In this study, we tried to validate and better characterize a previously introduced animal model (FFD mice), reported to simulate the whole spectrum of human NAFLD phenotype [18, 19]. In that model a diet high in saturated fats, cholesterol and fructose, resembling the composition of human “fast food” diet is used in order to simulate nutrition-induced NAFLD. The model was introduced as reproducing not only steatosis, but also NASH and hepatic fibrosis [18]. Its importance is supported by the fact that westernized diet is considered a major pathogenic contributor to the development and progression of NAFLD [2, 24]. Another key point of this model is the addition of fructose, a monosaccharide that is abundant in many processed foods, which is considered to be a major dietary contributor to the pathogenesis of NAFLD, by increasing de novo lipogenesis in hepatocytes, and upregulating tumor necrosis factor (TNF)-α, independently of insulin stimulation [25, 26]. Last but not least, the FFD mouse model is easily reproducible.

Although there are several animal models of NAFLD, in some of them NAFLD is developed without the reproduction of relevant metabolic parameters (e.g., IR, obesity, T2DM), whereas the full phenotypic histological range of NAFLD (steatosis, NASH, fibrosis and/or cirrhosis) is not developed in others. For example, the choline-deficient model leads to NASH and possible fibrosis, but it is also characterized by weight loss and low insulin levels, findings that do not characterize human NAFLD, which is accompanied by abdominal adiposity and high insulin levels (IR) [27]. Another well-known model, the ob/ob (leptin-deficient) mouse model, can hardly develop hepatic fibrosis when fed on CD [28, 29]. On the other hand, the FFD mouse was introduced to reproduce major hallmarks of human NAFLD, both metabolic (IR, obesity and T2DM) and histological, as opposed to other HFD models [18, 30]. As mentioned above, other relevant models, such as HFD, HFHC, HFHF, did not seem to reproduce the full spectrum of NAFLD, and/or they do not fully resemble human western diet [14, 17].

In this study, mice fed FFD developed hepatomegaly with greater liver-to-body weight ratio, as well as greater hepatic steatosis, hepatocellular ballooning and NAS than mice fed CD. Mean hepatocellular ballooning in FFD mice was similar to that of the introductory study of FFD mouse model (1.8 vs. 1.6, respectively) [18], but mean hepatic steatosis was lower in our study (1.0 vs. 2.7, respectively). Indeed, the validation of the development of hepatocellular ballooning, which is a hallmark of the intrahepatic inflammation, in FFD model is important, because it is difficult to develop in rodent models of NAFLD [31]. Furthermore, there was no difference between FFD and CD group in lobular inflammation, in contrast to the introductory study (mean values 1.1 vs. 2.5, respectively) [18], as well as in hepatic fibrosis (0.6 vs. 1.6, respectively). Of note, in our study, hepatic steatosis was not observed in any of the CD mice, while early fibrosis (F1) was observed in only five out of eight FFD mice. Six years after their introductory article [18], the same group conducted a longer-term study (36 weeks) with FFD and CD mice, housing two mice per cage [19], a condition slightly different from the introductory study, in which one mouse was housed per cage [18]. It is also important to highlight that, in their second study, only male mice were used, whereas female and male mice were used in equal numbers in the introductory study. In the second study, steatosis was greater in FFD mice from the first week, while inflammation was greater in FFD group in week 24 and onwards, and fibrosis was greater in FFD group from week 16 and onwards [19]. Contrary to the findings of the introductory study [18] and to those of our study, no evidence of hepatocellular ballooning was reported in that second study [19]. These findings underline the difficulty in reproducing characteristics of NASH and fibrosis in dietary animal models.

Other authors have published on similar dietary models of NAFLD, with varying results. Abe et al. showed greater steatosis and NAS, but they did not observe different inflammation score or hepatocellular ballooning even at week 30 in high-fat, high-fructose and high-cholesterol fed male mice; in their study, fibrosis was greater compared to CD mice at weeks 18 and 30, but not at week 24 [32]. Kristiansen et al. induced greater steatosis, inflammation, NAS and fibrosis, but not hepatocellular ballooning at week 26, in their high-fat, high-fructose and high-cholesterol model [28]. It is important to note, however, that in the studies of Abe and Kristiansen [28, 32], 2% w/w cholesterol was used, which represents a highly excessive amount, while we used 0.2% cholesterol, as corrected by Charlton et al. in 2015 [20]. This constitutes an important correction, since 0.2% rather than 2% simulates better the percentage of cholesterol in human western-type diet. A very high dietary cholesterol intake (1–2%) would radically surpass the daily recommended cholesterol consumption in humans and is, thus, far from a realistic intake [11]. On the other hand, higher dietary cholesterol intake in murine models may be required to reproduce NASH, and especially fibrosis, due to differences in metabolism between humans and mice, as the latter have a limited ability to metabolize cholesterol in the gut [33, 34]. In line with this, a FFD mouse model on high-fat, high-fructose diet and cholesterol developed greater NAS compared to the low-fat control group only when it was fed 2%, but not 0.2% cholesterol [35]. Kim et al. who used FFD with 0.2% cholesterol and approximately 22 g high-fructose corn syrup in the water, reported the development of steatosis and lobular inflammation without, however, evidence of fibrosis [36]. Based on the above, it seems that FFD may be provided for longer than 6 months to reproduce the full histological spectrum of NAFLD in a mouse model; this, however, remains to be investigated. All in all, it is important to note that different studies have used different types of FFD diets, and that creates serious comparability issues when it comes to their results. While it seems that the FFD mouse model resembles human disease to a considerable extent, it does need standardization in order to facilitate the interpretation of results and their comparability to those of other studies.

Regarding the metabolic features of NAFLD, we confirmed that FFD mice gained more weight and had higher glucose, cholesterol, insulin and liver function tests at the endpoint compared to CD mice. Compared with the introductory study by Charlton et al., in which the mean weight of mice at baseline was about 30 g, reaching 44.9 g on week 25 [18], our FFD mice had a mean weigh of 18.1 g at baseline, reaching 31.3 g on week 25 (Table 2); the histological differences observed between these studies may partly be attributed to the difference in baseline weight, while weight gain was similar. Charlton et al. showed higher circulating glucose, AST and cholesterol in FFD group in their introductory study [18], however, the elevations in glucose and AST were not different in their second study at week 24 [19]. Other studies using FFD reported various outcomes concerning biochemical parameters. A steady finding seems to be total cholesterol, which was shown to be increased after FFD in almost all relevant studies [28, 32, 37]. Some [17, 32, 37], but not all studies [28], agree with our results on ALT levels. As far as glucose levels are concerned, the results of different studies showed great variability, showing higher glucose levels in FFD mice at some time points, but not at the endpoints of 24-30 weeks in some studies [17, 28, 32]. TGs were higher in the CD group compared to the FFD group at the endpoint. This seemingly paradoxical finding was also observed by other authors [19, 32]. Hyperinsulinemia and IR were also observed in other NAFLD models, as well as in NASH patients even in the absence of obesity [18, 36, 38]. These results further emphasize the need for standardization of the FFD model, so as to lead to more reproducible results in relevant metabolic parameters, in line with the histological ones.

Lack of physical activity and a sedentary lifestyle promote the development and progression of NAFLD [39]. To enhance sedentary pattern, Charlton et al. used individual cages for mice (one mouse per cage) in the introductory study [18] and two mice per cage in their second study [19], whereas we placed four mice per cage, so as to restrict excessive movements without limiting the interaction with other mice and preventing stress associated with single housing. However, it would be useful to compare the size of the cages and the exact mean space given for each mouse, because it may probably account, at least partly, for the observed differences in different studies [18, 19].

Food consumption was greater in CD mice and stable during the study in this group, whereas FFD mice increased food consumption after the middle of the study, when they gained weight, more than the CD mice. On the contrary, water consumption was greater in FFD group, probably owing to the addition of fructose and glucose in the water, which may have made it more palatable to the mice. Caloric intake was not recorded in the introductory study [18]. The amount of food and water consumption is commonly not reported in animal studies, despite being critical variables, especially in studies in which nutrition and subsequent weight changes may have affected metabolic parameters or dosage of medications. Other C57BL/6 mice studies also reported lower food consumption in HFD group than in CD group [32, 40]. Interestingly, it has been indicated that sucrose-sweetened soft drinks reduce energy intake, thus increasing the possibility of a positive energy balance [41]. Of note, there is evidence demonstrating a negative effect of free fructose consumption on central appetite [42]. These data may explain the lower food intake at the first half of the study in FFD mice.

It should be highlighted that, with the exception of the introductory study of Charlton et al. [18], all relevant FFD studies used exclusively male mice. In our study, mice of both sexes were used to show differences in food and water consumption between male and female mice, which may partly explain a later weight gain in female compared to male mice (Fig. 2b, c, Table 2). This may have led to minimal biochemical and histological differences between subgroups of male and female mice (Table 3), which, however, could not explain the male predominance in the relevant studies. In humans, NAFLD prevalence is lower in premenopausal women compared to men of similar age, but after menopause, the prevalence of NAFLD becomes almost similar between men and women; this may be related to metabolic alterations following the abrupt decrease in estrogens in women after menopause [43,44,45]. These disparities between sexes have been demonstrated in some murine models confirming that female animals present with milder hepatic steatosis, inflammation and even fibrosis than male ones [46, 47]. The majority of preclinical studies use male mice, thus female mice are understudied, which may be misleading when the results of studies from male mice are extrapolated to female ones. In this regard, the effect of FFD on NAFLD in a longer-term, i.e., after menopause of female mice could be important. Alternatively, one could study the effect of FFD diet on NAFLD in ovariectomized female mice. We should also highlight that the dissemination of the negative results of studies is important to guide other researchers in the field in a global basis, so as not to invest their effort, time and resources on topics having previously provided negative results.

Last but not least, we could not overlook the recent change in the nomenclature of NAFLD to metabolic dysfunction-associated steatotic liver disease (MASLD) proposed by a multinational consensus [48]. This proposal is supported by the ascertainment that the use of “nonalcoholic” overemphasizes the absence of alcohol, whereas at the same time it underemphasizes the significance of multiple metabolic factors contributing to the development and progression of NAFLD [49]. The diagnosis of MASLD is based on definite criteria, i.e., hepatic steatosis along with at least one out of five metabolic risk factors, which may better reflect the pathophysiology of this highly heterogeneous disease and may facilitate research in the field with future clinical implications. However, in our opinion, the terms of NAFLD and MASLD, as well as NASH and metabolic dysfunction-associated steatohepatitis (MASH), may not be used invariably, since they are similar but not synonymous entities, until more evidence clarifies this issue.

This study has certain strengths and limitations. To our knowledge, this is the first study aiming to validate the model introduced by Charlton et al. [18], a western-type diet with 0.2% cholesterol and fructose in C57BL/6 J mice that simulates the fast food human diet, using both male and female mice. However, certain limitations of this study are: (1) the sample size was relatively small, especially for the analyses of subgroups stratified by sex (four mice per group), so the relevant findings should be cautiously interpreted, because they may be underpowered, e.g., findings on lobular inflammation; (2) we could not house one mouse per cage due to the induced stress that may have affected the outcomes of the study; (3) non-fasting blood sampling may have distorted blood tests results, e.g., glucose and TGs.

In conclusion, this study did not fully reproduce all elements of NAFLD in FFD male and female mice: hepatic steatosis was developed to a lower grade, and lobular inflammation and fibrosis were not developed so extensively as in the introductory study [18]. However, we could not overlook the development of hepatocellular ballooning and the association of histological evolution of NASH with morphological and metabolic changes characteristic of NAFLD. A longer duration of the study might have led to greater steatosis, lobular inflammation and fibrosis, but it remains to be shown. The FFD mouse model hold hope for the future of NAFLD research, but, until then, the diet should be standardized and some relevant issues (e.g., mice per cage, the selection of male and/or female mice) should be clearly determined.