this study, the delivery of exosome cargo containing viral RNA of HAV is mediated by the phosphatidylserine receptor HAV cellular receptor 1 (HAVCR1) and the cholesterol transporter Niemann-Pick disease, type C1 (NPC1), indicating that that viral infection, via this exosome mimicry mechanism, does not require an envelope glycoprotein [69].

A particular interest has been shown towards miRs in the context of viral hepatitis. Plasmid HBV-transfected hepatocytes release miR-21 and miR-29a-enriched EVs, which downregulate the proinflammatory IL-12 expression [61]. Furthermore, EVs from patients with chronic hepatitis B induce hepatocyte cancer cell line proliferation by transferring miR-25-3p [70], suggesting that EVs from infected host cells may promote the progression of viral hepatitis towards liver cancer. Regarding hepatitis C, EVs from HCV-infected hepatocytes are enriched with miR-19a which activates signal transducer and activator of transcription 3 (STAT3)-mediated tumor growth factor β (TGFβ) signaling in HSCs leading to matrix deposition [71].

Altogether, these studies suggest that hepatitis viruses utilize EV biogenesis machinery to promote their transmissibility and escape from the host immune system. This leads to the establishment of viral persistence, unproductive immune response, inflammatory microenvironment in the liver which is later translated in fibrosis, cirrhosis, and cancer [59] (see Fig. 4).

Fig. 4. EVs in HBV and HCV infections.

HBV and HCV infection cause release of hepatocyte EVs that can be taken up by immune cells. In HBV infection, these EVs target NK cells and macrophages, promoting decrease of cell cytotoxicity and T cell activation by macrophages and therefore escape from host immune system. Hepatocyte also emit EVs supporting cancer cell proliferation. In HCV infection, hepatocyte-derived EVs are taken up by monocytes, which supports expansion of anti-inflammatory T-reg cells and apoptosis of HCV-specific T cells, all supporting escape from host immune system. These EVs can also be taken up by HSC and contribute to increased extracellular matrix deposition. Prepared with Biorender.com.

2.4. Cholestatic liver disease

Cholestatic liver diseases include primary biliary cholangitis (PBC) and primary sclerosis cholangitis (PSC), which are characterized by cholestasis, inflammation and fibrosis. In cholestasis liver diseases, cholangiocytes are activated, with sustained proliferation leading to what is commonly known as ductular reaction. Depending on the cell of origin and physiological or pathological condition, EVs differently regulate cholangiocyte proliferation. In this regard, bile-derived small EVs interact with cholangiocyte primary cilia to decrease extracellular signal regulated protein kinase 1/2 (ERK1/2) phosphorylation, increase miR-15a expression and inhibit cholangiocyte proliferation [72]. Congruently, liver stem cell-derived EVs are enriched with miR let-7. They reduce the levels of ductular reaction mediators and biliary fibrosis through the inhibition of NF-κB and IL-13 signaling pathways [73]. On the other hand, serum levels of EV-associated long non-coding RNA (lncRNA) H19 correlate with the severity of PSC in patients [74]. Cholangiocyte-derived EVs mediate the transfer of lncRNA H19 into hepatocytes and promote cholestatic injury by suppressing small heterodimer partner (SHP) expression [74]. Moreover, EV-associated H19 induce cholangiocyte proliferation by upregulating high-mobility group AT-hook 2 (HMGA2) levels [75], HSC activation and matrix deposition [76], as well as macrophage activation, differentiation, and chemotaxis through C-C motif chemokine ligand 2 (CCL-2)/ C-C motif chemokine receptor 2 (CCR-2) signaling pathways [77] and S100 calcium binding protein A11 (S100A11) [78]. Further studies are thus needed to clarify the precise implication of different types of EVs in development and progression of cholestasic liver diseases in patients.

3. Role of EVs in cirrhosis development and complications

Cirrhosis can occur from several types of liver injury, such as high fat diet, excessive alcohol consumption, or viral infection. In this section we will comment on the role of EVs in the development of cirrhosis, progression of portal hypertension, and complications of cirrhosis.

3.1. Cirrhosis development

Cirrhosis development involves liver fibrogenesis, but also liver angiogenesis and intrahepatic activation of coagulation.

3.1.1. Fibrogenesis

Activation of hepatic stellate cells (HSCs) is crucial for the initiation of fibrogenesis [79]. Accordingly, most studies focus on how the microenvironment drives HSC activation and subsequent matrix deposition. In this regard, hepatocyte-derived lipotoxic EVs are enriched with miR-128-3p and can be internalized by HSCs [80]. miR-128-3p is subsequently involved in HSC activation and liver fibrosis by down-regulating the HSC quiescent marker peroxisome proliferator-activated receptor (PPAR) γ. Moreover, miR-128-3p levels in circulating EVs are associated with increased fibrosis in mouse models of liver injury [80]. In line with these findings, injured hepatocytes release EVs enriched with miR-192 and promote fibrogenic signaling in HSCs [81]. In a mouse model of carbon tetrachloride (CCl₄)-mediated liver fibrosis, endothelial cells have also been shown to promote HSC activation via transfer of sphingosine kinase 1 (SK1)-enriched EVs [82]. These SK1-enriched EVs are endocytosed in a dynamin-dependent manner and induce AKT phosphorylation in HSCs to promote subsequent cell migration [82].

Pro-fibrotic EVs are also released by activated HSCs which are thought to induce other HSC activation in a paracrine manner. Patients with cirrhosis have increased circulating concentrations of EV containing PDGF receptor α (PDGFRα) compared to controls [83]. Platelet-derived growth factor (PDGF)-activated HSCs release fibrogenic EVs enriched with PDGFRα [83]. More specifically, PDGF binding to PDGFRα promotes tyrosine 720 phosphorylation which recruits Src homology region 2 (SH2)-containing protein tyrosine phosphatase-2 (SHP2). Subsequently, SHP2 inhibits PDGFRα degradation and promotes its enrichment in EVs [83]. In addition, SHP2 promotes mTOR signaling, which inhibits the autophagic degradation of multivesicular bodies leading to increased fibrogenic EV release [84]. SHP2 and mTOR dependent EVs induce HSC migration in vitro and liver fibrosis in vivo [83,84]. In line with these studies, another group showed that EVs released by activated HSCs have decreased levels of Twist family basic helix-loop-helix transcription factor 1 (Twist1) and miR-214 leading to cellular communication network factor 2 (CCN2) expression in recipient HSCs and their subsequent activation [85]. Thus, it would be of therapeutic interest to investigate the role of the Twist1 inhibitor, harmine [86], on fibrogenic EV release and liver fibrosis (see Fig. 3).

Fig. 3. EVs in cirrhosis and portal hypertension.

In the hepatic parenchyma, injured hepatocyte and endothelial cells release EVs able to activate quiescent HSCs and their migration, contributing to fibrosis deposition. Activated HSCs also emit EVs contributing to activation of quiescent HSCs in a vicious circle. Active HSC, injured hepatocytes and cholangiocytes release EVs inducing angiogenesis. Patients with cirrhosis have an increase of blood EVs containing TF, contributing to a pro-coagulant state. Cirrhosis induces an increase of liver resistance to blood flow, leading to portal hypertension (PH), associated with cirrhosis complications. Patients with cirrhosis have blood EVs that can be taken up by endothelial cells and act as phospholipid donors; they are metabolized into vasodilatary prostaglandins hindering normal smooth muscle cell contraction. This leads to an increase in blood flow in the hepatic artery, further causing progression of PH. Prepared with Biorender.com.

3.1.2. Angiogenesis

Intrahepatic angiogenesis contributes to progression to cirrhosis and angiocrine signaling is crucial in this process [87,88]. However, how pathogenic EVs are related to angiogenesis remains incompletely understood. Lipotoxic hepatocyte-derived EVs can mediate angiogenesis in liver injury, which correlates with fibrosis severity. These EVs are enriched with vascular non-inflammatory molecule-1 (vanin-1) [89], a surface protein and important enzyme in converting pantetheine to pantothenic acid (vitamin B5). Vanin-1 mediates EV internalization by endothelial cells promoting subsequent endothelial cell migration, in vitro tube formation and in vivo angiogenesis in an experimental liver injury mouse model [89]. Treatment of methionine and choline-deficient (MCD) dietfed mice with vanin1 siRNA decreased pathogenic angiogenesis associated with NASH [89]. In line with these findings, shortterm administration of vanin-1 inhibitor RR6 completely inhibits circulating vanin-1 and improves insulin sensitivity; however it doesn’t affect hepatic steatosis [90]. It would be interesting to explore the long-term effect of this inhibitor on lipotoxic hepatocyte-derived EV release, liver injury-associated angiogenesis as well as hepatic steatosis.

In addition to EVs derived from hepatocytes, EVs released from cholangiocytes and PDGF-activated HSCs contain hedgehog ligands which promote a pro-angiogenic profile in LSECs and neovascularization in rats [91]. Likewise, EVs derived from portal myofibroblasts laden with vascular endothelial growth factor A (VEGFA) signal to liver endothelial cells and increase liver angiogenesis [92]. It would be of interest to develop inhibitors of the release of pro-angiogenic EVs strategies in patients with fibrosis and cirrhosis (see Fig. 3).

3.1.3. Coagulation

Coagulation activation associated with chronic liver diseases contributes to progression of cirrhosis [93]. Pro-coagulant large EVs contain anionic phospholipids on their outer membrane leading to the assembly of the components of the coagulation cascade. Moreover, large EVs expressing tissue factor at their surface are highly pro-coagulant because they can initiate the extrinsic pathway of blood coagulation. Circulating concentrations of tissue factor-enriched EVs are increased both in mice and patients with cirrhosis [94,95], suggesting that EVs might contribute to progression to cirrhosis via coagulation activation. Yet, this view is not supported by experimental data since mice deficient in tissue factor have similar levels of bile duct ligation-mediated liver fibrosis as compared to controls [94]. Testing other mouse models of cirrhosis would be useful to draw firm conclusions on this aspect.

Platelets are key players in hemostasis, but the role of plateletderived EVs in cirrhosis development remains uncertain. Plateletderived EVs have not been consistently found elevated in the setting of cirrhosis [6] and their circulating concentrations not associated with patients’ mortality [966]. However, a recent study reported that six-month transplant-free survival was lower in patients with low baseline level of circulating small Annexin V^± platelet-derived EVs [97]. Authors hypothesized that overconsumption of Annexin V platelet-derived EVs may promote coagulation activation and parenchymal extinction, leaving open the debate on the role of platelet-derived EVs in cirrhosis (see Fig. 3).

3.2. Portal hypertension

At early steps of cirrhosis development, portal hypertension results from increased hepatic vascular resistance to portal blood flow. This is due to the structural liver changes mentioned above, including fibrosis, angiogenesis and vascular occlusion, but also to a dynamic component related to an increased hepatic vascular tone [98–100]. Although the role of EVs in this intrahepatic vasoconstriction has not been assessed so far, indirect evidence suggests that they might be implicated in this process. In this regard, it has been recently demonstrated that Janus kinase 2 (JAK2) over-activation, in the context of myeloproliferative neoplasms, leads to the release of EVs able to transfer oxidative stress to endothelial cells leading to decreased nitric oxide (NO) availability and vasoconstriction [101]. JAK2 is overexpressed in the liver of patients with cirrhosis, mainly in activated hepatic stellate cells [102]. The same study suggested that ROCK signaling pathway is overexpressed in the cirrhotic liver, particularly in hepatic stellate cells [102]; this pathway is known to be implicated in EV release [12,84]. A possible hypothesis would thus be that JAK2/ROCK pathway in the liver would induce EV release, which in turn aggravates hepatic vascular resistance.

Splanchnic arterial vasodilatation associated with enhanced cardiac output and opening of portosystemic collaterals occurs following the establishment of portal hypertension. Splanchnic arterial vasodilatation causes an incremental portal blood flow which aggravates and perpetuates the portal hypertension syndrome [98–100]. Large EVs contribute to this arterial vasodilatation. Indeed, plasma large EVs from patients with Child-Pugh B or C cirrhosis impair the vasoconstriction ability of rat aortic rings ex vivo. Furthermore, these large EVs decrease arterial blood pressure in mice through transferring phospholipids to endothelial cells [96]. Phospholipid metabolism leads to the formation of prostacyclin, which causes vascular smooth muscle relaxation. Although the cell of origin of these large EVs causing this vasodilatation is not identified, the hyporeactive effect is specific to circulating EVs from patients with advanced cirrhosis and is not observed with EVs from patients with Child–Pugh A cirrhosis or healthy individuals [96].

Caspase activation, which enhances EV release [89], has been demonstrated to be involved in portal hypertension. Indeed, Emricasan, a pan-caspase inhibitor, has recently been shown to decrease hepatic venous pressure gradient in a phase II trial and then in a multicenter double-blinded study, but only in patients with severe portal hypertension [103,104]. This specificity for patients with severe portal hypertension is reminiscent of the effect on arterial vasodilatation observed only with EVs from patients with Child-Pugh B or C cirrhosis. This would suggest that Emricasan lowers portal hypertension by decreasing the release of certain EVs involved in the splanchnic arterial vasodilatation component of portal hypertension. This hypothesis seems contradicted by the study by Gracia-Sancho et al. on the effect of Emricasan in rats with CCl₄-mediated cirrhosis. Indeed, the authors observed that Emricasan reduces portal pressure without portal blood flow modulation, pointing towards an effect on the intrahepatic component of portal hypertension [105]. Authors show that Emricasan exerts its protective role in part through a paracrine mechanism on nonparenchymal cells mediated by hepatocyte-derived EVs [105]. However, it should be highlighted that the CCl₄ model used in this study by Gracia-Sancho et al., is not suitable for analysis of large EVs. Indeed, animals treated with CCl₄ have low circulating concentrations of large EVs which might be due to the high phospholipase A2 activity induced by CCl₄. Phospholipase A2 is able to hydrolyze phospholipids, such as phosphatidylserine or phosphatidylcholine, exposed at the surface of large EVs [13]. Altogether, we can speculate that, in advanced cirrhosis, large EVs might favor arterial vasodilation and small EVs intrahepatic vascular resistance (see Fig. 3).

3.3. Intestinal permeability, gut-liver axis and systemic inflammation

Gut-liver axis refers to the crosstalk between the gut and the liver including transport of gut-derived products to the liver, the transfer of liver-derived products, i.e. bile and antibodies to the intestine [106]. Recent evidence underlines the importance of the gut-liver axis in chronic liver disease. In particular, cirrhosis is associated with marked gut barrier impairment that contributes to portal hypertension [106]. Although bacteria can release EVs [107,108], their role in intestinal permeability in the context of cirrhosis has not been studied. In other settings, some studies demonstrate that EVs can modulate intestinal permeability. For example, EVs derived from the bacterium Akkermansia muciniphila reduce gut permeability in mice fed a high fat diet through increasing expression of tight junction proteins. This was confirmed in vitro where, Akkermansia muciniphila-derived EVs restored tight junction by increasing occludin expression in LPS-exposed Caco-2 cells [109,110]. Akkermansia muciniphila is known to be reduced in microbiota of patients with alcohol-associated liver disease, suggesting that a defect in Akkermansia muciniphila-derived EVs might contribute to gut permeability. Along the same lines, supplementation of mice with alcohol-associated liver disease with Akkermansia muciniphila restores intestinal barrier function leading to the reversal of liver disease features [111]. Besides bacteria-derived EVs, plasma EVs can also enhance intestinal permeability [112]. Bacterial-derived EVs enriched with LPS have been observed in patients with various causes of intestinal barrier dysfunction other than liver disease [113]. Given the key role of LPS in cirrhosis, this observation calls for dedicated studies in patients and animal models of cirrhosis [114].

Systemic inflammation, as reflected by elevated concentrations of C-reactive protein (CRP) in the circulation, is a key feature of decompensated cirrhosis [115,116]. When bound to EVs, CRP undergoes conformational changes leading to the expression of neoepitopes. This enhances its binding to C1q, activates the complement system, and thus increases recruitment of leukocytes to the site of damage [117]. The effect of circulating CRP-enriched EVs in cirrhosis deserves further investigations (see Fig. 3).

3.4. Complications of cirrhosis

Complications of cirrhosis include ascites, hepatic encephalopathy, hepato-pulmonary syndrome and porto-pulmonary hypertension. Ascites, defined as an accumulation of fluid in the peritoneal cavity, results from splanchnic arterial vasodilation and enhanced endothelial permeability in mesenteric beds [118]. Patients with cirrhosis and ascites have higher circulating concentrations of hepatocyte-derived and endothelium-derived EVs than patients with cirrhosis without ascites [119,120]. Moreover, concentrations of EVs present within ascites in patients with cirrhosis inversely correlate with 30-day mortality [121]. Interestingly, small EVs derived from ascites favor inflammation by inducing NF0κB in monocytes [122]. However, more studies are needed to decipher the role of EVs on ascites development.

Hepatic encephalopathy results from astrocytes swelling due to hyperammonemia and cerebral edema, and is aggravated by systemic inflammation [123]. Some recent data show that EVs contribute to hepatic encephalopathy. Indeed, Izquierdo-Alrejos et al. showed that compared to controls, plasma-derived EVs from hyperammonemic rats are enriched with proteins being mainly associated with immune system processes. In this study, when injecting EVs from hyperammonemic rats into control rats, these EVs entered Purkinje neurons and microglia, and induced motor incoordination [124]. In another study, systemic inflammation, induced by LPS or recombinant TNFα injection in mice, stimulated the release of miRNA-containing EVs by the choroid plexus epithelium in cerebro-spinal fluids. In this study, EVs were purified from cerebrospinal fluids of LPS-injected mice and injected into lateral ventricle of naive mice. These EVs were able to cross the ependymal cell layer and penetrate into the brain parenchyma to transfer their proinflammatory message to recipient brain cells [125]. These studies prove the concept that EVs can contribute to hepatic encephalopathy and pave the way for future, more detailed analyzes.

Hepatopulmonary syndrome is defined by the presence of a triad including chronic liver disease, dilatation of precapillary and postcapillary pulmonary vasculature and abnormal gas exchange [126]. These changes involve lung angiogenesis, vascular remodeling and excessive production of vasodilators, including nitric oxide (NO) and carbon monoxide (CO) [127]. As discussed above, portal myofibroblast can release EVs and activate VEGF signaling in endothelial cells. It might be possible that VEGF-laden EVs could reach pulmonary vasculature via the hepatic veins and contribute to hepatopulmonary syndrome. Moreover, a recent study by Chen et al. showed that rats with bile duct ligation and hepatopulmonary syndrome have higher levels of EV-associated miR-194. EVs containing miR-194 induce proliferation, migration and tube formation of pulmonary microvascular endothelial cells ex vivo. Inhibiting exosomal secretion in vivo significantly decrease angiogenesis [128].

Porto-pulmonary hypertension is the concomitant presence of pulmonary arterial hypertension and portal hypertension in patients with or without cirrhosis [129]. Both the liver and the portal system may regulate vasoreactivity and angiogenesis in the pulmonary circulation [130]. A contribution of EV to this complication is likely since injection of EVs from pulmonary hypertensive mice to healthy mice is able to induce pulmonary hypertension [131]. Further studies are needed to unveil the implication of EVs in porto-pulmonary hypertension.

4. Role of EVs in liver cancer

According to American Cancer Society, more than 800,000 people are diagnosed each year with primary liver cancer throughout the world. Liver cancer is a leading cause of cancer deaths worldwide, accounting for more than 700,000 deaths each year. Recently, several studies have demonstrated the importance of EVs in primary liver cancer progression by modulating the immune tolerance and supporting cancer cell survival and resistance. The most common type of liver cancer is hepatocellular carcinoma (HCC). EVs have been shown to derive from HCC cells or other cell types to support cancer progression. The expression of ultra-conserved long non-coding RNAs (ucRNA) is altered in HCC. Interestingly, ucRNAs can be horizontally transferred through EVs. Kogure et al. demonstrated that the most highly significantly expressed ucRNA in HCC cell-derived EVs is TUC339, involved in HCC cell proliferation and spread [132]. Another long non-coding RNA transferred through liver cancer cell EVs is lncRNA H19, which promotes angiogenesis and tumor progression [133]. EVs from HCC cells promote HCC cell proliferation and epithelial-tomesenchymal transition (EMT) by transferring miR-3129 that targets thioredoxin interacting protein (TXNIP) [134]. Moreover, high-metastatic HCC cells secrete exosomal miR-1247-3p that directly targets Beta-1,4-Galactosyltransferase 3 (B4GALT3), leading to activation of β1-integrin-NF-κB signaling in fibroblasts [135]. Activated cancer-associated fibroblasts further promote cancer progression by secreting pro-inflammatory cytokines, including IL-6 and IL-8 [136]. Other miRs have been reported to be differentially expressed in HCC patients when compared to patients with cirrhosis. Indeed, pro-proliferative and promigratory exosomal miR-222, miR-221, and miR-18a were upregulated in the serum of patients with HCC, whereas anti-apoptotic and anti-growth miR-101, miR-122 and miR-195 were downregulated [137,138]. In addition, HCC-derived EVs transform normal cells such as HSCs through miR-21, macrophages through miR-23a-3p and adipocytes into cancer-supporting cells [139–141].

HCC cells respond to EVs from other cell types, such as adipocytes. In patients with high body fat ratio, miR-23a was transferred through EVs from most likely adipocytes to promote HCC cell growth and migration [142]. In line with this, adipose-secreted circular RNAs through EVs were found to regulate deubiquitination in HCC, thus facilitating cell growth and reducing DNA damage via the suppression of miR-34a and the activation of deubiquitination-related USP7 [143].

Cholangiocarcinoma (CCA) is a malignancy with features of the biliary epithelium. Although an uncommon cancer, CCA is the second most common primary liver tumor after HCC [144]. Several studies have shown the importance of EVs, and especially their RNA cargo, in the development of CCA. A recent study demonstrated that cholangiocarcinoma-associated circular RNA 1 (circ-CCAC1) levels are increased in cancerous bile-resident EVs and tissues. In CCA cells, circ-CCAC1 increased cell progression through miR-514a-5p and ying yang 1 (YY1). Moreover, CCA-derived EVs was transferred to endothelial monolayer cells, disrupting endothelial barrier integrity and inducing angiogenesis [145]. Other non-coding RNAs, such as miR-195 and miR-30e, have beneficial roles. Indeed, CCA cells inhibit miR-195 expression in LX2 cells, which are immortalized HSCs. MiR-195 overexpressing LX2 release miR-195-enriched EVs which inhibit CCA cell growth and invasiveness [146]. In line with this study, miR-30e is also reduced in CCA cells. Its overexpression in CCA cells and its transfer through EVs inhibited TGFβ-mediated endothelial-to-mesenchymal transition, cell invasion and migration in recipient CCA cells [147].

In summary, cancer EVs decrease the immune response and remodel tumor microenvironment to enhance cancer cell proliferation, migration and invasiveness, highlighting the importance of targeting tumor EVs in cancer treatment (see Fig. 5).

Fig. 5. EVs in HCC.

Cancer cell-derived EVs can induce transformation of HSCs and macrophages into cancer-supporting cells. They promote tumor angiogenesis, cancer cell proliferation, epithelial-to-mesenchymal transition and secretion of pro-inflammatory cytokines, all contributing to tumor progression. Prepared with Biorender.com.

5. Role of EVs in acute liver diseases

5.1. Acute liver injury

Acute liver injury results from a short insult to the liver that can rapidly progress and lead to systematic consequences and eventually to multi-organ failure. A well-established cause of acute liver injury is acetaminophen (APAP)-induced acute hepatotoxicity. It has been shown that administration of circulating EVs from APAP-treated induce hepatocyte apoptosis with elevated TNFα/IL-1β expression in recipient mice [148]. CD133⁺ hematopoietic stem cell-derived EVs are elevated in APAP-treated mice in a CD39-dependent manner. Moreover, patients with liver injury display elevated levels of CD39⁺ EVs in their circulation [149]. However, the role of hematopoietic stem cell-derived EVs in acute liver injury remains unknown. Umbilical cord mesenchymal stem cell-derived plasma EVs might display a beneficial role in acute liver injury. Indeed, in vitro these EVs protect APAP-treated hepatocytes by inhibiting cell death [150]. It would be interesting to know the role of mesenchymal stem cell-derived EVs in an in vivo model of acute liver injury.

5.2. Hepatectomy and ischemia-reperfusion

Hepatectomy, the surgical removal of all or a portion of the liver, is a main therapeutic strategy for HCC or liver failure which is accompanied by ischemia-reperfusion. Partial hepatectomy triggers hepatocyte proliferation to restore the lost liver mass. In a rat model of partial hepatectomy, the inflammation-associated miR-150 and miR-155 and the regeneration-associated miR-21 and miR-33 are up-regulated in serum-derived extracellular vesicles [151]. It has been shown that EVs derived from thymus cell antigen 1 (Thy1)^± cells induce hepatocytic progenitor cell proliferation through IL17B receptor signaling. Indeed, administration of Thy1-EVs in vivo increased the number and size of hepatocyte progenitor clusters in partial hepatectomy rat livers [152]. Platelets also participate to hepatocyte proliferation through interacting with LSECs and HSCs. However, this effect is independent of platelet-derived EVs [153]. In a mouse model of hepatectomy, serum mononuclear cell-derived large EVs transfer miR-142-3p to endothelial cells decreasing TNFα levels and endothelial apoptosis [154]. Moreover, hepatocyte EVs, but not other cell type-derived EVs, induce dose-dependent hepatocyte proliferation in vitro and in vivo in mouse models of partial hepatectomy and ischemia/reperfusion. Mechanistically, hepatocyte-derived EVs transfer neutral ceramidase and sphingosine kinase 2 (SK2) causing increased synthesis of S1P in recipient hepatocytes [155]. Under mechanical stress due to hepatectomy, hepatocytes and Kupffer cells also release ATP-enriched vesicles from a lysosomal compartment to promote hepatocyte proliferation and liver regeneration [156]. In a more therapeutic perspective, EVs from human umbilical cord blood mesenchymal stem cells enhance liver regeneration after hepatectomy in rats by transferring miR-124 and inhibiting Forkhead box protein G1 (FOXG1) [157].

Ischemia/reperfusion is observed in liver surgeries such as hepatectomy and liver transplantation. Ischemia/reperfusion promotes interferon regulatory factor 1 (IRF-1)-mediated Rab27a transcription and EV release leading to oxidized phospholipids, activation of neutrophils and subsequent hepatic IR injury [158]. Human liver stem cell and human mesenchymal stromal cellderived EVs reduce liver injury and inflammation and increase hepatocyte proliferation in a mouse experimental model of ischemia/reperfusion [159,160]. These studies point towards the potential of EVs in attenuating liver injury during hepatectomy and ischemia/reperfusion, which can be further alleviated at the human level.

6. Therapeutic potential of EVs

Therapeutic strategies involving EVs have been applied in clinical trials for several diseases (NCT04356300, NCT04173650, NCT04657458, NCT02931045, NCT04281901). In the field of liver diseases, therapeutic efforts involving EVs are nascent. There are two main strategies: 1. EVs as a therapeutic target, and 2. EVs as tools to treat liver diseases.

6.1. EVs as a therapeutic target

As described above, EVs can participate to disease establishment. Therefore, strategies targeting EV biogenesis, release and cargo present a potential interest. In a pre-clinical model of alcohol-associated liver disease, hepatocyte-derived small EVs are taken up by other hepatocytes and HSCs. In vitro, the exosome biogenesis inhibitor GW4869, which blocks neutral sphingomyelinase-dependent exosome biogenesis [161], inhibited the intercellular transfer of RNA [162]. In vivo, GW4869 administration to a mouse model of non-alcoholic steatohepatitis alleviated the high fat diet-induced fatty livers [163].

In liver diseases, the release of EVs can be enhanced in response to various stimuli, such as ER stress [24], lipotoxicity [12,27] and autophagy [84]. Exposure of hepatocytes to palmitate induces pro-inflammatory EV release in an IRE1α and ROCK1 dependent manner and their enrichment in S1P [12,24]. Utilization of sphingosine kinase inhibitors, of S1P receptor inhibitors and of the ROCK1 inhibitor fasudil abolish macrophage chemotaxis [12,24]. Fasudil also inhibits the release of CD40L-enriched EVs from alcohol-injured hepatocytes and attenuates in vivo inflammation and liver injury in a mouse model of alcohol steatohepatitis [14]. Similar to these studies, MLK3 inhibitors URMC099 and CLFB1134 also reduced pro-inflammatory EV release from hepatocytes [27].

Autophagy has been reported to be closely linked to EV release. Indeed, inhibition of autophagy by mTOR, downstream SHP2, leads to increased fibrogenic EV release in HSC, and these EVs participate to liver fibrosis in vivo [84]. Blocking SHP2 by its allosteric specific inhibitor SHP099 abolish EV release in vitro and CCl₄-mediated liver fibrosis in vivo [83,84]. Interestingly, SHP099 also affects the cargo of activated HSC-derived EVs by inhibiting PDGFRα enrichment in these EVs. PDGFRα is known to be fibrogenic and to be regulated by autophagy [164]. In vivo, EVs derived from SHP099-treated HSCs reduced liver fibrosis compared to EVs derived from activated HSCs [83].

These studies demonstrate that pharmacological inhibition of EV biogenesis, release and cargo ameliorates liver diseases. However, these inhibitors might target other processes and their path towards clinical use requires thorough studies of their toxicity and side effects.

6.2. EVs as tools to treat liver diseases

Due to many of their biological properties, EVs can be utilized as tools to treat diseases. Translation to clinical practice requires production of clinical-grade EVs in compliance with Current Good Manufacturing Practice (CGMP). This requires the generation of a qualified cell bank (if natural EVs are being used), a scaled-up manufacturing process, and a specific release criterion for the EVs. [165,166] In this section we will include cell-derived EVs as well as engineered EVs as tools in the treatment of liver diseases.

Stem cell-derived EVs have been recently utilized in pre-clinical models of liver diseases. Intravenous injection of induced pluripotent stem cell (iPSC)-derived EVs in CCl₄ and bile duct ligation-induced liver fibrosis results in anti-fibrotic effects at protein and gene levels. In this same study, genomics analysis of these iPSC-derived EVs revealed that the most abundant microRNA is miR-92a-3p. [167] Another group utilized EVs derived from human iPSC-derived mesenchymal stromal cells (iPSC-MSC), which were intravenously administered to a murine ischemia/reperfusion model. EVs from iPSC-MSCs exert a protective and proliferative effect on hepatocytes [168]. In a murine model of liver fibrosis, mesenchymal stem cell-derived EVs attenuate CCl₄-dependent liver injury and increase hepatocyte proliferation [169]. This beneficial effect of mesenchymal stem cell-derived EVs is also confirmed in rats, where the administered EVs attenuate CCl₄-mediated liver fibrosis and high fat diet-mediated steatohepatitis by inhibiting HSC and Kupffer cell activation [170]. In addition to these stem cell-derived EVs, human liver stem cellderived EVs have been shown to promote liver regeneration. Indeed, in vitro, human liver stem cell-derived large extracellular vesicles are internalized by hepatocytes in an integrin α4-dependent mechanism. Treatment of hepatectomized rats with human liver stem cell-derived large extracellular vesicles accelerates regeneration [171]. More recently, EVs from human liver stem cells have been shown to display anti-fibrotic and antiinflammatory effects in a murine model of NASH. Out of 29 fibrosis-associated genes upregulated in NASH liver, 28 are significantly downregulated by EV treatment [172]. In addition to stem cell-derived EVs, normal hepatocyte-derived EVs exert some regenerative potential. Indeed, hepatocyte-derived small EVs deliver the synthetic machinery to form S1P in target hepatocytes resulting in cell proliferation and liver regeneration after ischemia/reperfusion injury or partial hepatectomy [155].

EVs can be found in biological fluids such as serum and milk. Liver fibrosis is suppressed by serum EVs from normal mice but not from fibrotic mice. Moreover, normal serum EVs reduce levels of hepatocyte death, inflammatory infiltration, circulating aspartate transaminase (AST)/ alanine aminotransferase (ALT) levels and hepatic or circulating pro-inflammatory cytokines [173]. In addition to serum EVs, bovine milk EVs have been shown to have a therapeutic potential as a drug delivery tool. It has been published that bovine milk EVs can be easily engineered to be loaded with small interfering RNAs to target cancer cells [174]. More specifically, the same group loaded bovine milk EVs with small interfering RNA to β-catenin and decorated with RNA scaffolds to incorporate RNA aptamers capable of binding to epithelial cell adhesion molecule (EpCAM). EpCAM is a surface marker of liver cancer stem cells. Cellular uptake of these EpCAM-targeting therapeutic milk-derived nanovesicles in vitro resulted in loss of b-catenin expression and decreased proliferation. In vivo, these engineered biological nanotherapeutics were taken up by tumor xenografts and reduced tumor weight [175]. In addition, EVs secreted from human adipose tissue-derived mesenchymal stromal which overexpress miR-125b reduce HCC cell proliferation in vitro [176]. In line with these studies, exosomal lncRNA termed SUMO Specific Peptidase 3-Eukaryotic Translation Initiation Factor 4A1 (SENP3-EIF4A1) transferred from normal cells to HCC cells inhibits the in vitro and in vivo development of HCC [177]. Other methods to bioengineer EVs have been demonstrated and a myriad of approaches have been applied in pre-clinical models of several diseases [178]. It would be of great interest in the field of liver diseases to continue to develop therapeutic strategies involving EVs and take advantage of their multiple assets.

7. Conclusion

Despite the numerous pre-clinical studies performed in the field of liver diseases, the passage to the clinical phase remains uncertain. There are several unanswered questions regarding the use of EVs as therapeutics strategies, such as their kinetics, their toxicity, their off-target effects and the delivery to a precise type of cell. Many aspects of the EV biology in health and disease remain to be discovered, with the aspiration to shed some light onto novel therapies in the perpetual efforts to cure liver diseases.