The present study illustrates that the Ldlr−/− mice are a promising long term NASH model, while in contrast to our expectations, APOE2ki mice are not. The HFC diet led to sustained hepatic inflammation, apoptosis and fibrosis in Ldlr−/− mice, but not in APOE2ki mice. This difference was most likely a consequence of the increased sensitivity for oxLDL-induced inflammation in Ldlr−/− mice compared to APOE2ki mice. These novel observations indicate that hyperlipidemia and steatosis are not sufficient for maintaining the inflammatory response in the liver. In addition, our data demonstrate that the Ldlr−/− mouse model can be used as an excellent physiological model particularly vulnerable for investigating hepatic inflammation in the context of fatty liver disease.
Lack of a suitable animal model that faithfully recapitulates the pathophysiology of human NASH is a major obstacle in delineating mechanisms responsible for the progression of steatosis to NASH. The best characterized and most widely used genetic model for NASH is the leptin-deficient mouse (ob/ob). The ob/ob mice develop hepatic steatosis but not hepatic inflammation or fibrosis, possibly due to the loss of normal leptin signalling 
. Therefore, these ob/ob mice need a pro-injurious stimulus, such as an endotoxin (LPS) 
. The most well-known nutritional model for NASH is a diet deficient in methionine and choline (MCD). These mouse models display all of the hallmarks of NASH, from steatosis to inflammation and fibrosis development 
. However, mice fed an MCD-deficient diet tend to lose weight and display lowered plasma TG levels and are therefore very different from NASH in human metabolic syndrome or diabetes patients who are mostly obese and/or hyperlipidemic 
. Therefore, the currently available genetic and nutritional models are especially useful to investigate late stages of NASH but have some limitations restricting their use to serve as a physiological model to study the development of hepatic inflammation in the context of NASH 
Based on the analogy between the mechanisms of NASH and atherosclerosis, an emerging trend in NASH research is to utilize the mouse models traditionally targeted for studies of atherosclerosis, including APOE2ki
mice. The APOE2ki
mouse, a humanized mouse model of hyperlipidemia, has outstanding potential as it is highly responsive to dietary factors and pharmacological interventions 
. As these mice are not commonly used, the effect of feeding a long-term HFC diet on the liver of APOE2ki
mice had so far not been investigated. In contrast, several studies have been performed to determine the consequences of a long-term HFC diet on Ldlr−/−
mice. The Ldlr−/−
mice are mildly hypercholesterolemic due to the absence of LDL receptors, which prolongs the plasma half-life of VLDL and LDL 
. These mice were utilized by Kong et al.
, who revealed that 5 months of HFC feeding in male Ldlr−/−
mice induced macrovesicular steatosis, but not inflammatory cell infiltration 
. In contrast, Yoshimatsu et al.
demonstrated that 3 months of HFC feeding of female Ldlr−/−
mice resulted in steatosis, infiltration of neutrophils into the liver and increased serum aminotransferases (ALT) levels. However, the diet was highly enriched with cholesterol (1.25%) and cholic acid (0.5%), and the latter is a primary bile acid that is known to cause hepatic toxicity. In addition, the effect on hepatic fibrosis was not investigated 
. Finally, Gupte et al.
demonstrated that middle-aged (12-month-old) male Ldlr−/−
mice fed an HFC diet for 3 months developed steatosis, inflammation, fibrosis, oxidative stress and elevated liver injury markers 
. However, the authors concluded that the advanced age of these mice exacerbated the HFC-induced fibrosis 
. Thus, the current data regarding the effect of a long-term HFC diet on NASH progression in hyperlipidemic mice is partial and inconclusive. Besides, elevated liver enzymes such as ALT in the blood usually suggest liver damage, however, there is not always a correlation between these liver enzymes and NAFLD or NASH 
. By using young mice with a physiological diet, we proved for the first time that the Ldlr−/−
mice are a promising model for investigating the onset of hepatic inflammation in NASH and can be a valuable tool for conducting interventional studies; these findings should eventually lead to a better understanding of human NASH and the development of an efficient therapy for this condition.
Studies in apoE−/−
mice show that ApoE influences several inflammatory processes due to the fact that it is produced by a wide variety of cell types, including macrophages 
. However, our data suggest that the APOE2 isoform is not directly involved in inflammation as the inflammatory response was not sustained in the APOE2ki
mice. Unlike APOE2ki
mice have been extensively used as models for atherosclerosis. It was shown that the mean cross-sectional area of the plaque in the APOE2ki
mice is approximately half that seen in age-matched apoE−/−
mice with a similar genetic background and the plaques of the APOE2ki
mice are less mature 
. Moreover, it was also demonstrated that apoE has allele-specific effects in protecting cells from oxidative cell death, with E2 the most effective one 
. In addition, the circulation time of atherogenic particles is also reduced in APOE2ki
mice compared to apoE−/−
mice , which can result in an decreased inflammatory response in the body 
. Altogether, these evidences suggest that APOE2ki
mice are less inflammatory compared to the apoE−/−
mice require an atherogenic diet to develop atherosclerosis, APOE2ki
mice spontaneously develop the full spectrum of atherosclerotic lesions, even on a regular chow diet. Moreover, an atherogenic diet, high in fat and cholesterol, further exacerbates the development of atherosclerosis and xanthomas in these APOE2ki
mice. However, it was shown that the lesions in these APOE2ki
mice mainly consisted of foam cells and had relatively fewer fibrous caps, cholesterol clefts and necrotic cores 
. Thus, in line with our observations, it is likely that the foamy macrophages from the atherosclerotic plaques of APOE2ki mice are less inflammatory compared to those in Ldlr−/−
mice. Currently, APOE2ki
mice are still accepted as an established model for atherosclerosis. Altogether, based on our data and data from the literature, this paradigm should be re-evaluated.
Although hepatic inflammation was completely abolished in APOE2ki
mice after 3 months of HFC feeding, plasma and liver lipid levels were still elevated. Previously, we reported that elevated plasma cholesterol levels can trigger hepatic inflammation 
. In addition, omitting cholesterol from the diet even resulted in a dramatic inhibition of hepatic inflammation, without affecting the levels of steatosis. In line with these findings about steatosis, recent reports have also raised doubts about steatosis as a precondition for the development of inflammation during NASH progression 
,  
. In the present study, neither plasma cholesterol, steatosis, nor anti-oxLDL antibodies were correlated with hepatic inflammation. These observations suggest that the differences in inflammation are not related to systemic difference in lipids and oxidation, but rather to differences in the activity of intracellular inflammatory pathways. In line with this hypothesis, KCs from both hyperlipidemic models still had a foamy appearance after 3 months of HFC diet. Moreover, BMM of Ldlr−/−
mice showed increased expression of Il-6
compared to macrophages from APOE2ki
mice after oxidized LDL loading. Relevantly, IL-6 was shown to dictate the transition from acute to chronic inflammation by changing the nature of leukocyte infiltration (from polymorphonuclear neutrophils to monocyte/macrophages). In addition, IL-6 exerts stimulatory effects on T- and B-cells, thus favoring chronic inflammatory responses 
. Thus, these data suggest that, unlike macrophages from APOE2ki
mice, macrophages of Ldlr−/−
mice produce higher levels of IL-6 and therefore are more prone to develop chronic inflammation. Relevantly, IL6 is also increased in serum of patients with NAFLD and is linked with insulin resistance 
. Gene expression of Cd36
, the main scavenger receptor responsible for the uptake of oxidized cholesterol and involved in inflammatory signal transduction, was also higher in BMM of Ldlr−/−
mice after oxLDL loading compared to APOE2ki and C57Bl6
mice. Interestingly, oxLDL up-regulates CD36 expression via the peroxisome proliferator-activated receptor (PPAR) γ, and this may initiate a feed forward loop of CD36 expression that amplifies the inflammatory response 
. In support of this view, the expression levels of several target genes for PPARγ (ABCG1, LPL, CD36) were also reduced in the livers of APOE2ki mice compared to Ldlr−/−
mice. Thus, it is possible that this loop mechanism does not occur in APOE2ki mice and therefore these mice are less sensitive to the inflammatory cascade.
So far, there is no suitable physiological model for studying the hepatic inflammation in a metabolic context that faithfully recapitulates the pathophysiology of human NASH. In this study, the hepatic inflammatory response induced by prolonged HFC feeding in APOE2ki and Ldlr−/− mice was investigated. We demonstrated that Ldlr−/− mice have increased sensitivity for oxLDL-induced inflammation, apoptosis and fibrosis compared to APOE2ki mice. Therefore, the Ldlr−/− mouse model is particularly useful for understanding the relationships between lipid metabolism and inflammatory recruitment in the context of NASH. This model may therefore be an excellent platform for the assessment of therapeutic strategies for hepatic inflammation.