Portal vein thrombosis (PVT) is a frequent complication in patients with cirrhosis, yet its pathogenesis remains incompletely understood. Although traditionally regarded as a thrombotic disorder, accumulating evidence indicates that cirrhotic PVT represents a primary vascular disease characterized by endothelial dysfunction and progressive intimal remodeling of the portal vein wall [1,2]. In this context, Anton et al. demonstrated clear features of endothelial-to-mesenchymal transition (EndMT) in portal vein endothelial cells, leading to eccentric intimal hyperplasia and suggesting that portal vein occlusion in cirrhosis is predominantly fibrotic rather than fibrin-rich [3]. These findings support the concept that portal vein obstruction in cirrhosis reflects a maladaptive vascular response rather than a classical thrombotic event.
While inflammation, portal hypertension, systemic endotoxemia, and hepatic dysfunction are recognized contributors to PVT development, none adequately explain why thrombosis occurs selectively within specific segments of the portal vein. This unresolved question highlights a gap in understanding the focal and segmental nature of PVT [2]. In this regard, bile acids (BAs), as key regulators of inflammation and fibrogenesis through FXR-mediated signaling, provide a plausible mechanistic link [4]. Emerging evidence from NASH models suggests that alterations in enterohepatic BA composition, together with their downstream signaling effects, may directly influence portal endothelial biology and contribute to the segmental susceptibility observed in cirrhotic PVT [5]. Preclinical analyses by Gillard et al. demonstrated a marked reduction in total and, in particular, secondary bile acids in portal blood, while systemic BA levels remained unchanged [5]. This dissociation suggests that portosystemic BA profiles differ substantially in both concentration and composition, and systemic sampling does not reflect the biochemical environment encountered by the portal endothelium [5]. Given the continuous enterohepatic recirculation of bile acids through the portal venous system, portal endothelial cells are exposed to a specialized local bile acid microenvironment regulated by FXR- and TGR5-dependent signaling pathways [6]. Given that portal BA concentrations are substantially higher than systemic levels, endothelial BA sensing is largely dependent on this localized microenvironment [5]. Secondary bile acids—especially deoxycholic acid (DCA) and lithocholic acid (LCA)—play a central role in activating anti-inflammatory and antifibrotic receptors such as FXR and TGR5. Their reduction leads to diminished receptor activation, impaired endothelial homeostasis, and increased susceptibility to inflammatory and stress-related responses [5].
Mechanistically, FXR signaling enhances endothelial nitric oxide (NO) production via eNOS activation and suppresses inflammatory pathways, thereby regulating vascular tone and resistance [7,8]. Similarly, TGR5 activation contributes to endothelial protection by promoting NO production and maintaining vascular homeostasis [8]. Experimental data further demonstrate that TGR5 activation suppresses macrophage-driven vascular and hepatic inflammatory responses through NF-κB inhibition, including β-arrestin2–IκBα interaction, while reducing vascular lesion formation and supporting a vasculoprotective role for bile acid signaling [9,10]. Loss of these signaling pathways may therefore facilitate endothelial dysfunction, promote macrophage infiltration, and create a microenvironment permissive for EndMT. Histological observations of CD68⁺ macrophage accumulation and myofibroblast-like cell presence in the portal vein wall support this interpretation [3].
Within this framework, immune activation may further amplify vascular remodeling. Neutrophil extracellular traps (NETs) have been shown to enhance fibrotic thrombus remodeling and endothelial injury through macrophage–myofibroblast transition (MMT), reinforcing an EndMT-prone phenotype [11].Additionally, extracellular DNA derived from NETs can activate the TGF-β1/Smad3 signaling pathway, linking thrombus organization with progressive intimal fibrosis [11]. These findings suggest that immunothrombotic processes may act in concert with endothelial signaling disturbances to drive structural changes within the portal vein.
Despite the central role of EndMT in cirrhotic PVT, the segmental distribution of this process remains unexplained. One plausible explanation is that FXR and TGR5 signaling operate in a threshold-dependent manner, such that even subtle changes in local BA composition may lead to significant reductions in receptor activation [5]. This may render specific portal vein segments more vulnerable to endothelial dysfunction. In these susceptible regions, EndMT activation may lead to eccentric intimal hyperplasia, extracellular matrix deposition, and progressive luminal narrowing, thereby providing a mechanistic basis for the focal nature of PVT.
The limitations of Virchow’s triad in explaining cirrhotic PVT further support this paradigm. Hypercoagulability alone does not independently predict PVT development, and the low prevalence of inherited or acquired thrombophilia suggests a limited role for classical thrombotic mechanisms [12]. Consistently, prospective data have shown that inherited/acquired hemostatic abnormalities and inflammatory markers, including neutrophil extracellular trap formation, do not independently predict PVT development, whereas factors associated with severe portal hypertension remain dominant predictors [13]. Although reduced portal flow velocity (<15 cm/s) is a recognized risk factor, the segmental occurrence of PVT indicates that hemodynamic factors are insufficient to fully account for disease development [12]. Notably, intimal thickening has been demonstrated even in cirrhotic patients without radiologically evident PVT, supporting the possibility that vascular remodeling may precede overt thrombosis [14]. Moreover, the portal venous system exhibits a distinct biochemical environment compared to the systemic circulation, characterized by elevated levels of vWF, FVIII, sulfated glycosaminoglycans, and endothelial-derived microparticles [12], all indicative of endothelial dysfunction. When integrated with evidence of EndMT ³ and predominant intimal fibrosis [1], these findings support the concept that PVT represents a localized vascular disorder rather than a classical thrombotic event. The addition of bile acid–mediated signaling disruption introduces a metabolic dimension that may trigger segmental endothelial responses [5].
Taken together, these findings support a metabolic–endothelial model of cirrhotic PVT in which disruption of secondary bile acid–driven FXR/TGR5 signaling promotes EndMT-associated vascular remodeling within the portal venous microenvironment [5,6,9,10]. This framework may explain the fibrotic nature, segmental distribution, and limited anticoagulant responsiveness of PVT, consistent with histological evidence demonstrating predominant intimal fibrosis rather than fibrin-rich thrombus [1] and with emerging data supporting substantial regional heterogeneity in portal hemostasis [15].
FundingThe author declares that no funds, grants, or other support were received during the preparation of this manuscript.
Author contributionsÖzlem Beydaş conceived the idea for the manuscript, conducted the literature review, and wrote the manuscript.
The author read and approved the final version of the manuscript.
None.
