More than eight hundred million people suffer from chronic liver disease, which accounts for approximately two million deaths per year worldwide [1], with cirrhosis, viral hepatitis, and hepatocellular carcinoma (HCC) being the leading causes of liver-related deaths. Liver biopsy remains the gold standard diagnostic tool to assess the stage of liver diseases despite its significant disadvantages (low acceptance, highly invasive, and heterogeneous). The lack of non-invasive tools constitutes a significant barrier to the clinical management of liver diseases.

Circulating extracellular vesicles (EVs), a heterogeneous population of small membrane-encapsulated particles identified in several body fluids such as blood, saliva, and urine, have been proposed as markers for liquid biopsies in several diseases [2,3]. In the past decade, several shreds of evidence have suggested that EVs have a key role in liver disease, since they have been recognized as potent vehicles of intercellular communication due to their capacity to transfer proteins, lipids, and nucleic acids, thereby influencing various physiological functions of the recipient cells [4,5]. In this way, they may contribute to the pathogenesis, initiation, and progression of different liver diseases [6], emerging as a key player in cell–cell communication during acute and chronic liver disease [7]. Previous studies report changes in the number, surface markers, and cargos in the circulating EVs after liver injury, thus representing a potential biomarker for liver disorders [8]. Interestingly, changes in EVs have been demonstrated before histological signs appear, providing strong evidence of their utility as non-invasive tools even for early diagnosis [9].

NAFLD is the most common chronic liver disease worldwide, affecting about a quarter of the world’s adult population [30]. The diagnosis of steatohepatitis (NASH), a more severe form of the disease, requires a histological examination to confirm the presence of ballooning and inflammation; thus, the development of non-invasive biomarkers in NAFLD has generated considerable attention in the scientific community. Previous studies reported the role of EVs in NAFLD development, such as hepatocyte-derived sphingosine 1-phosphate (S1P)-containing EVs with pro-inflammatory activity in NASH [31]. Indeed, hepatocyte-derived EVs released from damaged hepatocytes in experimental models of NASH activate non-parenchymal cells, such as endothelial, stellate, and hepatic macrophages, contributing to the progression of liver disease [32]. A few studies support the use of several types of EVs for diagnostic purposes in patients with NASH. Circulating CD14+ MVs in NAFLD correlated with alanine aminotransferase (ALT) levels and NAScore [6]. Recently, Povero et al. [33] analyzed hepatocyte-derived EVs from pre-cirrhotic and cirrhotic NASH patients, using asialoglycoprotein receptor 1 (ASGPR1) and long-chain fatty acid transport protein 5 (FATP-5) as hepatocyte markers. Levels of ASGPR1 EVs were found to be increased twofold in pre-cirrhotic NASH and threefold in cirrhotic NASH, compared to healthy controls. Furthermore, ASGPR1+ EVs were also found to correlate with the hepatic vein portal gradient (HVPG), being a potential non-invasive biomarker of portal hypertension in patients with NASH and cirrhosis. Finally, a proteomic analysis of circulating EV cargo revealed two feature pairs (IL13Ra1 > TNR4; WISP-1 > BMP-14) with strong predictive power for liver fibrosis and cirrhosis (80% sensitivity and 80% specificity). Platelet-derived MVs have also been proposed for NASH diagnosis, with levels of CD61+ MVs correlating with fat fraction, ballooning, and fibrosis stage in NAFLD patients [34]. Another study used ASGR2 (asialoglycoprotein receptor 2) and CYP2E1 (cytochrome P450 2E1) antibodies as markers of hepatocyte-derived EVs and observed a decrease in ASGR2+ and CYP2E1+ MVs levels after bariatric surgery in 28 biopsy-proven NAFLD patients [8].

Uncontrolled alcohol consumption results in a liver injury characterized by steatosis, inflammation, hepatitis, and cirrhosis [35]. Alcohol exposure increases the number of circulating EVs of hepatic origin (mostly hepatocytes and hepatic stellate cells (HSCs)). EVs act on target cells (macrophages, endothelial cells, and HSCs) promoting inflammation and fibrosis [3]. Its release is related to a partial inhibition of autophagy promoted by a decreased level of lysosomal-associated membrane proteins LAMP1 and LAMP2 through miR-155 expression [36]. In that sense, it has been demonstrated that alcohol-treated hepatocytes cross-talk with immune cells via microRNAs (miRNAs) contained in exosomes. In alcoholic liver disease (ALD), serum/plasma miR-122 and miR-155 levels are increased and predominantly associated with exosomes [37]. Thus, hepatocyte-derived exosomes with miR-122 increase the secretion of pro-inflammatory cytokines in monocytes [38]. Also, in hepatocytes, alcohol increases the delivery and formation of exosomes containing mitochondrial double-stranded RNA (mtdsRNA) that participate in the production of pro-inflammatory cytokines (IL1B) in Kupffer cells [39]. Related to this, hepatic EVs derived from alcoholic hepatitis mice are able to activate primary HSCs, inducing a-SMA and collagen through upregulating miRNAs and increasing IL1B and IL17 production in a TLR9-dependent manner in macrophages [40]. In this regard, CD40 ligand (CD40L) in EVs, in a caspase-dependent manner in response to alcohol exposure, has a critical role as a mediator of macrophage activation [41]. Furthermore, alcohol-exposed monocytes can communicate with naive monocytes via miR-27a-loaded EVs that program naive monocytes into M2 macrophages [42]. Also, protein cargo in EVs can be important, since, in animal models of alcoholic liver disease, macrophage activation was induced by hepatocyte-derived EVs harboring heat shock HSP90 protein [43].

Alcohol has been implicated in fibrosis through the release of a major fibrogenic cytokine, transforming growth factor-beta-1 (TGF-β1), and HSCs activation [44]. A study associated alcohol with the increase of profibrogenic factors through the levels of miR-19b in HSCs and derived exosomes. Interestingly, decreased miR-19b levels in activated HSCs resulted in a change in the expression of other miRNAs (miR-17–92 cluster). However, miR-19b was induced at the plasma and exosomal levels in this alcohol-induced hepatic fibrogenesis model [45]. In addition, elevated levels of EVs have been observed containing CYP2E1 derived from the liver in patients with alcoholism and in alcohol-exposed animals. CYP2E1 activity is associated with oxidative and endoplasmic reticulum stress after alcohol consumption, leading to the activation of apoptotic pathways and toxicity to monocytes and hepatocytes. Thus, these EVs with CYP2E1 cargo could act as a biomarker for liver damage from long-term alcohol exposure [46].

Moreover, EVs secreted from other organs such as intestinal epithelial cells have effects on hepatocytes during acute alcohol injury, highlighting the importance of the gut-liver axis in ALD progression [47].

Lastly, some studies have tried to find soluble markers to diagnose alcoholic hepatitis in a non-invasive manner. The plasma levels and EVs of cytokeratin-18 fragments (M30 and M65) are reliable non-invasive markers of alcoholic hepatitis [48] High levels of CD34+ and ASGPR1+ EVs can be used as markers of non-response to corticosteroid therapy in severe alcoholic hepatitis [49]. Recently, plasma EV concentration and sphingolipid cargo were found to correlate with the severity and mortality of alcoholic steatohepatitis [50].

EVs are potent modulators of the immune response. In vitro studies showed that hepatocytes infected with replicating HBV release EVs that induced a programmed cell death 1 ligand 1 (PD-L1) expression in monocytes, possibly suppressing host antiviral activity [51]. Notably, Montaldo et al. analyzed EVs in the plasma of HCV patients after direct-acting antiviral therapy, finding that miR204-5p, miR181a-5p, miR143-3p, and miR-122-p were decreased in the EVs from HCV patients compared to healthy donors. After that, EV cargo was determined after 6 months of therapy, and miR204-p and miR143-3p were still different between healthy and HCV-treated patients, indicating that EV-mediated signals could play a causal role in fibrosis progression despite viral eradication [52]. Another study showed that patients with active hepatitis C (ALT > 100 IU/mL) had an elevated number of T cell-derived MPs compared to patients with mild hepatitis C (ALT< 40U/mL) and healthy controls [53]. Our group recently reported a decrease in endothelial and platelet apoptotic MV levels after a sustained virological response in HCV patients, concluding that this may be directly involved in the improvement of inflammation and endothelial dysfunction observed in these patients after HCV eradication [54].

Besides amplifying inflammation and modulating injury, EVs have also been demonstrated to promote liver fibrosis in NAFLD and ALD [55]. HSCs regulate the establishment and sustaining of liver fibrosis [56] partly due to their ability to chronically secrete EVs. Previous studies indicate that lipotoxic hepatocyte-derived pro-inflammatory miRNA-rich EVs could activate TLR-3 in HSCs [57], inducing their activation and migration [58]. These activated HSCs (aHSCs) in turn release EVs that contain various profibrotic proteins, lipids, and nucleic acids [59]. Moreover, the release of these EVs increases in response to liver injury [60]. Unlike aHSCs, EVs secreted by quiescent HSCs display antifibrotic properties since their cargo is shown to suppress HSC activation. Furthermore, they reduce inflammation, promote cell viability, inhibit hepatocyte apoptosis, and decrease liver transaminase levels, indicating their therapeutic potential [61,62]. Also, in the progression of NAFLD, exosomes from visceral adipose tissue (VAT) were related to fibrosis through TGF-B dysregulation in the hepatocytes and HSCs [63].

Although HCV does not replicate in HSCs, EVs from HCV-infected hepatocytes induce the expression of profibrogenic genes. miR-19a in these hepatocyte-derived HCV-EVs was able to promote fibrosis by targeting SOCS3 which caused the activation of the STAT3–TGF-B signaling pathway [64]. Previous studies aimed at determining the role of platelet and monocyte-derived MVs as biomarkers of fibrosis in biopsy-proven NAFLD patients [65]. CD14+ CD16+ EVs improved the ability of liver fibrosis scores to identify patients with F3/F4 fibrosis in a small preliminary cohort. Weil et al. reported 2.5-fold higher levels of platelet-derived MVs in 10 healthy subjects compared with 90 cirrhotic patients [66]. Finally, another study found that hepatocyte-derived MVs were 4.0-fold and 2.2-fold higher in patients with Child–Pugh C compare with those with Child–Pugh A or B respectively. Indeed, hepatocyte-derived MVs correlated with HVPG and were able to predict 6-month mortality independently of the Child–Pugh score or Model for End-Stage Liver Disease (MELD) [67].

HCC is the most common primary liver cancer, being the fourth cause of cancer-related deaths worldwide [68]. HCC has a poor prognosis due to the lack of early symptoms and the low sensitivity and specificity of available diagnostic tools. Early detection is essential to improving surveillance and the adoption of curative surgical therapies. In the same scenario, the earlier the cholangiocarcinoma (CCA) detection, the more opportunities there are for curative treatments. In the context of cancer, the role of EVs has emerged as another promising strategy for liver cancer surveillance. Besides cell-to-cell contact, intercellular communication also happens through EVs to set up and modify tumor microenvironments. EVs are released by cancer cells in order to promote tumor growth and improve the tumor microenvironment for the spreading of these cells [69]. EVs are present in circulation at the early and advanced stages of the disease. The stability and integrity of EVs and their molecular cargos may serve as useful early-stage cancer diagnostic biomarkers and therapeutic approaches [70,71,72].

Current guidelines suggest the need for non-invasive tools for the diagnosis of HCC and CCA [73]. Circulating MV levels were found to increase in HCC patients in comparison to cirrhotic patients. Furthermore, they correlated with HCC tumor size, pathological type, and TNM stages, tending to a decrease after surgical intervention [74]. Several HCC-associated surface markers have been used to isolate and quantify liver tumor EVs. Julich-Haertel and colleagues showed that the combination of annexin V+ EpCAM+ ASGPR1+ CD133+ taMPs allowed one to distinguish liver malignancies and cirrhosis. Furthermore, EpCAM+ ASGPR1+ and annexin V+ were increased in liver cancer (HCC and CCA) compared to cirrhotic patients. In addition, 7 days after tumor resection, EpCAM+ ASGPR1+ annexin V+ levels significantly decreased, showing a strong association with tumors [75]. A high expression of MMP-7-EVs could be a marker for the differential diagnosis of CCA [76]. Regarding early diagnosis, three EV subpopulations, EpCAM+ CD63+, CD147+ CD63+, and GPC3+ CD63+ were highly associated with the early diagnosis of HCC (AUROC of 0.95 (95% CI = 0.90–0.99) with a sensitivity of 91% and a specificity of 90%) [77]. Another pan-cancer marker was proposed for the diagnosis of HCC and CCA, such as EpCAM+ CD147+ EVs which were increased in HCC, CCA, and other cancers [75]. A complex technique integrating covalent chemistry-mediated EV capture/release, multimarker antibody cocktails, nanostructured substrates, and microfluidic chaotic mixers showed that purified EpCAM, ASGPR1, and CD147 EVs have a 10-gene HCC-specific signature that allows one to distinguish HCC patients from at-risk cirrhotic patients (AUROC: 0.93 (95% CI, 0.86–1.00; S: 94.4% and S: 88.5%)) [78]. Early recurrence after liver resection was related to higher levels of Hepar-1+ microparticles before surgery, suggesting its potential role as a prognostic biomarker [79]. An emerging body of evidence supports the idea that platelets have an important role in carcinogenesis, mainly in HCC development [80,81]. In fact, platelet-derived EVs have also been related to colon cancer [82], but further studies are needed in liver tumors.

The content of cancer-derived EVs significantly differs from that of healthy cells, including different types of RNA such as miRNA, lncRNA, and cancer-specific proteins [73]. miR-122 EVs allowed the differentiation of HCC from liver cirrhosis (AUC:0.990, 95% CI, 0.945–1.00). In addition, the combination of miRNA-122, miRNA-148a, and Alpha-fetoprotein (AFP) increases diagnostic accuracy (AUC:0.931, 95% CI, 0.857–0.973), suggesting that the serum vesicle microRNA signature alone or in combination with available markers could be used as a screening tool for HCC [83]. Higher miR-21 in the circulating EVs of HCC patients are better markers than serum miR-21 in differentiating HCC from cirrhotic and healthy patients [84,85]. Another cluster of miRNAs, miR-18a, miR-221, miR-222, miR-224, miR-101, miR-106b, miR-122, and miR-195, were found to have an increase in exosomes from HCC [73]. Inside purified ASGPR1+ EVs, four miRNAs, miR-10b-5p, miR-21-5p, miR-221-3p, and miR-223-3p were found to be increased in those patients with lower AFP levels [86]. The lncRNA LINC00853 in EVs showed a good diagnostic capacity for HCC (AUC:0.934, 95% IC 0.887–0.966) [87]. Another study found that lnc85 was higher in the exosomes of HCC patients with high and low levels of AFP compared to the healthy control and liver cirrhosis (AUC:0.869) [88]. Recently, a study showed that cancer-associated fibroblast (CAF)-derived MVs can be implicated in HCC progression. The survival rate in patients with low antitumoral miR-150-3p levels in plasma CAF-derived exosomes was significantly poor compared to patients with high miR-150-3p levels. [89]

Regarding CCA, it is well known that bile EVs were significantly higher in CCA patients [90]. A proteomic study showed that EVs from CCA patients expressed a specific protein profile showing potential usefulness as a diagnostic tool [91]. Many studies have shown that EVs are involved in the development and progression of liver cancer. Further pieces of evidence to better understand the role of EVs in diagnosis and prognosis are needed.