Mechanistic Studies of Cholesterol Lipotoxicity Pertinent to NASH

Abstract

Non-alcoholic fatty liver disease (NAFLD) affects ~30% of the world population with similar or higher prevalence in Australia. The pathology of NAFLD ranges from benign simple steatosis (SS) to non-alcoholic steatohepatitis (NASH), with fibrosis that can progress to cirrhosis. Understanding the pathogenesis of NASH remains a challenge. In both humans and mice, obesity, diabetes and metabolic syndrome are associated with NAFLD, but free cholesterol (FC) accumulates in livers showing NASH but not in those with SS. Such cholesterol-loaded livers are sensitised to cytokine-mediated mitochondrial injury. However, at the time this research was conducted, there was no direct evidence that linked FC lipotoxicity to hepatocyte cell death or inflammatory recruitment.

In this thesis, primary murine hepatocytes were loaded with FC by exposing them to human low-density lipoprotein (LDL), and these cells were used to characterise the mechanisms of hepatocellular injury and cell death (apoptosis and necrosis). In particular we tested the hypothesis that c-Jun N-terminal kinase (JNK) activation and mitochondrial injury are essential steps in FC hepatocellular lipotoxicity. We also examined how FC-injured hepatocytes could promote activation of Kupffer cells (KC), which is a key feature of liver inflammation in NASH.

The background to NAFLD and NASH as an important public health problem, and as a liver disease is introduced in Chapter 1. Concepts about NASH pathogenesis are discussed in light of available knowledge up to the start of this PhD in 2011. The common research materials and methods are discussed in Chapter 2.

In Chapter 3, the novel in vitro model of FC-loaded primary murine hepatocytes is described. Briefly, primary murine hepatocytes (C57B6/J wild type [WT]) were incubated with LDL (0–40 µM), and shown to be loaded with FC. The subcellular sites of primary hepatocyte FC were determined by co-localising filipin fluorescence with organelle markers. The results were compared with the intracellular distribution of FC seen in atherogenic diet-fed foz/foz mouse livers, an in vivo model of NASH. In mice with NASH, FC co-localised to plasma membrane (PM), mitochondria and endoplasmic reticulum (ER) compartments. This pattern was replicated in hepatocytes incubated with LDL to dose-dependently increase hepatocyte FC. Further, FC loading reduced PM fluidity and caused cell surface blebbing, with release of extracellular vesicles (EVs), as evident on scanning and transmission electron microscopy (EM).

In Chapter 4, the role of JNK1 in FC-mediated hepatocellular injury was explored using primary hepatocytes from WT, Jnk1-/-and Jnk2-/- mice. These cells were incubated with LDL (0–40 μM), and molecular pathways of FC-mediated cell death determined by western blot and immunofluorescence. Separate experiments were performed with chemical specific JNK1 inhibitors (CC-401, CC-930 and CC-003) in WT hepatocytes. Supernatant was collected from FC-loaded WT, Jnk1-/- and Jnk2-/- hepatocyte experiments and assayed for high mobility group box 1 (HMGB1) and EVs. Supernatant or EVs from WT FC-injured primary hepatocytes were added to primary KC cultures from WT and Tlr4-/- mice. Ultrastructural changes were assessed by electron microscopy (EM), while TNF and IL-1β release into the supernatant was quantified by enzyme-linked immunosorbent assay (ELISA).

FC loading caused dose-dependent LDH leakage, apoptosis, necrosis and HMGB1 release. At 40 μM LDL, hepatocellular cell death was associated with JNK1 activation, c-Jun phosphorylation, mitochondrial membrane pore transition, cellular oxidative stress (increased GSSG with reciprocal decrease in GSH concentration) and ATP depletion. Administration of JNK inhibitors (CC-401, CC-930 and CC-003) ameliorated hepatocellular apoptosis and necrosis, while Jnk1-/- hepatocytes were refractory to FC-induced injury. Cyclosporine A (inhibits mitochondrial membrane permeability transition [MPT] pore opening) and caspase-3 inhibitors abrogated FC-mediated hepatocellular cell death. Importantly, there was no increase of ER stress proteins in vitro or in vivo, while inhibitors of ER stress-mediated cell death, 4-phenylbutyric acid failed to protect FC-loaded hepatocytes.

In Chapter 5, the supernatant and EVs isolated from pervious experiments were studied, in particular their ability to active KCs and proinflammatory pathways. Addition of HMGB1-enriched culture medium from FC-loaded hepatocytes activated KCs, as assessed by increased nuclear NF-κB (p65) fluorescence, release of IL-1β and TNF-α, and ultrastructural changes. These effects were mitigated by administration of HMGB1-neutralising antibody, and were absent in Myd88-/- knockout hepatocytes. As mentioned above, deposition of FC within the PM of hepatocytes also released EVs and these were shown here to contain HMGB1. The results allowed us to conclude that FC loading of hepatocytes stimulates HMGB1 secretion and release of PM-derived EVs. In turn, HMGB1 activates KCs through a TLR4-MyD88 dependent process.

In Chapter 6 we sought to characterise EVs from both human and experimental NASH, with a particular focus on their cell-of-origin and protein composition. To achieve this, EVs were isolated from healthy human controls, NAFLD patients with simple SS, NASH but with no or mild-moderate fibrosis (F0-F2), and NAFLD with advanced fibrosis (F3-F4), as well as atherogenic diet-fed foz/foz mice with NASH and wildtype (WT) mice with SS. Composition of EVs harvested from the circulation was studied using a combination of western blotting and flow cytometry. EVs were found to circulate in both experimental and human NAFLD, with significantly higher levels in patients with clinical NASH and advanced fibrosis compared with healthy controls or those with SS. Furthermore, these EVs were highly enriched with HMGB1 and TLR4, in addition to CD4-, CD8-, CD36- and CD147-positive markers. A significant proportion of circulating EVs were hepatocellular in origin, as shown by their “tags” of asialoglycoprotein receptor 1 (ASGRP1) and solute carrier family 10 member 1 (SCL10A1). Chapter 7 summarises the key experimental findings from Chapters 3 to 6 in a broader context, and proposes several important directions for future research.

Collectively, the research findings presented here demonstrate that FC deposition in mitochondria and PM causes hepatocyte cell death, confirm the role of JNK1 activation as an important pathway for hepatocyte lipotoxic injury and reveal a link between HMGB1 and EVs with lipotoxicity and engagement of KC activation in a TLR4-dependent manner. It is proposed that this is a likely causal link in the transition from steatosis to NASH. Additionally, in both human and experimental NASH, distinct EV populations circulate, and this provides a potential for development of novel non-invasive diagnostic tests, as well as molecular targets.