Abstract

Background: Metabolic dysfunction-associated steatohepatitis (MASH) pathogenesis involves both lipotoxicity (toxic lipid-induced cellular stress) and a myeloid cells-associated sterile inflammatory response. Liver sinusoidal endothelial cells (LSEC) dysfunction (impaired vasorelaxation) leading to subclinical portal hypertension in non-cirrhotic MASH is a driver of liver fibrosis. However, the role of lipotoxicity in LSEC dysfunction, and the LSEC molecular mediators of the inflammatory response in MASH are still obscure.

Methods: We isolated primary mouse LSECs from wild type (WT) C57BL/6J mice on chow or fat, fructose, and cholesterol (FFC) rich diet, and subjected the LSECs to multi-omics studies. We generated mice with endothelial deletion of GSK3β (GSK3β^∆end) by crossing the GSK3β^flox/flox mice with the Cdh5(PAC)-CreER^T2 mice. We employed the FFC diet and the choline-deficient high-fat diet (CDHFD) to induce MASH in WT and mice with GSK3β^∆end, or GSK3β ^flox/flox.

Results: An integrated pathway analysis of the mouse LSEC proteomics and transcriptomics indicates that leukocyte transendothelial migration and focal adhesion are the major pathways altered in MASH. We performed kinome profiling of the LSEC phospho-proteomics and identified GSK3β as the major hub, and the top kinase in MASH LSECs. We then confirmed that GSK3β is activated in primary human LSECs treated with the toxic lipid palmitate. Palmitate promoted the LSEC pro-inflammatory response by upregulating the chemokine CXCL2 and the adhesion molecule intracellular adhesion molecule 1 (ICAM-1) in a GSK3β-dependent manner. GSK3β activation in LSEC under lipotoxic stress phosphorylated focal adhesion kinase (FAK) and myosin light chain 2 (MLC2), generating contractile stress fibers. Using a flow-based adhesion assay and a collagen gel-based transmigration assay we demonstrated that primary human neutrophil and THP-1 monocyte adhesion and migration through an LSEC monolayer treated with palmitate were reduced by GSK3 inhibition. We employed two mouse models of diet-induced MASH and two structurally distinct GSK3 inhibitors [LY2090314, elraglusib (9-ING-41)] treatment. Both treatment ameliorated liver inflammation, injury, and fibrosis in the two models. Furthermore, Cytometry by the time of flight (Cytof) on isolated intrahepatic leukocytes indicated reduced proinflammatory monocyte-derived macrophages and monocyte-derived dendritic cells in CDHFD-fed, elraglusib (ING)-treated mice (Figure 1A). Portal pressure was also reduced in GSK3β^∆end mice with diet-induced MASH when compared to GSK3β ^flox/flox mice on the same diet (Figure1B). Taken together these data support the role of GSK3 in myeloid cells-associated liver inflammation and portal hypertension in murine MASH.

Conclusion: GSK3 inhibition attenuates toxic lipid induced LSEC pro-inflammatory phenotype and vasoreactivity and may serve as a potential therapeutic strategy in human MASH.