Maintenance of the highly specialized phenotype of LSECs is crucial for liver homeostasis. During liver fibrosis and cirrhosis, loss of LSEC phenotype and functions amplify liver damage by undergoing a process called ‘capillarization’, which is clarified by the loss of fenestrae (cytoplasmic pores) and the manifestation of a basement membrane, leading to LSEC dysfunction (20).

Su et al (24) mapped the spatial distribution of heterogeneous liver endothelial cells in normal and cirrhotic mouse liver and identified zone-specific transcriptomic changes of LSECs associated with liver cirrhosis using small conditional RNA-seq technology, finding that zone 3 LSECs are more susceptible to capillarization and that CD34 is more useful than conventional CD31 as a marker of capillarization in LSECs. In 2008, Deleve et al (25) employed cell co-culture to demonstrate that capillarized LSECs are unable to maintain HSC quiescence due to loss of VEGF-stimulated release of nitric oxide (NO), while differentiated LSECs can prevent HSC activation and promote reversion of activated HSCs to quiescent HSCs. This study highlighted the critical role of LSECs in regulating HSC status. In addition, capillarization of LSECs occurs in the early stage of NASH. An in vivo study showed that loss of the unique fenestrated phenotype is an initial step for fibrogenesis, preceding the activation of HSCs and onset of fibrosis in alcoholic liver injury (26), suggesting that signals from LSECs may be one of the earliest triggers of HSC activation.

The exact mechanisms regulating the loss of fenestrae have not been fully elucidated, but several molecules and pathways have been identified. Maintenance of the fenestrated LSEC phenotype requires VEGF, which is derived from hepatocytes and HSCs. VEGF is a key regulator for the LSEC phenotype (27). It has been established from various in vitro and in vivo studies that VEGF operates in two pathways, the NO-dependent and NO-independent (28). In the NO-dependent pathway, endothelial nitric oxide synthase (eNOS)-NO-cyclic guanosine monophosphate signaling is critical for regulating the formation of fenestrae in LSECs (29). Endothelial Notch activation leads to LSEC dedifferentiation and promotes liver fibrogenesis through eNOS-soluble guanylate cyclase (sGC) signaling. Regarding Notch ligands, Delta-like ligand 4 (DLL4) is predominantly expressed in endothelial cells under hypoxic conditions during liver fibrosis (30). It is reported that DLL4 overexpression accelerated defenestration of LSECs and production of basement membrane in both human fibrotic livers and carbon tetrachloride (CCl₄)-induced mouse livers (31). In addition to Notch signaling, the Hedgehog (Hh) pathway also induces LSEC capillarization and formation of vascular tubes in vitro by Hh ligands binding to their transmembrane receptors [Patched (Ptc)] to activate Smoothened (32). Besides, miR-511-3p was found to positively regulate Hh signaling by targeting Ptch1 in hepatic sinusoidal obstruction syndrome (33). Long noncoding RNA (lncRNA) Airn was identified to interact with enhancer of zeste homolog 2 to maintain LSEC differentiation through Krüppel-like factor 2 (KLF2)-eNOS-sGC pathway, thus indirectly inhibiting HSC activation (34). Studies have found that several transcription factors in liver endothelial cells also play key roles in regulating the characteristics and function of LSECs. Genetic endothelial GATA4 deletion leads to liver fibrosis and hepatopathy by GATA4-MYC-PDGFB axis (35). In 2015, transcriptional factor KLF2 was found to upregulate eNOS expression, which is essential for maintaining a functional endothelial phenotype and also explains the molecular mechanisms of statins (36). In 2022, de Haan et al (37) found that LSEC-enriched zinc-finger E-box-binding homeobox2 (Zeb2) can preserve the liver angioarchitecture and prevent liver fibrosis.

On the other hand, capillarization is thought to be the repair of damaged LSECs by bone marrow endothelial progenitors that engraft but fail to fully mature through cell autonomous pathways that inhibit the NO-dependent pathway. Maretti-Mira et al (38) identified heparin-binding EGF-like growth factor (HB-EGF) as a signal to maintain HSCs quiescence while the immature LSECs from bone marrow were unable to shed HB-EGF from the cytosolic membrane. These data demonstrate that capillarization is an early event in the fibrotic process and is a perfect target for the treatment of liver fibrosis. The number of fenestrae was shown to be significantly decreased in LSECs from BMP9-knockout mice and addition of BMP9 to the LSECs culture recovers fenestrae and elevates the expression of GATA4 and PLVAP, suggesting that BMP9 is a key paracrine regulator of LSEC fenestration (39). Conversely, Haristi et al found that high BMP9 in combination with LPS stimulation induced the expression of certain capillarization markers in LSEC (40). Desroches-Castan et al (41) found that the role of BMP9 in LSEC differentiation depends on genetic background of C57BL/6, BALB/c and 129/Ola mice. Further studies may warrant the exact roles of BMP9 in LSEC capillarization.

In addition, cell-specific junctional adhesion molecule A (JAM-A) is reported to exert crucial functions in hepatic fibrogenesis (10). Loss of endothelial JAM-A induces LSEC capillarization and HSC activation by triggering Hh signaling and loss of VEGFR1/2 with no alterations in myeloid recruitment (10). Elevated portal blood lipopolysaccharide (LPS) due to leaky gut induces LSEC capillarization in mice and in vitro, while antimicrobial treatment lowers portal blood LPS concentration and inhibits LSEC capillarization (42). Overall, the signaling pathways that regulate LSEC fenestration varies under physiological and pathological conditions and the mechanisms involved by various factors are summarized in Table I.

HSCs are the main source of ECM and thus have been the central to the development of liver fibrosis. The interaction between LSECs and HSCs forms a positive feedback loop. The starting point of the loop is VEGF. As described above, VEGF from hepatocytes and HSCs maintains the phenotype of fenestrated LSECs and, in turn, NO from VEGF-stimulated LSECs prevents HSC activation. Therefore, VEGF may have a dual role in fibrosis. It induces fibrosis-associated angiogenesis while recruiting monocytes mainly through the CXC chemokine (CXCL)9-MMP13 axis to control the secretion of multiple MMPs and tissue inhibitor of MMPs during fibrosis resolution (43). Therefore, as a key mediator of angiocrine signaling, the regulatory role of VEGF in liver diseases remains to be elucidated. Endothelin-1 from HSCs also plays a key role in the regulation of eNOS activation in LSECs (44). In addition, capillarized LSECs activate HSCs through secreting PDGF, TGF-β and lowering a transcription factor KLF2, which acts as a vasoprotective molecule of the liver endothelium, subsequently activating HSCs (45). Activated HSCs further act on quiescent HSCs and LSECs by autocrine TGF-β1 (46). Thrombospondin-1 can be induced by HSC activation, enhancing LSEC capillarization by blocking the NO-dependent pathway (47). In addition, LSECs stimulate HSC migration by producing CXCL12/stromal-derived factor-1. CXCL12 binds to CXC chemokine receptor (CXCR)4 on HSCs, thereby inducing HSC proliferation and collagen production perpetuating fibrosis, while CXCR7-expressing LSECs activated by CXCL12 initiate liver regeneration (48). Enhancement of intracellular adipocyte fatty acid-binding protein (A-FABP) potentiates the LSEC capillarization by inducing Hh signaling, leading to impairment of the gatekeeper function of LSECs on HSC activation. On the other hand, LSEC derived A-FABP releases and acts on HSCs in a paracrine manner to potentiate the transactivation of TGF-β1 by activating JNK/c-Jun signaling. Elevated TGF-β1 subsequently promotes the expression of ECM in both paracrine and autocrine manners. These findings define that A-FABP is a novel therapeutic target of liver fibrosis (49). The contribution of exosomes and their cargoes has been proposed as a new model of intercellular communication between LSECs and HSCs. LSEC-derived sphingosine kinase 1-containing exosomes regulate HSC signaling and migration through fibronectin-integrin-dependent exosome adherence and dynamin-dependent exosome internalization (50).

Loss of fenestration of LSECS protects the liver from ongoing damage by limiting toxins to specific areas (19). However, the formation of a continuous basement membrane lining the sinusoids disrupts the bidirectional exchange of oxygen and nutrients between hepatocytes and sinusoids (51), as well as the lipoprotein secretion of hepatocytes and the de novo lipogenesis in hepatocytes which aggravates hepatic steatosis in NASH (52). LSECs can also secrete Wnt and hepatocyte growth factor (HGF), which together act as key hepatocyte mitogens with a potent liver regenerative capacity that is readily switched to a pro-fibrotic phenotype, in which the Erk1/2/Akt axis in LSECs acts as a switch (53). Growth factors from LSECs, such as bone morphogenetic protein (BMP)2, BMP6 and TGFβ1 act on hepatocytes and HSCs to control systemic iron homeostasis and fibrotic processes in a paracrine manner, respectively (54). Notch activation can alter the angiocrine profile of LSECs to compromise hepatocyte proliferation and liver regeneration through downregulating critical hepatocyte mitogens, including Wnt2a, Wnt9b and HGF (31). A recent study found that lncRNA Airn, a nuclear localized and highly unstable in the form of non-splicing 108 kb ncRNA from LSECs, promoted hepatocyte proliferation by increasing paracrine secretion of Wnt2a and HGF from LSECs (34). lncRNA Airn is inherent and might be a serum biomarker for liver fibrogenesis. Indeed, after acute injury, endothelial cells determine if the liver undergoes regeneration or fibrosis by angiocrine signaling (55). LSECs and hepatocytes also communicate with each other through the VEGF-A/VEGFR-2 signaling which is regarding to angiogenesis. Yan et al (56) investigated the biological roles of cluster of differentiation (CD)147 in a CCl₄-induced liver fibrosis mouse model and found that CD147 promoted liver fibrosis progression via VEGF-A/VRGFR-2 signaling, mediating crosstalk between hepatocytes and LSECs. A combination of leukocyte derived chemokine2 expressed by hepatocytes and Tie1 produced by LSECs can promote LSECs capillarization and reduce portal angiogenesis (57).

Hepatic macrophages are important to the pathogenesis of chronic liver injury. Activated KCs can secrete a wide variety of proinflammatory activating HSCs. KCs and infiltrating macrophages can usually be divided into two subtypes, including M1 macrophages, which produce proinflammatory cytokines such as TNF-α, IL-1b, CCL2 and CCL5, and M2 macrophages which secrete a distinct set of mediators including IL-13, IL-10, IL-4 and TGF-b (58,59). Imbalance of M1/M2 ratio may be the key to the progression of NASH to liver fibrosis (52). Pro-inflammatory KCs may also produce cytokines and chemokines which increase the expression of adhesion molecules by LSECs, leading to leukocyte infiltration and activation (60). In a murine model of acute liver injury caused by overdose of acetaminophen, You et al (61) found that liver resident macrophages and infiltrating macrophages express an array of angiogenic factors and induce LSEC proliferation and migration to repair liver blood vessels, while in NASH, the capillarization of LSECs is necessary for activating KCs in that LSECs injury appeared during the simple steatosis phase and preceded the appearance of activate KCs and HSCs (62). Ford et al (63) investigated the effects of KCs on human liver sinusoidal endothelial cells (hLSECs) phenotype on soft materials, which indicate that TNF-α secreted by KCs undermines the effects of a soft matrix that is representative of healthy tissues and leads to loss of LSEC fenestrae. The relationship between LSECs and macrophages in liver fibrosis needs further investigation.

LSECs belong to nonprofessional antigen presenting cells, plasmacytoid DCs and are primary immunologic mediators underlying hepatic tolerance. They express major histocompatibility complex (MHC) class I, MHC class II and the co-stimulatory molecules CD80, CD86 and CD40 (11). The LSECs mediate hepatic immune tolerance toward self or foreign antigens by expression of anti-inflammatory mediators under physiological conditions, but they gain pro-inflammatory functions upon viral infection (64). LSECs release cytokines via toll-like receptors and contribute to viral infection of the liver, such as hepatitis B or hepatitis C virus (65). The physiologic consequences of higher antigen-presenting cell-related immunogenicity is associated with elevated intrahepatic inflammatory milieu in fibrosis. These LSECs gain enhanced capacity to capture antigens and induce the immunogenic T cell to sensitize endogenous cytotoxic T lymphocytes (66). During liver fibrosis, Th1 and Th2 cells recruit and adhere to liver sinusoids using α4β1-integrin and vascular adhesion protein-1 (VAP-1), respectively (67). However, the specific mechanisms by which T cells and LSECs crosstalk regulate sinusoidal capillarization remain largely unknown. Emerging evidences suggested B-cell activation is involved in NAFLD (68), while crosstalk between LSECs and B-cell remains to be elucidated. In addition, chemokine receptor CXCR6 on LSECs and its ligand CXCL16 on natural killer T cells (NKTs) control NKT cell migration (69).

Inflammation is an essential part involved in the process of liver fibrosis. Hepatitis and NAFLD, especially NASH, have recently received increasing attention (70). LSECs interact with portal circulation via expression of surface receptors and release of paracrine factors, such as intercellular adhesion molecule, vascular cell adhesion molecule-1, VAP-1, chemokines and cytokines (71), recruiting the circulating leucocytes to enter the liver parenchyma. Exploring these inflammatory mediators may help to discover new therapeutic targets for the treatment of liver fibrosis. Gola et al (72) considered LSECs as a microbiome sensor, sustaining commensal-induced MyD88-dependent signaling to regulate the composition of the peri-cellular matrix involved in chemokine gradient formation. In vitro and in vivo findings demonstrate that specific focal adhesion proteins recruited by phosphofructokinase can parlay LSEC mechanotransduction into stiffness-induced angiocrine signaling and mediate neutrophil infiltration and promote liver fibrosis (73). By contrast, Brozat et al found that LSEC-derived JAM-A had no effect on cell migration from sinusoids or adhesion to LSECs (10), suggesting a complex role of leukocyte-endothelial cell interactions in promoting liver fibrosis.

At present, an increasing number of individuals realize that the platelets are no longer bystanders in liver diseases and anti-platelet therapy has been proven effective in treating chronic liver injury (74,75). In vitro studies show that platelet adhesion is partly mediated by integrin (glycoprotein IIb/IIIa and αvβ3) with human hepatic sinusoidal endothelial cells (76), leading to platelet and endothelial activation and leukocyte recruitment. On the other hand, in wild-type mice after chronic CCl₄ challenge plasma Von Willebrand factor levels are increased (77), which could be associated with the interaction between platelets and LSECs. However, it is not clear whether platelet adhesion is associated with capillarization. Platelets have well appreciated roles in physiological and pathological processes including inflammation, tissue repair, angiogenesis and tumor growth (78). A few studies have focused on the release of platelet granule content or platelet RNA transfer to activated HSCs (79,80), but the influence of platelets on LSECs needs more research. Although platelets have been shown to simulate proliferation of cultured hepatocytes (81), this is a long way to elucidating the role of modulators of platelets in chronic liver diseases.

While most researchers focused on intercellular crosstalk of LSECs, Brougham-Cook et al (82) redirected attention to identify the unique phenotypes that LSECs exhibit due to ECM composition, stiffness and soluble factor alterations in fibrotic models by a high-throughput cellular microarray. LSECs play a crucial role in hepatic fibrogenesis by participating in ECM deposition in the form of collagen and fibronectin synthesis and regulating ECM metabolism. Extra domain A from injured LSECs promotes stellate cell motility and parenchymal liver fibrosis (83). Liu et al (84) used the principles of mechanical mechanics to explain capillarization of LSECs, which might contribute to early-stage collagen contraction, then collagen fibers function as mechanical transducer to activate HSCs. Notably, researchers found that capillarized LSECs undergo a partial endothelial-mesenchymal transition characterized by increased ECM production without activating cell mobility, leading to perisinusoidal ECM deposition preferentially in liver sinusoids but not septal/portal scars (85) (Fig. 1)