Liver sinusoidal endothelial cell fenestrations in metabolic liver disease: from molecular mechanisms to therapeutic perspectives Authors: Akram S. Al-Mughalles, Junhu Ma, Hong Ma, Cuiyu Zhang, Huangjin Xiao, Xiaoyuan Ma, Qian Zhang, Chenggang Jia, Hua Han, Bi-sen Ding, Shuya Zhang Published: 2026 ABSTRACT Metabolic dysfunction-associated steatotic liver disease (MASLD) and its progressive inflammatory form, metabolic dysfunction-associated steatohepatitis (MASH), constitute major causes of hepatic morbidity worldwide. Central to the pathophysiology of these conditions is the progressive impairment of liver sinusoidal endothelial cells (LSECs), whose characteristic fenestrated architecture is indispensable for hepatic homeostasis. Capillarization, defined as the loss of LSEC fenestrations, propels disease progression by disrupting the equilibrium between pro-fenestration and pro-capillarization signaling cascades. This review critically appraises the LSEC Fenestration Signaling Nexus as a central, targetable mechanism in metabolic liver disease and assesses emerging therapeutic strategies engineered to restore fenestrations. We present a translational roadmap that prioritizes combinatorial pharmacological approaches and LSEC-specific biomarkers for precision medicine. This integrative framework furnishes new perspectives for rational drug development and individualized therapeutic strategies in metabolic liver disease.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12964-026-02997-8. Introduction Metabolic dysfunction-associated steatotic liver disease (MASLD) is projected to become the foremost indication for liver transplantation by 2030. This escalating prevalence across all geographic regions poses a formidable clinical challenge [1–6]. At the mechanistic core of this disease spectrum lies a critical cell type: the liver sinusoidal endothelial cells (LSECs), which sustain hepatic homeostasis through a specialized vascular interface [7–9]. LSECs are unique cells, forming a highly porous endothelial barrier perforated by dynamic transcellular pores termed fenestrations (50 to 300 nm in diameter). These fenestrations assemble into specialized structures known as sieve plates. Rather than constituting a static barrier, this “liver sieve” functions as a dynamic regulatory interface that selectively mediates molecular exchange between the hepatic parenchyma and sinusoidal circulation [10, 11] (Fig. 1).Fig. 1Ultrastructural architecture and functional zonation of LSECs. A, B Representative scanning electron microscopy (SEM) images from original experimental data depicting (A) fenestrated and (B) capillarized LSECs, scale bars = 1um. Fenestrated LSECs display characteristic sieve plates, whereas capillarized LSECs exhibit loss of fenestrae. C Schematic representation of the liver sinusoid illustrating the ultrastructural and functional zonation of LSEC fenestrations along the porto-central axis, with key cellular interactions highlighted. Created with BioRender (https://BioRender.com) Ultrastructural architecture and functional zonation of LSECs. A, B Representative scanning electron microscopy (SEM) images from original experimental data depicting (A) fenestrated and (B) capillarized LSECs, scale bars = 1um. Fenestrated LSECs display characteristic sieve plates, whereas capillarized LSECs exhibit loss of fenestrae. C Schematic representation of the liver sinusoid illustrating the ultrastructural and functional zonation of LSEC fenestrations along the porto-central axis, with key cellular interactions highlighted. Created with BioRender (https://BioRender.com) The early loss of fenestrations, a process designated as capillarization, constitutes a primary pathogenic event in MASLD. This structural transformation triggers a pathological cascade that promotes steatosis, steatohepatitis, systemic metabolic dysfunction, and hepatocellular carcinoma (HCC) [12–14]. Historically, capillarization was regarded as passive scarring; however, current evidence demonstrates that it actively drives disease progression [15]. This review introduces the LSEC Fenestration Signaling Nexus as a unifying conceptual model. The phenotypic fate of LSECs is governed by the balance between two opposing signaling arms: the pro-fenestration vascular endothelial growth factor (VEGF)-nitric oxide (NO) pathway and the pro-capillarization Ras homolog family member A (RhoA)-Rho-associated coiled-coil-containing protein kinase (ROCK) pathway. When this balance is perturbed, disease progression accelerates. We evaluate therapeutic strategies designed to restore fenestration through Nexus rebalancing and outline a roadmap for precision medicine in MASLD management. The LSEC fenestration signaling Nexus: a multi-layered regulatory system A multi-layered regulatory system governs the integrity of LSEC fenestrations. Within the LSEC Fenestration Signaling Nexus, molecular pathways, environmental cues, and intercellular interactions converge in a coordinated hierarchy to maintain the delicate balance between fenestration and capillarization [16, 17] (Fig. 2).Fig. 2The LSEC Fenestration Signaling Nexus: an integrated signaling network governing the fenestration-capillarization equilibrium. A Homeostatic Nexus maintenance. The balance between pro-fenestration and pro-capillarization pathways determines the LSEC phenotypic state. VEGF activation of VEGFR2 initiates the PI3K/AKT/eNOS cascade, generating nitric oxide (NO) that sustains fenestration integrity through Rac1 activation and suppression of the RhoA-ROCK-MLC-P axis, thereby preventing excessive actomyosin contraction. Simultaneously, β-catenin stabilization, modulated by VEGFR2 and c-Myc signaling, drives transcriptional programs essential for fenestration maintenance through TCF/LEF-mediated gene expression. B Metabolic stress-induced Nexus collapse. Cholesterol and free fatty acid (FFA) overload induce lipid raft expansion and disruption, leading to sequestration and functional impairment of pro-fenestrative receptors including VEGFR2 and Frizzled. This consequently attenuates PI3K/AKT/eNOS/NO signaling and β-catenin/TCF/LEF transcriptional activity. Concurrently, metabolic injury activates multiple capillarization drivers: CD36- and FABP4-mediated lipid uptake promotes inflammatory signaling and insulin resistance, while pathological agonists including endothelin-1 (ET-1) and lysophosphatidic acid (LPA) drive sustained activation of the RhoA-ROCK-MLC-P cascade. Combined with elevated intracellular Ca.2⁺, this signaling convergence results in irreversible actin hyper-polymerization and cytoskeletal stabilization, culminating in the capillarized phenotype. Created with BioRender (https://BioRender.com) The LSEC Fenestration Signaling Nexus: an integrated signaling network governing the fenestration-capillarization equilibrium. A Homeostatic Nexus maintenance. The balance between pro-fenestration and pro-capillarization pathways determines the LSEC phenotypic state. VEGF activation of VEGFR2 initiates the PI3K/AKT/eNOS cascade, generating nitric oxide (NO) that sustains fenestration integrity through Rac1 activation and suppression of the RhoA-ROCK-MLC-P axis, thereby preventing excessive actomyosin contraction. Simultaneously, β-catenin stabilization, modulated by VEGFR2 and c-Myc signaling, drives transcriptional programs essential for fenestration maintenance through TCF/LEF-mediated gene expression. B Metabolic stress-induced Nexus collapse. Cholesterol and free fatty acid (FFA) overload induce lipid raft expansion and disruption, leading to sequestration and functional impairment of pro-fenestrative receptors including VEGFR2 and Frizzled. This consequently attenuates PI3K/AKT/eNOS/NO signaling and β-catenin/TCF/LEF transcriptional activity. Concurrently, metabolic injury activates multiple capillarization drivers: CD36- and FABP4-mediated lipid uptake promotes inflammatory signaling and insulin resistance, while pathological agonists including endothelin-1 (ET-1) and lysophosphatidic acid (LPA) drive sustained activation of the RhoA-ROCK-MLC-P cascade. Combined with elevated intracellular Ca.2⁺, this signaling convergence results in irreversible actin hyper-polymerization and cytoskeletal stabilization, culminating in the capillarized phenotype. Created with BioRender (https://BioRender.com) Core signaling equilibrium At the center of the Nexus lies a fundamental molecular equilibrium. The VEGF-VEGFR2-eNOS-NO axis functions as the primary pro-fenestration pathway, stabilizing the actin cytoskeleton [18–21] (Fig. 2A). Conversely, the RhoA/ROCK pathway promotes actin polymerization and contractility, thereby driving capillarization [22–27]. A state of dynamic antagonism prevails, in which NO exerts tonic inhibition on RhoA/ROCK activity [28, 29]. The actin cytoskeleton serves as the central effector integrating both pathways, and this balance is inherently vulnerable to metabolic stressors. Notably, the ROCK2 isoform, rather than ROCK1, is selectively upregulated in human LSECs in fibrotic livers, and selective ROCK2 inhibition has been shown to restore fenestrations in preclinical models [30]. For instance, cholesterol overload disrupts lipid raft microdomains, impairing VEGF signaling [31–33]. Concurrently, intracellular lipid accumulation activates peroxisome proliferator-activated receptor (PPAR) signaling and upregulates pro-capillarization mediators such as fatty acid-binding protein 4 (FABP4) and CD36 (Fig. 2B) [34, 35], which further destabilize the Nexus through inflammatory feedback loops [36, 37]. A paramount unanswered question concerns the identification of the irreversible tipping point in this system, specifically the transition from a dynamic, reversible state to a stable, pathological capillarization. Cellular crosstalk: the Nexus in its microenvironmental context Intercellular communications are essential for the physiological regulation of the hepatic sinusoid [38] Under homeostatic conditions, a reciprocal signaling loop between fenestrated LSECs [17, 39, 40], hepatocytes, quiescent hepatic stellate cells (HSCs), and Kupffer cells (KCs)maintains metabolic exchange and an anti-inflammatory microenvironment (Fig. 3A) [41–43]. LSECs support hepatocyte function through angiocrine factors such as HGF and, in return, receive maintenance signals such as VEGF. This reciprocal signaling preserves fenestration integrity. Furthermore, quiescent HSCs and KCs contribute additional support to fenestration maintenance [41, 42]. However, upon capillarization, this crosstalk shifts decisively toward a pro-fibrotic, pro-inflammatory state. The paracrine action of capillarized LSECs promotes fibrogenic reprogramming through secretion of key factors such as transforming growth factor-beta (TGF-β), platelet-derived growth factor (PDGF), and FABP4, which activate the pro-fibrotic phenotype of hepatic stellate cells (HSCs) and result in collagen deposition, thereby establishing a self-sustaining fibrogenic loop. Simultaneously, LSECs acquire a pro-inflammatory phenotype, upregulating adhesion molecules (e.g., intercellular adhesion molecule-1 [ICAM-1] and vascular cell adhesion molecule-1 [VCAM-1]) and chemoattractants that recruit circulating immune cells, thereby amplifying hepatic injury [9, 44, 45]. Additionally, scavenger receptors such as stabilin-2 facilitate leukocyte adhesion and infiltration [9, 44, 45]. This positions the capillarized LSEC as a central hub coordinating pathological remodeling (Fig. 3B) [7, 46]. Additional details are provided in the supplementary file and Table S1. pro-fibrotic, pro-inflammatory state.Fig. 3Cellular crosstalk in the hepatic sinusoid under physiological and pathological conditions. A Homeostatic intercellular communications in the healthy liver. Fenestrated LSECs maintain a sophisticated signaling network that sustains hepatic function. They receive hepatocyte-derived VEGF and Hedgehog signals that preserve fenestrae porosity through eNOS/NO-dependent mechanisms, while reciprocally secreting WNT2 and HGF to regulate hepatocyte metabolic function. LSEC-derived NO preserves the quiescent phenotype of HSCs, which in turn secrete VEGF and BMP9 to support endothelial homeostasis. Simultaneously, LSECs establish immunosuppression by inhibiting KC activation through MYD88-dependent pathways and blocking neutrophil recruitment. This coordinated signaling maintains an anti-inflammatory milieu facilitated by IL-10 and PGE2 release. The intact sinusoidal architecture permits bidirectional transport across the Space of Disse, with non-adherent blood cells (red blood cells, platelets, lymphocytes, neutrophils, B cells) circulating unimpeded. B Pathological remodeling in capillarized LSEC. Metabolic stress triggers fundamental reorganization of sinusoidal communication. Capillarized LSECs undergo TGF-β/PDGFB-mediated HSC activation, resulting in collagen I/IV deposition that obliterates the Space of Disse. FABP4 upregulation drives monocyte infiltration via CXCL12 chemotaxis and enhances adhesion of neutrophils, lymphocytes, and natural killer (NK) cells through ICAM-1 and VAP1 overexpression. Activated KCs release pro-inflammatory mediators (TNF-α, PAF, PDGF) that generate vWF-dependent blood clots, while capillarized LSECs secrete Ang-2 and LECT2, promoting Tie1 receptor dysfunction. Activated blood cells (platelets, neutrophils, lymphocytes, and NK cells) adhere to capillarized LSECs, obstructing sinusoidal flow and abolishing bidirectional transport, thereby creating a pro-fibrotic microenvironment. Created with BioRender (https://BioRender.com) Cellular crosstalk in the hepatic sinusoid under physiological and pathological conditions. A Homeostatic intercellular communications in the healthy liver. Fenestrated LSECs maintain a sophisticated signaling network that sustains hepatic function. They receive hepatocyte-derived VEGF and Hedgehog signals that preserve fenestrae porosity through eNOS/NO-dependent mechanisms, while reciprocally secreting WNT2 and HGF to regulate hepatocyte metabolic function. LSEC-derived NO preserves the quiescent phenotype of HSCs, which in turn secrete VEGF and BMP9 to support endothelial homeostasis. Simultaneously, LSECs establish immunosuppression by inhibiting KC activation through MYD88-dependent pathways and blocking neutrophil recruitment. This coordinated signaling maintains an anti-inflammatory milieu facilitated by IL-10 and PGE2 release. The intact sinusoidal architecture permits bidirectional transport across the Space of Disse, with non-adherent blood cells (red blood cells, platelets, lymphocytes, neutrophils, B cells) circulating unimpeded. B Pathological remodeling in capillarized LSEC. Metabolic stress triggers fundamental reorganization of sinusoidal communication. Capillarized LSECs undergo TGF-β/PDGFB-mediated HSC activation, resulting in collagen I/IV deposition that obliterates the Space of Disse. FABP4 upregulation drives monocyte infiltration via CXCL12 chemotaxis and enhances adhesion of neutrophils, lymphocytes, and natural killer (NK) cells through ICAM-1 and VAP1 overexpression. Activated KCs release pro-inflammatory mediators (TNF-α, PAF, PDGF) that generate vWF-dependent blood clots, while capillarized LSECs secrete Ang-2 and LECT2, promoting Tie1 receptor dysfunction. Activated blood cells (platelets, neutrophils, lymphocytes, and NK cells) adhere to capillarized LSECs, obstructing sinusoidal flow and abolishing bidirectional transport, thereby creating a pro-fibrotic microenvironment. Created with BioRender (https://BioRender.com) Nexus collapse in metabolic liver disease The LSEC fenestration system constitutes a highly efficient biological sieve that facilitates selective bidirectional transport while excluding cellular components through three key structural features: sieve plates, absence of a basal lamina, and a collagen-free Space of Disse [47] (Fig. 1A). Under physiological conditions, fenestrated LSECs enable unimpeded substrate exchange; solute transfer depends primarily on flow rate rather than membrane permeability [31]. Capillarization, by contrast, represents a profound structural and functional disruption characterized by three cardinal changes (Fig. 1B): (1) loss or severe reduction of fenestrations (the transcellular pores themselves), (2) deposition of a thickened basement membrane and collagen IV accumulation that obliterates the Space of Disse, and (3) endothelial cell thickening with actin stress fiber stabilization. Together, these changes reduce substrate flux by more than 50%0 [31]. Moreover, capillarized LSECs undergo a fundamental phenotypic shift, adopting pro-fibrogenic characteristics through increased secretion of TGF-β and PDGF, which consequently initiates self-perpetuating cycles of hepatic injury [35, 48]. MASLD and Metabolic and alcohol-associated liver disease (MetALD) In human and rodent studies, [13, 49, 50] early fenestration loss impairs two critical hepatic functions. First, it reduces the clearance of triglyceride-rich lipoproteins. Second, it disrupts insulin transcytosis. Together, these defects promote lipid accumulation and hepatic insulin resistance [15, 51]. The resulting lipid overload within LSECs instigates metabolic reprogramming, activating PPAR signaling and FABP4 expression [34]. This FABP4 expression further destabilizes the Nexus, establishing a vicious cycle of endothelial dysfunction [52, 53]. As disease progresses to MASH, capillarized LSECs amplify inflammation by secreting C–C motif chemokine ligand 2 (CCL2) and upregulating adhesion molecules such as ICAM-1, consequently recruiting leukocytes and exacerbating hepatic injury [9, 52, 54]. Both MASLD and alcoholic liver disease (ALD) converge on a shared pathological endpoint [55]. LSECs metabolize ethanol directly via cytochrome P450 2E1 (CYP2E1), generating oxidative stress that potently activates the pro-capillarization RhoA/ROCK pathway [56–58]. Ethanol simultaneously disrupts the gut-liver axis, compounding direct hepatic injury through promotion of endotoxemia [59]. This mechanistic convergence is clinically embodied in metabolic and alcohol-associated liver disease (MetALD), where metabolic and alcoholic insults synergistically disrupt the Nexus, frequently resulting in accelerated fibrosis progression [55]. Elevated serum markers of LSEC dysfunction (e.g., VCAM-1 and CD146) underscore the clinical relevance of this shared pathway and predict fibrosis severity across both etiologies [60]. The profound therapeutic implication is that strategies directed at restoring the Nexus hold promise across the etiological spectrum of steatohepatitis. The limited efficacy of monotherapies further supports the development of LSEC-centric therapies capable of mitigating injury across multiple etiologies [61, 62]. Systemic metabolic dysregulation: LSEC capillarization as a central driver The LSEC Fenestration Signaling Nexus serves not only as a governor of intrahepatic health but also as a critical interface between the liver and systemic metabolic homeostasis [52, 63]. In the healthy state, the fenestrated LSEC phenotype facilitates efficient nutrient exchange [64]. However, chronic metabolic insults trigger Nexus dysregulation and capillarization, which in turn initiates and amplifies systemic metabolic disorders through several interconnected pathways [65, 66] The initial consequences of capillarization are concentrated within the liver yet carry immediate systemic implications. Capillarized LSECs exacerbate hepatic insulin resistance by impairing insulin transcytosis and secrete pro-inflammatory cytokines (IL-6, TNF-α) that disrupt systemic glucose homeostasis [67, 68]. Furthermore, capillarized LSECs activate HSCs via PDGF/TGF-β pathways, promoting fibrosis while impairing lipoprotein clearance, which contributes to dyslipidemia [69] (Fig. 4).Fig. 4LSEC-driven systemic metabolic dysregulation. This schematic depicts how LSEC dysfunction serves as a nexus linking local hepatic pathology to systemic metabolic disease. Capillarization and LSEC inflammation, driven by factors such as LPS (from gut dysbiosis), insulin resistance, and hyperlipidemia, lead to elevated release of vasoactive and inflammatory mediators (e.g., FABP4, Ang-2, CXCL10). These factors enter the systemic circulation, contributing to peripheral tissue insulin resistance, impaired pancreatic β-cell function, and adverse effects on cardiovascular health (CVD) and PCOS. Created with BioRender (https://BioRender.com) LSEC-driven systemic metabolic dysregulation. This schematic depicts how LSEC dysfunction serves as a nexus linking local hepatic pathology to systemic metabolic disease. Capillarization and LSEC inflammation, driven by factors such as LPS (from gut dysbiosis), insulin resistance, and hyperlipidemia, lead to elevated release of vasoactive and inflammatory mediators (e.g., FABP4, Ang-2, CXCL10). These factors enter the systemic circulation, contributing to peripheral tissue insulin resistance, impaired pancreatic β-cell function, and adverse effects on cardiovascular health (CVD) and PCOS. Created with BioRender (https://BioRender.com) This intrahepatic dysfunction is powerfully reinforced by inter-organ crosstalk, primarily through the gut-liver and liver-adipose axes. The gut-liver axis generates a deleterious feedback loop: increased intestinal permeability promotes endotoxin (lipopolysaccharide [LPS]) translocation, which activates KCs to release inflammatory mediators that suppress the pro-fenestration VEGF signaling pathway [70, 71]. Concurrently, gut-derived oxidative stress directly activates the pro-capillarization RhoA/ROCK pathway [72]. Simultaneously, the liver-adipose axis establishes a self-perpetuating cycle. Adipose-derived mediators, including leptin, IL-6, and notably FABP4, accelerate LSEC capillarization. In return, the dysfunctional liver promotes atherosclerosis through very-low-density lipoprotein (VLDL) overproduction and contributes to pancreatic β-cell dysfunction via signaling molecules such as angiopoietin-2 (ANGPT2) and C-X-C motif chemokine ligand 10 (CXCL10) [72]. Notably, emerging evidence now extends the role of LSEC dysfunction to endocrine pathologies, such as polycystic ovary syndrome (PCOS) [73]. Capillarized LSECs contribute to PCOS through both direct and indirect mechanisms. FABP4 secretion directly correlates with PCOS clinical features [74], and this protein also induces systemic insulin resistance and pancreatic β-cell dysfunction, which together promote hyperandrogenism [75, 76]. This positions LSEC capillarization as a previously unrecognized link between hepatic vascular pathology and reproductive endocrine dysfunction [77]. Therapeutically, this integrated perspective highlights promising targets within the bile acid-FXR/FGF21 and GLP-1/GLP-1R pathways, which are modulated by gut microbiota and demonstrate potential for simultaneously addressing LSEC dysfunction and its systemic metabolic consequences [78, 79]. In summary, the capillarized LSEC is not merely a consequence of metabolic disease but a central hub that propagates injury across multiple organ systems. Consequently, therapeutic strategies that target and restore the LSEC Nexus represent a promising, unified approach to mitigate the interconnected pathologies of metabolic syndrome, diabetes, and their extrahepatic complications. Therapeutic targeting: strategies to rebalance the Nexus The LSEC Fenestration Signaling Nexus provides a structured and mechanistically grounded framework for therapeutic development in metabolic liver disease. Effective interventions must concurrently inhibit pro-capillarization pathways (e.g., RhoA/ROCK) and reactivate pro-fenestration signaling (e.g., VEGF-NO), advancing beyond single-target agents toward multi-modal strategies. This integrated approach recognizes that capillarization constitutes a stabilized pathological state unlikely to be reversed by isolated pharmacological manipulation. Effective therapies must combine two approaches: (1) direct molecular and cellular interventions, and (2) foundational lifestyle modifications encompassing nutrition and physical activity [80]. Such lifestyle measures contribute to LSEC preservation through systemic metabolic improvement, anti-inflammatory effects, and enhanced redox homeostasis, thereby serving as essential adjuncts to pharmacotherapy [81]. The overarching therapeutic objective is restoration of the fenestrated LSEC phenotype, which in preclinical studies reverses metabolic liver disease [82]. The following sections evaluate a spectrum of emerging strategies to rebalance the Nexus, from molecular and nanotechnological approaches to regenerative medicine, alongside the supportive role of lifestyle intervention. Lifestyle interventions as a foundational LSEC therapy Lifestyle modifications constitute a foundational therapeutic strategy for metabolic liver disease by directly targeting the LSEC Fenestration Signaling Nexus [83, 84]. Nutritional interventions trigger precise molecular changes. Caloric restriction enhances hepatic perfusion by activating Kruppel-like factor 2 (KLF2) to increase fenestration density, while Mediterranean diets rich in omega-3 fatty acids remodel LSEC lipid rafts and restore VEGFR2-eNOS coupling [85–87]. In parallel, physical activity induces beneficial hemodynamic and biochemical adaptations; aerobic exercise elevates sinusoidal shear stress to stabilize the LSEC phenotype via mechanotransduction, and resistance training reduces visceral adipose-derived FABP4 release, thereby mitigating LSEC lipotoxicity [88–90]. Collectively, these interventions create a “Nexus-primed” hepatic microenvironment that not only preserves endothelial function but also critically synergizes with pharmacotherapy. Clinical studies demonstrate this synergy: weight loss potentiates farnesoid X receptor (FXR) agonist efficacy, leading to significantly improved outcomes. To definitively establish LSEC fenestration as a therapeutic endpoint, future prospective studies must incorporate direct sinusoidal imaging and dynamic serum biomarker assessment, such as FABP4 measurement, to validate the LSEC-specific efficacy of lifestyle intervention. The therapeutic landscape for MASLD, MASH, and MetALD converges on a shared pathogenic mechanism: LSEC capillarization. Whether driven by metabolic insults, ethanol exposure, or their synergistic combination in MetALD, restoring the fenestrated LSEC phenotype represents a unified therapeutic objective. The following molecular strategies target this convergent endpoint across etiologies. Pharmacological therapies: molecular strategies Direct Nexus modulation (core pathways) The LSEC Fenestration Signaling Nexus, with its defined molecular constituents and hierarchical organization, provides a rational blueprint for therapeutic intervention. The most promising strategies are those that directly engage this core machinery, evolving from single-pathway agents to systemic multi-node modulators that collectively aim to fenestration restoration. Pro-fenestration reactivation (VEGF, No donors) The most direct therapeutic strategy involves reactivating the pro-fenestration arm by targeting its primary molecular switch. For instance, administering recombinant VEGF-A165 shows preclinical promise in reducing portal pressure and fibrosis [91, 92]. However, a critical limitation immediately emerges: systemic VEGF delivery carries significant risks of hemorrhage and tumorigenesis, consequently creating a therapeutic imperative for targeted delivery systems [93, 94]. In contrast to the complexities of growth factor therapy, complementary strategies centered on nitric oxide (NO) signaling offer a more direct route to stabilize the LSEC cytoskeleton. Direct NO donors aim to bypass upstream signaling defects and reverse capillarization [95]. Nevertheless, systemic hemodynamic instability constrains the clinical utility of NO donors. Consequently, the field is advancing toward more sophisticated, liver-specific solutions such as the NO prodrug V-PYRRO/NO, which is engineered for selective activation by hepatic cytochrome P450 enzymes to maximize local therapeutic effects [96]. Furthermore, a deeper understanding of the Nexus reveals that NO bioavailability is frequently compromised by oxidative stress. Therefore, a synergistic approach involves combining NO donors with antioxidants such as recombinant manganese superoxide dismutase (MnSOD), which scavenges mitochondrial superoxide to protect NO from inactivation and directly counter the oxidative drivers of capillarization [97]. Clinical studies corroborate the translational relevance of enhancing NO pathways. Ascorbic acid (vitamin C) improves intrahepatic endothelial dysfunction in cirrhotic patients, underscoring the therapeutic potential of this axis [95, 98, 99]. Summarized in (Table 1).Table 1Pro-fenestration therapeutic strategies for LSEC fenestration restorationPro-fenestration AgentMechanism of ActionPrimary Target PathwayPreclinical EvidenceClinical StatusKey AdvantagesMajor Limitations/Safety ConcernsDelivery StrategyReferencesRecombinant VEGF-A165Direct activation of VEGFR2-PI3K-AKT-eNOS-NO signalingVEGFR2-eNOS-NO pathwayReduces portal pressure and fibrosis in rodent modelsPreclinicalDirectly targets primary pro-fenestration switch; potent angiogenic effectSystemic VEGF carries hemorrhage and tumorigenesis risks; off-target vascular effectsTargeted delivery systems (nanoparticles, prodrugs) [91, 92, 100, 101]NO Donors (direct)Bypass upstream signaling defects; direct eNOS pathway activationVEGFR2-eNOS-NO pathwayReverses capillarization in LSEC modelsPreclinical/Early translationalCircumvents VEGF-specific delivery challengesSystemic hemodynamic instability; limited bioavailabilityLiver-selective formulations [95, 102–104]V-PYRRO/NO (NO prodrug)P450-activated hepatic NO release; selective liver activationVEGFR2-eNOS-NO pathwayRestores fenestrations with reduced systemic effectsPreclinical/Early developmentHepatic selectivity; improved bioavailability vs. direct donorsProdrug activation efficiency; variable P450 expressionHepatic P450-mediated activation [96]Manganese Superoxide Dismutase (MnSOD)Scavenges mitochondrial superoxide; protects NO from oxidative inactivationAntioxidant/NO bioavailability axisReduces oxidative stress; synergizes with NO donorsPreclinicalAddresses dual pathology (fenestration loss + oxidative stress); synergistic with NOProtein degradation; mitochondrial uptake challengesMitochondrial targeting; nanoparticle encapsulation [97, 105]Ascorbic Acid (Vitamin C)Antioxidant; enhances NO bioavailability; improves eNOS couplingNO bioavailability/antioxidantImproves intrahepatic endothelial dysfunction in cirrhotic patientsClinical (proven)Safe, well-tolerated; clinically validated in cirrhosis; oral availabilityLimited intrinsic anti-fibrotic activity; requires combination therapyOral or IV supplementation [95, 98, 99, 106] Table 1Pro-fenestration therapeutic strategies for LSEC fenestration restorationPro-fenestration AgentMechanism of ActionPrimary Target PathwayPreclinical EvidenceClinical StatusKey AdvantagesMajor Limitations/Safety ConcernsDelivery StrategyReferencesRecombinant VEGF-A165Direct activation of VEGFR2-PI3K-AKT-eNOS-NO signalingVEGFR2-eNOS-NO pathwayReduces portal pressure and fibrosis in rodent modelsPreclinicalDirectly targets primary pro-fenestration switch; potent angiogenic effectSystemic VEGF carries hemorrhage and tumorigenesis risks; off-target vascular effectsTargeted delivery systems (nanoparticles, prodrugs) [91, 92, 100, 101]NO Donors (direct)Bypass upstream signaling defects; direct eNOS pathway activationVEGFR2-eNOS-NO pathwayReverses capillarization in LSEC modelsPreclinical/Early translationalCircumvents VEGF-specific delivery challengesSystemic hemodynamic instability; limited bioavailabilityLiver-selective formulations [95, 102–104]V-PYRRO/NO (NO prodrug)P450-activated hepatic NO release; selective liver activationVEGFR2-eNOS-NO pathwayRestores fenestrations with reduced systemic effectsPreclinical/Early developmentHepatic selectivity; improved bioavailability vs. direct donorsProdrug activation efficiency; variable P450 expressionHepatic P450-mediated activation [96]Manganese Superoxide Dismutase (MnSOD)Scavenges mitochondrial superoxide; protects NO from oxidative inactivationAntioxidant/NO bioavailability axisReduces oxidative stress; synergizes with NO donorsPreclinicalAddresses dual pathology (fenestration loss + oxidative stress); synergistic with NOProtein degradation; mitochondrial uptake challengesMitochondrial targeting; nanoparticle encapsulation [97, 105]Ascorbic Acid (Vitamin C)Antioxidant; enhances NO bioavailability; improves eNOS couplingNO bioavailability/antioxidantImproves intrahepatic endothelial dysfunction in cirrhotic patientsClinical (proven)Safe, well-tolerated; clinically validated in cirrhosis; oral availabilityLimited intrinsic anti-fibrotic activity; requires combination therapyOral or IV supplementation [95, 98, 99, 106] Pro-fenestration therapeutic strategies for LSEC fenestration restoration Pro-capillarization inhibition (ROCK inhibitors) Concurrently, direct inhibition of the pro-capillarization arm is equally critical. In preclinical models, pharmacological ROCK inhibitors (e.g., Y-27632) reduce actomyosin contractility to increase fenestration porosity in rodent LSECs [107] (Fig. 2B). More recently, the highly selective ROCK2 inhibitor TDI01 restored LSEC fenestrations in a minipig MASH model, normalized key fenestration genes (NOS3, PPARG, KLF2) while suppressing capillarization markers (MMP2, COL1A1), and demonstrated safety and a trend toward reduced fibrosis with improved histology after 24 weeks in a phase 1 trial (ChiCTR2200058868) and a six‑patient extended study (ChiCTR2400082056), providing the first clinical proof‑of‑concept that ROCK2 inhibition can restore LSEC fenestrations in humans [30]. Pleiotropic modulators (statins, lanifibranor) Building on these targeted interventions, pleiotropic modulators offer the advantage of concurrently engaging multiple nodes. Statins (atorvastatin, simvastatin) demonstrate mechanistically relevant effects on LSEC biology: endothelial nitric oxide synthase (eNOS) upregulation, RhoA/ROCK inhibition, and improved NO bioavailability [108–111]. These properties theoretically align with Nexus restoration. However, critical limitations must be acknowledged. While statins provide robust cardiovascular risk reduction in MASLD patients, making them indicated for comorbid cardiovascular disease (CVD) prevention, their direct impact on liver histology is inconsistent and unsupported by dedicated Phase III trials for MASH resolution. Preclinical models provide the only evidence for LSEC fenestration restoration; human validation remains lacking [112]. Consequently, current guidelines from the European Association for the Study of the Liver (EASL) and the American Association for the Study of Liver Diseases (AASLD) recommend statins specifically for cardiovascular risk management in MASLD, not for liver disease modification [113, 114]. Their inclusion here reflects the mechanistic relevance of their pleiotropic effects to the Nexus model, not a therapeutic endorsement for MASLD/MASH. Ultimately, the pinnacle of direct modulation is represented by agents capable of systemic Nexus recalibration. The pan-PPAR agonist lanifibranor epitomizes a system-resetting approach. By concurrently activating PPARα, PPARγ, and PPARδ, it orchestrates a multi-organ response: enhancing hepatic fatty acid oxidation, promoting adipose tissue lipid storage and insulin sensitization, and exerting broad anti-inflammatory and anti-fibrotic effects [115]. This holistic modulation addresses the intertwined metabolic, inflammatory, and fibrotic axes of MASLD. While the Phase II NATIVE trial demonstrated efficacy at 24 weeks, achieving MASH resolution without fibrosis worsening in 49% of patients versus 21% on placebo, and fibrosis improvement of at least one stage in 48% versus 29% on placebo, its clinical adoption requires a clear-eyed assessment of its limitations [116]. Key limitations include: (i) class-associated adverse events, such as weight gain (24% vs. 11% placebo) and peripheral edema (19% vs. 9% placebo), which may affect long-term adherence and pose potential cardiovascular risks in a metabolically vulnerable population; (ii) response heterogeneity, where robust histological response may be linked to factors such as patatin-like phospholipase domain-containing 3 (PNPLA3) genotype or baseline metabolic severity; and (iii) the unknown sustainability of mechanistic benefits, including the reversal of LSEC capillarization and reduction in portal pressure, beyond the initial treatment period [117, 118]. The absence of long-term data leaves critical questions regarding the durability of histological improvement and the consequences of sustained pan-nuclear receptor activation over decades. The ongoing Phase III NATIV3 trial (NCT04849728), designed for 72 weeks of treatment, is essential to validate these long-term efficacy and safety concerns [119]. Summarized in (Table 2).Table 2Comparison of anti-MASH agents across 12 key efficacy and safetyParameterLanifibranorOCASemaglutideResmetiromTirzepatideTDI01MechanismPan-PPAR agonist (α/γ/δ)FXR agonistGLP-1 RATHR-β agonistGLP-1/GIP RASelective ROCK2 inhibitorClinical StagePhase 2b (NATIVE)Phase 3 (REGENERATE)Phase 3FDA-Approved (2024)Phase 3Phase 1 + extended (n = 6 fibrosis)MASH Resolution w/o Fibrosis Worsening49% (1200 mg) vs 22% PBO6.5% (25 mg) vs 3.5% PBO (not statistically significant)Up to 62.9%55%Up to 60%Not yet assessed (mixed etiology)Fibrosis ≥ 1 Stage Improvement w/o MASH Worsening48% (1200 mg) vs 29% PBO22.4% (25 mg) vs 9.6% PBOLimited efficacy26.4%Superior to OCATrend in 5/6 patients (biopsy)Combined MASH Resolution + Fibrosis Improvement35% (1200 mg) vs 9% PBO ~ 6–7% vs placeboVariable ~ 21%Highest ratesNot applicableWeight Loss + 2.4–2.7 kg (increase) ~ 0.5 kg lossUp to 15% loss ~ 2–3% lossUp to 20% lossNot reportedHbA1c Reduction (T2DM patients) > 0.5% (T2DM subset)Minimal1.5–1.8% ~ 0.5–0.8%2.0–2.5%Not reportedPruritus Risk < 5% overall16.55% (RR 1.75‡)RareRareRareNone reportedLipid ChangesFavorable improvementsDyslipidemia risk (↑ TC, LDL)FavorableFavorableFavorableNot reportedGI Adverse EventsDiarrhea, nausea (common)Constipation riskNausea, vomiting (common)Nausea, diarrhea (common)GI events (manageable)MinimalDiscontinuation Rate < 5% for AEs ~ 1–10% (due to pruritus) ~ 10–15%Low (~ 2%) ~ 5–10%0% (no SAEs)Study Duration24 weeks18–48 months24 weeks-6 months52–96 weeks24 weeks24 weeksTrial Population247 (76% F2/F3)931–1,607 (F2/F3)2,000 + 1,500 + 1,200 + 6 patients (F2-F4 fibrosis)References [30, 116, 120–127]Legend: PBO Placebo, RA Receptor Agonist, THR-β Thyroid Hormone Receptor-beta, GLP-1 Glucagon-Like Peptide-1, GIP Glucose-dependent Insulinotropic Polypeptide, AE Adverse Events, w/o without, TC Total Cholesterol, LDL Low-Density Lipoprotein, GI Gastrointestinal Table 2Comparison of anti-MASH agents across 12 key efficacy and safetyParameterLanifibranorOCASemaglutideResmetiromTirzepatideTDI01MechanismPan-PPAR agonist (α/γ/δ)FXR agonistGLP-1 RATHR-β agonistGLP-1/GIP RASelective ROCK2 inhibitorClinical StagePhase 2b (NATIVE)Phase 3 (REGENERATE)Phase 3FDA-Approved (2024)Phase 3Phase 1 + extended (n = 6 fibrosis)MASH Resolution w/o Fibrosis Worsening49% (1200 mg) vs 22% PBO6.5% (25 mg) vs 3.5% PBO (not statistically significant)Up to 62.9%55%Up to 60%Not yet assessed (mixed etiology)Fibrosis ≥ 1 Stage Improvement w/o MASH Worsening48% (1200 mg) vs 29% PBO22.4% (25 mg) vs 9.6% PBOLimited efficacy26.4%Superior to OCATrend in 5/6 patients (biopsy)Combined MASH Resolution + Fibrosis Improvement35% (1200 mg) vs 9% PBO ~ 6–7% vs placeboVariable ~ 21%Highest ratesNot applicableWeight Loss + 2.4–2.7 kg (increase) ~ 0.5 kg lossUp to 15% loss ~ 2–3% lossUp to 20% lossNot reportedHbA1c Reduction (T2DM patients) > 0.5% (T2DM subset)Minimal1.5–1.8% ~ 0.5–0.8%2.0–2.5%Not reportedPruritus Risk < 5% overall16.55% (RR 1.75‡)RareRareRareNone reportedLipid ChangesFavorable improvementsDyslipidemia risk (↑ TC, LDL)FavorableFavorableFavorableNot reportedGI Adverse EventsDiarrhea, nausea (common)Constipation riskNausea, vomiting (common)Nausea, diarrhea (common)GI events (manageable)MinimalDiscontinuation Rate < 5% for AEs ~ 1–10% (due to pruritus) ~ 10–15%Low (~ 2%) ~ 5–10%0% (no SAEs)Study Duration24 weeks18–48 months24 weeks-6 months52–96 weeks24 weeks24 weeksTrial Population247 (76% F2/F3)931–1,607 (F2/F3)2,000 + 1,500 + 1,200 + 6 patients (F2-F4 fibrosis)References [30, 116, 120–127]Legend: PBO Placebo, RA Receptor Agonist, THR-β Thyroid Hormone Receptor-beta, GLP-1 Glucagon-Like Peptide-1, GIP Glucose-dependent Insulinotropic Polypeptide, AE Adverse Events, w/o without, TC Total Cholesterol, LDL Low-Density Lipoprotein, GI Gastrointestinal Comparison of anti-MASH agents across 12 key efficacy and safety Legend: PBO Placebo, RA Receptor Agonist, THR-β Thyroid Hormone Receptor-beta, GLP-1 Glucagon-Like Peptide-1, GIP Glucose-dependent Insulinotropic Polypeptide, AE Adverse Events, w/o without, TC Total Cholesterol, LDL Low-Density Lipoprotein, GI Gastrointestinal Adjunctive molecular strategies (supporting pathways) Beyond direct cytoskeletal modulation, adjunctive therapies target the pathological microenvironment and systemic signaling that reinforce capillarization, creating a permissive setting for core therapies to function. A key strategy involves modulating systemic signaling hubs to indirectly support LSEC health. Metabolic regulators (FXR agonists) FXR agonists, exemplified by obeticholic acid (OCA), function as central metabolic regulators that suppress de novo lipogenesis, enhance biliary bile acid excretion, and induce hepatoprotective fibroblast growth factor 19 (FGF19). This factor in turn attenuates TGF-β1-mediated pro-fibrotic signaling and collagen IV synthesis, while directly promoting a favorable sinusoidal phenotype [128, 129]. Clinical evidence demonstrates both potential and limitations. In the REGENERATE trial, OCA at 25 mg showed fibrosis improvement in 23% versus 12% on placebo at 18 months, though MASH resolution was not significantly improved [130]. However, pruritus occurred in 51% of patients at this dose, severely impacting quality of life and leading to high discontinuation rates [131]. The pathogenesis of pruritus involves off-target receptor activation, while class-effect dyslipidemia poses theoretical cardiovascular risks [132]. Development of more selective agonists such as tropifexor continues, though dose-dependent pruritus and lipid alterations persist in trials [133, 134]. Response heterogeneity further complicates risk–benefit assessments, underscoring the need for predictive biomarkers. While FXR agonists represent a mechanistically promising approach to Nexus modulation, their clinical utility remains constrained by adverse effects that limit tolerability and widespread adoption. Anti-inflammatory/antioxidant agents (lactoferrin) Concurrently, mitigating the inflammatory and oxidative insults that accelerate Nexus disruption is crucial. The multifunctional glycoprotein lactoferrin exemplifies this approach, exerting protective effects by reducing oxidative stress and downregulating defenestrating cytokines such as TNF-α, thus preserving the integrity of the VEGF-VEGFR2 signaling complex [135]. ECM remodeling inhibitors (MMP-9 ASOs) A primary physical adjunctive target is the vicious cycle of pathological ECM remodeling. Capillarization-driven HSC activation deposits a stiff ECM that further perpetuates endothelial dysfunction via integrin signaling. Matrix metalloproteinase (MMP) inhibitors, particularly MMP9-specific antisense oligonucleotides (ASOs), aim to prevent this structural consolidation by targeting a key enzyme in basement membrane deposition [136, 137]. However, the dual role of MMPs in physiological tissue turnover necessitates spatiotemporally controlled inhibition, likely via advanced delivery systems, to avoid compromising beneficial tissue remodeling. Immunomodulatory agents (adenosine agonists) The adenosine A2A receptor agonist CGS21680 targets sinusoidal inflammation by polarizing macrophages toward an anti-inflammatory phenotype and suppressing cytokines such as TNF-α [138]. Furthermore, the link between coagulation and endothelial dysfunction underscores the potential of recombinant thrombomodulin, which exerts concurrent anticoagulant, anti-inflammatory, and cytoprotective effects to stabilize the LSEC phenotype [139]. Enabling technologies: overcoming barriers to delivery In advanced disease, capillarization and fibrosis create a physical barrier that impedes therapeutic access. Nanoparticle (NP)-based delivery systems are engineered to overcome this obstacle. LSEC-targeted NPs can restore fenestration density in HCC, enhancing drug penetration [140]. For fibrotic livers, sequential “nano-penetrator” strategies that first degrade ECM barriers (e.g., with losartan) followed by ligand-conjugated NPs (e.g., targeting scavenger receptor class A member 3 [Scara3]) enable the restoration of LSEC permeability and targeted delivery of antifibrotic agents [141–143]. A paradigm-shifting advancement is the development of artificial intelligence (AI)-designed spherical nucleic acids (SNAs). By integrating multi-omics data with machine learning, miR-325-3p was identified as a key mediator of fenestration restoration. These SNAs exploit Scara3 for cell-specific delivery, and a single dose has been shown to reverse capillarization, reduce collagen deposition by more than 50%, and normalize LSEC ultrastructure across multiple fibrosis models. Moreover, this approach synergistically combines AI-driven target discovery with precision nanocarrier design [144, 145]. Regenerative strategies: stem cell-based therapies Regenerative medicine aims to repair or replace capillarized LSECs. Current strategies focus on mesenchymal stem cell (MSC) transplantation and human induced pluripotent stem cell (hiPSC)-derived endothelial cells (iECs). MSC therapies demonstrate a unique capacity to reverse capillarization, improve microcirculation, and reduce fibrosis [146, 147]. In contrast, hiPSC-derived iECs can engraft and differentiate into LSEC-like populations in vivo; however, functional maturation remains a significant challenge, with only approximately 5% of engrafted cells developing fenestrations [148]. Consequently, critical priorities for clinical translation include advancements in bioreactor-based 3D culture systems and precision gene-editing tools to enhance engraftment efficiency and functional maturation [149–152]. Challenges and future prospects: the path to clinical translation The translation of LSEC-targeted therapies faces substantial biological, technical, and clinical hurdles. Fundamental biological and technical hurdles A paramount challenge is the inadequacy of existing preclinical models. Animal models carry interspecies limitations, and conventional 2D cell cultures fail to recapitulate the biomechanical niche essential for the fenestrated phenotype, a fact underscored by the rapid loss of this phenotype in immortalized LSEC lines in vitro. Emerging human-centric platforms, such as 3D bioprinted sinusoids and organ-on-chip systems, offer improved physiological relevance. However, they require further development to fully incorporate disease-specific ECM remodeling and mechanotransduction pathways [149, 151, 152]. Indeed, these advanced models often oversimplify disease-specific ECM remodeling and lack full integration of key mechanotransduction pathways (e.g., Piezo1/TRPV4), necessitating the development of more dynamic, multicellular systems [153, 154]. As summarized in (Table 3), the field’s reliance on these models has likely contributed to the repeated failure of single-pathway therapies in human trials, highlighting an urgent need for more human-relevant, multicellular experimental systems.Table 3Comparative analysis of in vitro LSEC culture model systemsParameter2D Monoculture3D Liver OrganoidsLiver-on-Chip (LoC)iPSC-Derived LSECsMulti-Organ ChipsReferencesPhysiological RelevanceLowHighHighModerateVery High [150, 151, 154, 155]Fenestration Maintenance < 3 days ~ 5 days > 7 daysVariableDependent on integrated cues [155–158]Multicellular CrosstalkLSEC-onlyclusters of organ-specific cellsLSEC-hepatocyte-stellate-KupfferLow (limited to co-culture)Very High (Adipose-gut and liver tissue)Biomechanical CuesNoneECMYes (flow, mechanical strain)Partial (ECM stiffness)Yes (flow, mechanical strain) [140, 154, 155]Disease ModelingLimitedModerateHigh (fibrosis, MASLD/MASH)Low (immature ECM/disease)systemic diseaseScalabilityHighModerateLowHighVery LowCostLowModerateHighModerateVery HighTranslational PotentialLowModerateHighHighVery High Table 3Comparative analysis of in vitro LSEC culture model systemsParameter2D Monoculture3D Liver OrganoidsLiver-on-Chip (LoC)iPSC-Derived LSECsMulti-Organ ChipsReferencesPhysiological RelevanceLowHighHighModerateVery High [150, 151, 154, 155]Fenestration Maintenance < 3 days ~ 5 days > 7 daysVariableDependent on integrated cues [155–158]Multicellular CrosstalkLSEC-onlyclusters of organ-specific cellsLSEC-hepatocyte-stellate-KupfferLow (limited to co-culture)Very High (Adipose-gut and liver tissue)Biomechanical CuesNoneECMYes (flow, mechanical strain)Partial (ECM stiffness)Yes (flow, mechanical strain) [140, 154, 155]Disease ModelingLimitedModerateHigh (fibrosis, MASLD/MASH)Low (immature ECM/disease)systemic diseaseScalabilityHighModerateLowHighVery LowCostLowModerateHighModerateVery HighTranslational PotentialLowModerateHighHighVery High Comparative analysis of in vitro LSEC culture model systems Compounding the challenge of disease modeling are the significant barriers to targeted drug delivery. In the fibrotic liver, capillarization and dense ECM deposition create a physical barrier that can reduce nanoparticle transit efficiency by up to 72% [141]. This problem is exacerbated by profound LSEC heterogeneity; receptor expression varies dramatically by zonation and disease stage (e.g., CD36 overexpression in MASLD vs. stabilin-2 loss in cirrhosis), which challenges the efficacy of ligand-specific targeting strategies. Established LSEC cell lines:utility and limitations as in vitro models The establishment of immortalized LSEC lines from human, mouse, and rat livers represented a critical advance in endothelial cell biology, providing a renewable and accessible platform for controlled in vitro research. Subsequent studies generated immortalized LSEC lines from multiple sources: human lines include SK Hep1 and TMNK-1; mouse lines include TSEC; and rat lines include NP11, NP26, NP31, and NP32. Researchers created these cell lines through two complementary approaches: spontaneous immortalization or genetic introduction of SV40 T-antigen and human telomerase reverse transcriptase (hTERT) [159–164]. As a result, these models have served as invaluable tools for controlled, reductionist investigations into LSEC biology [165–167]. Notably, some lines, including SK Hep1, have demonstrated a capacity to maintain fenestrations under specific culture conditions, highlighting their intrinsic potential [163]. However, a significant and persistent knowledge gap remains, as these fenestration-competent models have not been systematically leveraged to define how core signaling pathways, specifically the pro-fenestrative VEGF-VEGFR2 axis and the cytoskeletal control exerted by the RhoA-ROCK equilibrium, directly govern fenestration dynamics using immortalized LSEC lines. This underutilization reflects a fundamental constraint: conventional 2D monoculture causes rapid loss of the fully differentiated LSEC phenotype. This loss demonstrates that LSECs are not cell-autonomous; rather, they depend critically on the permissive sinusoidal niche [64]. Consequently, while these immortalized lines have been instrumental for studying general processes such as angiogenesis, their underutilization for studying fenestration dynamics represents a critical void and an immediate opportunity. Future research must systematically employ these accessible models, under optimized culture conditions, to deconstruct the precise hierarchical control of the VEGF-NO and RhoA-ROCK pathways over the sieve plate. This very shortcoming, however, proved catalytic, driving the field toward developing more physiologically relevant systems that incorporate essential biomechanical and paracrine cues to authentically recapitulate the liver’s vascular signaling Nexus. Therapeutic and safety bottlenecks Several interconnected bottlenecks hamper the clinical application of LSEC-targeted therapies. Temporal heterogeneity of LSEC capillarization poses a major challenge; interventions effective in early MASLD may fail in advanced cirrhosis, where the Nexus is irreversibly disrupted [67, 168, 169]. This zonation-dependent heterogeneity of LSEC phenotypes is summarized in (Table 4) Moreover, pharmacological agents often exhibit problematic pleiotropy [122]. For instance, while VEGF reduces portal pressure, its pro-angiogenic nature poses a risk accelerating HCC progression [170–174]. Similarly, systemic NO donors may exacerbate hemodynamic instability in decompensated cirrhosis [175, 176].Table 4LSEC heterogeneity across zonation and disease progressionLSEC FeatureHealthy Liver – Zone 1 (Periportal)Healthy Liver – Zone 3 (Pericentral)MASH (Zone 3 most affected)Cirrhosis (Pan-lobular, Zone 3 worst)ReferencesFenestrationsLarge fenestrae, low densitySmall fenestrae, high densityProgressive loss (starting in zone 3)Complete loss in zone 3; pan-lobulaor absence [177, 178]MechanotransductionVariable shear responsiveness; Piezo1/TRPV4 channels sense flow gradientsEnhanced shear sensing via eNOS/NO and YAP/TAZ pathwaysImpaired sensing (zone 3 earliest)Non-functional in all zones [103, 179, 180]Metabolic FunctionLipid scavenging, CR1-mediated uptake, active antigen presentation/immune surveillanceEndocytosis of growth factors, enhanced eNOS metabolic flux, fenestration-dependent sieving, lipid homeostasisImpaired lipid clearance, immune activation, CD36/SR-A overexpressionChronic inflammation, systemic oxidative stress [101, 169, 177, 181–184]Molecular Markers – Capillarization (CD34, vWF, F8, PECAM1)Higher baseline expression (physiological)Lower baseline expressionModerate upregulation (zone 3 begins to rise)Very high in all zones, but highest in zone 3 (FC ~ 8–10 × vs. control) [168, 185, 186]Molecular Markers – LSEC Identity (Lyve1, Stabilin-1, Stabilin-2)Low/ModerateModerateDownregulation (zone 3 earliest)Severe downregulation (FC ~ 0.2–0.4 for Stab1/2 in zone 3) Table 4LSEC heterogeneity across zonation and disease progressionLSEC FeatureHealthy Liver – Zone 1 (Periportal)Healthy Liver – Zone 3 (Pericentral)MASH (Zone 3 most affected)Cirrhosis (Pan-lobular, Zone 3 worst)ReferencesFenestrationsLarge fenestrae, low densitySmall fenestrae, high densityProgressive loss (starting in zone 3)Complete loss in zone 3; pan-lobulaor absence [177, 178]MechanotransductionVariable shear responsiveness; Piezo1/TRPV4 channels sense flow gradientsEnhanced shear sensing via eNOS/NO and YAP/TAZ pathwaysImpaired sensing (zone 3 earliest)Non-functional in all zones [103, 179, 180]Metabolic FunctionLipid scavenging, CR1-mediated uptake, active antigen presentation/immune surveillanceEndocytosis of growth factors, enhanced eNOS metabolic flux, fenestration-dependent sieving, lipid homeostasisImpaired lipid clearance, immune activation, CD36/SR-A overexpressionChronic inflammation, systemic oxidative stress [101, 169, 177, 181–184]Molecular Markers – Capillarization (CD34, vWF, F8, PECAM1)Higher baseline expression (physiological)Lower baseline expressionModerate upregulation (zone 3 begins to rise)Very high in all zones, but highest in zone 3 (FC ~ 8–10 × vs. control) [168, 185, 186]Molecular Markers – LSEC Identity (Lyve1, Stabilin-1, Stabilin-2)Low/ModerateModerateDownregulation (zone 3 earliest)Severe downregulation (FC ~ 0.2–0.4 for Stab1/2 in zone 3) LSEC heterogeneity across zonation and disease progression Regenerative approaches face their own translational hurdles [187]. While iPSC-derived LSECs show therapeutic potential, engraftment rates remain below 5% in preclinical models, and the rarity of fully differentiated, fenestrated cells highlights persistent challenges in scalability and functional integration. Finally, regulatory pathways for advanced therapies such as CRISPR-Cas9-mediated gene editing, despite their preclinical promise, face stringent scrutiny due to legitimate concerns over off-target effects and potential systemic toxicity [188–192]. Phenotypic plasticity and the biomarker imperative MASLD represents a continuum of disease states shaped by the plasticity of liver endothelial cells. In the early stages, LSECs retain a restoration-competent phenotype, supported by an open chromatin configuration that maintains expression of key regulators like KLF2 and NOS3, preserving their characteristic fenestrations. This plasticity, the ability to revert toward a healthy endothelial identity defines a critical window where therapeutic intervention is most effective [193]. The reversibility of early LSEC defenestration hinges on preserved chromatin accessibility and the absence of permanent epigenetic silencing. Dietary intervention studies reveal biphasic recovery dynamics: rapid functional improvement precedes slower structural restoration of fenestrations [194, 195]. In patients with early-stage disease (F0–F1), therapies aimed at reversing LSEC dysfunction such as ROCK inhibitors (e.g., fasudil/Y-27632), nitric oxide donors, and activation of the VEGF–NO axis can deliver both rapid functional gains and support structural regeneration. ROCK inhibitors, already in clinical development, have shown promise in mitigating capillarization and inflammation during early disease [196–198]. As disease progresses, mechanotransduction pathways like YAP/TAZ, activated by increasing extracellular matrix stiffness, act as epigenetic switches that lock LSECs into a capillarized state [199, 200] These insights define distinct therapeutic windows across disease stages: aggressive, reversal-focused strategies in early disease; combination therapies during transitional phases; and regenerative approaches in advanced disease. This stage-dependent framework supports precision management of MASLD, tailoring treatment intensity and mechanism to the individual’s disease stage. The shift from a plastic to a fixed phenotype represents a molecular point of no return, where disease duration becomes the key determinant of reversibility. This reality underscores the urgent need for biomarkers capable of capturing phenotypic plasticity a task that conventional bulk disease markers cannot accomplish. While markers like von Willebrand Factor (vWF) and hyaluronic acid (HA) reflect endothelial stress and overall severity, they fail to distinguish between patients with reversible endothelial plasticity and those with irreversible capillarization [119, 201, 202]. FABP4 emerges as a promising early biomarker in MASLD, not only detecting initial capillarization but also playing an active role in disease progression [203]. Its dual function makes FABP4 particularly useful for identifying at-risk patients before irreversible epigenetic changes occur. A precision stratification approach for MASLD/MASH integrates biomarker profiles with disease stage, as outlined below: Early-stage patients (high FABP4, low vWF) likely retain restoration-competent LSECs and may benefit from metabolic interventions such as PPAR agonists aimed at disease reversal.Intermediate patients with elevated vWF show signs of progressive capillarization and may require combination therapies.Advanced patients (high FABP4 and markedly elevated vWF) have likely reached an epigenetically fixed state, warranting integrated multi-modal strategies that combine metabolic modulation with regenerative approaches, rather than reversal-focused therapies alone. Early-stage patients (high FABP4, low vWF) likely retain restoration-competent LSECs and may benefit from metabolic interventions such as PPAR agonists aimed at disease reversal. Early-stage patients (high FABP4, low vWF) likely retain restoration-competent LSECs and may benefit from metabolic interventions such as PPAR agonists aimed at disease reversal. Intermediate patients with elevated vWF show signs of progressive capillarization and may require combination therapies. Intermediate patients with elevated vWF show signs of progressive capillarization and may require combination therapies. Advanced patients (high FABP4 and markedly elevated vWF) have likely reached an epigenetically fixed state, warranting integrated multi-modal strategies that combine metabolic modulation with regenerative approaches, rather than reversal-focused therapies alone. Advanced patients (high FABP4 and markedly elevated vWF) have likely reached an epigenetically fixed state, warranting integrated multi-modal strategies that combine metabolic modulation with regenerative approaches, rather than reversal-focused therapies alone. The precise epigenetic and molecular mechanisms distinguishing restoration-competent from epigenetically fixed disease states in human MASLD/MASH remain incompletely understood and warrant prospective validation (Table 5).Table 5Clinical biomarkers of LSEC dysfunctionMarkerDisease StageLSEC-Specific Clinical SignificanceReferencesHyaluronic Acid (HA)Fibrosis/CirrhosisDirectly reflects impaired LSEC endocytic capacity and reduced clearance function; elevated HA indicates loss of LSEC fenestration and dysfunction of the endocytic machinery that normally captures and catabolizes serum hyaluronic acid, making it a marker of LSEC-specific disease burden [204]CD34Early MASLD → Advanced MASHA marker of LSEC capillarization indicating loss of fenestrations and transition to continuous endothelium. Increased CD34 staining directly correlates with degree of endothelial dedifferentiation and porosity loss. CD34 intensity increases from periportal to pericentral zones and worsens with fibrosis progression [119]vWF (von Willebrand Factor)Progressive MASLD → CirrhosisSelectively elevated in capillarized LSECs due to endothelial injury and loss of normal intracellular storage; released into circulation as LSECs acquire a pro-inflammatory and prothrombotic phenotype. Predicts severity of portal hypertension and LSEC vascular dysfunction [205]VCAM-1 and ICAM-1MASH, Advanced DiseaseEpigenetically upregulated adhesion molecules specific to activated/capillarized LSECs; mediate pathological monocyte and leukocyte adhesion and transendothelial migration. Indicate LSEC-driven pro-inflammatory endotheliopathy and loss of immunoregulatory function [206]Endothelial Microparticles (EMPs)Portal Hypertension, MASHShed from apoptotic and activated LSECs; mark acute endothelial injury and loss of LSEC structural integrity. EMPs incorporate tissue factor and procoagulant phosphatidylserine, driving both microvascular thrombosis and hepatic vascular dysfunction [207]Endoglin (ENG)/Soluble Endoglin (sENG)MASLD → CirrhosisENG expression is dysregulated in LSECs during metabolic and cholestatic stress. sENG, cleaved by MMP-14 from activated LSECs, reflects extent of LSEC inflammation, endothelial dysfunction, and loss of proper sinusoidal integrity [208]Type IV Collagen and LamininProgressive FibrosisDirectly reflect LSEC capillarization with thickened basement membrane deposition and loss of the Space of Disse; represent physical consolidation of the defenestrated phenotype. Indicate irreversibility of LSEC structural changes [209]Angiopoietin-2 (Ang-2)MASH, Portal HypertensionElevated in capillarized LSECs; acts as a vascular destabilization factor and promotes LSEC inflammation and dysregulation of the Ang-Tie2 axis. LSEC-derived Ang-2 indicates loss of endothelial stabilization and progression toward pathological angiogenesis [210]FABP4Early MASLD → MASHA master lipotoxicity sensor and pro-capillarization mediator expressed in LSECs; dual function as active driver of disease progression and early biomarker. FABP4 identifies at-risk patients before irreversible epigenetic silencing of LSEC fenestration genes occurs [203]ADMA (Asymmetric Dimethylarginine)Portal Hypertension, Advanced DiseaseElevated in capillarized LSECs with impaired eNOS/NO signaling; correlates with reduced LSEC NO bioavailability and increased vasoconstriction. ADMA levels directly reflect dysfunction of the VEGF-NO axis that normally maintains fenestration [211]LSEC-derived Extracellular Vesicles (EVs)Protective (in vitro model)Inhibit activation of hepatic stellate cells (HSCs) and Kupffer cells in vitro, suggesting anti-fibrotic and anti-inflammatory potential. May represent a homeostatic LSEC function lost during capillarization [212] Table 5Clinical biomarkers of LSEC dysfunctionMarkerDisease StageLSEC-Specific Clinical SignificanceReferencesHyaluronic Acid (HA)Fibrosis/CirrhosisDirectly reflects impaired LSEC endocytic capacity and reduced clearance function; elevated HA indicates loss of LSEC fenestration and dysfunction of the endocytic machinery that normally captures and catabolizes serum hyaluronic acid, making it a marker of LSEC-specific disease burden [204]CD34Early MASLD → Advanced MASHA marker of LSEC capillarization indicating loss of fenestrations and transition to continuous endothelium. Increased CD34 staining directly correlates with degree of endothelial dedifferentiation and porosity loss. CD34 intensity increases from periportal to pericentral zones and worsens with fibrosis progression [119]vWF (von Willebrand Factor)Progressive MASLD → CirrhosisSelectively elevated in capillarized LSECs due to endothelial injury and loss of normal intracellular storage; released into circulation as LSECs acquire a pro-inflammatory and prothrombotic phenotype. Predicts severity of portal hypertension and LSEC vascular dysfunction [205]VCAM-1 and ICAM-1MASH, Advanced DiseaseEpigenetically upregulated adhesion molecules specific to activated/capillarized LSECs; mediate pathological monocyte and leukocyte adhesion and transendothelial migration. Indicate LSEC-driven pro-inflammatory endotheliopathy and loss of immunoregulatory function [206]Endothelial Microparticles (EMPs)Portal Hypertension, MASHShed from apoptotic and activated LSECs; mark acute endothelial injury and loss of LSEC structural integrity. EMPs incorporate tissue factor and procoagulant phosphatidylserine, driving both microvascular thrombosis and hepatic vascular dysfunction [207]Endoglin (ENG)/Soluble Endoglin (sENG)MASLD → CirrhosisENG expression is dysregulated in LSECs during metabolic and cholestatic stress. sENG, cleaved by MMP-14 from activated LSECs, reflects extent of LSEC inflammation, endothelial dysfunction, and loss of proper sinusoidal integrity [208]Type IV Collagen and LamininProgressive FibrosisDirectly reflect LSEC capillarization with thickened basement membrane deposition and loss of the Space of Disse; represent physical consolidation of the defenestrated phenotype. Indicate irreversibility of LSEC structural changes [209]Angiopoietin-2 (Ang-2)MASH, Portal HypertensionElevated in capillarized LSECs; acts as a vascular destabilization factor and promotes LSEC inflammation and dysregulation of the Ang-Tie2 axis. LSEC-derived Ang-2 indicates loss of endothelial stabilization and progression toward pathological angiogenesis [210]FABP4Early MASLD → MASHA master lipotoxicity sensor and pro-capillarization mediator expressed in LSECs; dual function as active driver of disease progression and early biomarker. FABP4 identifies at-risk patients before irreversible epigenetic silencing of LSEC fenestration genes occurs [203]ADMA (Asymmetric Dimethylarginine)Portal Hypertension, Advanced DiseaseElevated in capillarized LSECs with impaired eNOS/NO signaling; correlates with reduced LSEC NO bioavailability and increased vasoconstriction. ADMA levels directly reflect dysfunction of the VEGF-NO axis that normally maintains fenestration [211]LSEC-derived Extracellular Vesicles (EVs)Protective (in vitro model)Inhibit activation of hepatic stellate cells (HSCs) and Kupffer cells in vitro, suggesting anti-fibrotic and anti-inflammatory potential. May represent a homeostatic LSEC function lost during capillarization [212] Clinical biomarkers of LSEC dysfunction A roadmap to precisely restore LSEC fenestrations Addressing these substantial challenges requires a paradigm shift toward comprehensive, mechanistically informed research aimed at restoring fenestrations through targeted modulation of the LSEC Fenestration Signaling Nexus. We propose a three-pronged strategic roadmap. First, comprehensive characterization of Nexus heterogeneity Systematic resolution of LSEC heterogeneity across disease progression is essential. Single-cell and spatial multi-omics technologies provide unprecedented opportunities to map the LSEC Fenestration Signaling Nexus with cellular resolution. This profiling must address zonally distinct LSEC phenotypes, across disease stages. Critically, age- and sex-stratified analyses must elucidate demographic disparities in disease susceptibility and therapeutic responsiveness. Second, advanced human-relevant experimental systems Current preclinical platforms inadequately capture the biomechanical and paracrine signaling governing LSEC fenestration maintenance. Next-generation systems must incorporate: (i) flow-responsive mechanotransduction pathways; (ii) multicellular architecture (hepatocytes, stellate cells, Kupffer cells); and (iii) dynamic extracellular matrix composition reflecting disease-stage remodeling. These platforms will enable accurate prediction of therapeutic responses across heterogeneous patient populations, reducing late-stage clinical trial failures. Third, rational Nexus-priming combinatorial therapies Single-target approaches have proven clinically insufficient. We propose sequential, mechanistically complementary interventions: (i) a “Nexus-priming” agent that creates a permissive microenvironment, exemplified by ROCK inhibitors or FXR agonists, followed by (ii) a “fenestration-restoring” agent promoting fenestration reassembly, such as nitric oxide donors or pro-fenestration microRNAs. Artificial intelligence can model the Nexus network in silico, integrating multi-omics data to identify optimal drug combinations and dosing sequences while minimizing off-target toxicity. This data-informed approach represents the most promising pathway toward personalized, effective fenestration-restoring therapies. Conclusion The LSEC Fenestration Signaling Nexus model reframes capillarization as a dynamic, targetable driver of metabolic liver disease. The path to clinical translation is fraught with challenges, including model limitations and drug pleiotropy. However, the convergence of spatial multi-omics, AI-driven discovery, and precision nanocarriers presents an unprecedented opportunity. The future lies in rational combinatorial strategies that simultaneously suppress pro-capillarization drivers and reactivate pro-fenestration pathways. By embracing this integrated approach, LSEC-targeted therapy can progress from concept to clinical cornerstone for metabolic liver disease. Supplementary Information Supplementary Material 1. Supplementary Material 1.