The present data confirmed our previous report that MS-NASH mice possess all features of metabolic disorders [18, 19] and NAFLD/NASH when on WDF , however, with a relatively long induction duration (~ 20 weeks), and moderate liver fibrosis (pathology score ~ 1). Low dose CCl4 (0.08 mL/kg) accelerated the progression of NASH in ~ 8 weeks and exacerbated liver fibrosis by raising the pathology score to ~ 4, but not significantly affecting other features of liver pathology measured by NAS in MS-NASH mice on WDF. Consistent with preclinical and clinical reports [16, 29–31], OCA treatment reduced NASH pathology in this model and surpassed the therapeutic effects on previous WDF only model . Thus, MS-NASH mice on WDF and CCl4 is an appropriate translational model for testing novel therapeutics targeting NASH/NAFLD.
The pathogenesis of NAFLD/NASH is initiated by systemic dysmetabolism, leading to lipid accumulation, hepatosteatosis, hepatocyte ballooning , inflammatory cell infiltration, increase in oxidative stress [5, 6], hepatocytes injury, fibrosis, etc. [7–9]. CCl4 is a known liver toxin directly causing hepatocyte injury, leading to liver fibrosis, cirrhosis and carcinoma, which often requires relatively high doses from 0.2 to 5 mL/kg . In order to avoid aggressive direct hepatocyte injury overshadowing development of hepatosteatosis, a critical component of NAFLD/NASH in MS-NASH mice, CCl4 at 0.2 mL/kg was selected in the first experiment, which, however, caused excessive animal death in the first 3 weeks, probably due to acute liver toxicity evidenced by significant weight loss, reduction of food and caloric intake as well as massive elevation of ALT and AST, but did not significantly affect liver pathology measured by steatosis, ballooning and overall NAS or fibrosis score. When the dose of CCl4 was further reduced to 0.08 mL/kg, although the animals still lost weight, most survived with ALT and AST levels not as high as that in high dose CCl4 group.
Although several noninvasive imaging methods have been used to assist diagnoses of NAFLD/NASH in clinical and preclinical research [32–35], histopathological examination is still the gold standard, especially to differentiate NASH from simple steatosis [28, 36, 37]. In preclinical research with rodent models, postmortem histopathological examination of liver tissue is still a commonly used method, in which the NAFLD Activity Score (NAS) is semi quantitatively evaluated by pathologists for assessment of NAFLD to distinguish steatosis from NASH . NAS provides a composite score based on the degree of steatosis, lobular inflammation, and hepatocyte ballooning with a score < 2 unlikely and ≥ 5 likely representing NASH, respectively. The present liver histopathology in MS-NASH mice on WDF exhibited persisting macrovesicular steatosis, hepatocyte ballooning degeneration, and inflammatory cell infiltration with the NAS scores ~ 5, thus, it qualifies as a NSAH model (Figs. 2 & 4). The NAS does not include fibrosis score, with the latter being reported separately on a scale from 0 (without fibrosis) to 4 (cirrhosis) , which does not always correlate with each other . In the present experiment, the fibrosis scores were relatively low, only ~ 1.2 in MS-NASH mice for a total duration of 16 weeks on WDF, which, however, was raised to ~ 3 by combination of low dose CCl4 for 8 weeks (Figs. 4 & 6), but not by high dose CCl4 for 3 weeks (Fig. 2h). We suspect a longer duration may be needed for fibrosis development to the level detectable by the pathology scoring system due to its low sensitivity and resolution. The Pathology fibrosis score criteria from 1 to 3 also depends on zonal distribution of fibrotic findings, with a score of 1 being periportal fibrosis; a score of 2 being periportal and perisinusoidal fibrosis; and a score of 3 being bridging fibrosis between multiple fibrotic areas. MS-NASH mice on WDF develop perilobular fibrosis initially, while CCl4-induced liver fibrosis exhibits centrilobular distribution characteristics. As showed in the present results, a correct combination of these 2 insults (WDF + CCl4) can yield aggravated fibrosis at multiple zones and accelerated bridging between areas.
Furthermore, the semiquantitative pathology scoring system with manually slide reading by pathologists is not only time-consuming and labor-intensive, but also subjective with person to person deviation, and lacks resolution and sensitivity to detect subtle difference in different animal models, different disease stages, and subtle changes by therapeutic intervention. With development of computerized imaging analysis and machine learning technology, commercial software is now available for automatic quantification of histopathology images, including liver pathology for NASH research [31, 38, 39]. As shown in Fig. 2h, the semi-quantitative fibrosis score ~ 1 may be too low to distinguish any differences among treatment groups. However, quantitative image analysis is able to measure the relative area of fibrosis over the entire section as ~ 2% in the control group, and increased to ~ 8% by high (Fig. 2h) or low (Fig. 4f & 6 k) dose CCl4 in MS-NASH mice on WDF. Furthermore, the present data also showed high correlation between a computer quantification and the pathology scores for steatosis, ballooning, inflammation and fibrosis (Fig. 8). Thus, the computerized image analysis is a valid method that is more efficient and consistent, providing higher sensitive and less subjective quantification of histopathology changes in NASH research.
Farnesoid X receptors like OCA has been used in preclinical treatment for diet  or chemical  induced liver fibrosis and NASH. The present data demonstrated a more robust efficacy of OCA treatment in MS-NASH mice on both WDF and CCl4 compared to those on WDF alone , which might be due to the model with a more robust liver fibrosis or OCA with dual alleviation to both diet/chemical-induction.
Mitochondrial metabolism dysregulation has been implicated in NALFD pathogenesis and progression with reduced capacity to compensate for increased oxidative stress, a key factor in hepatic injury and fibrosis, although the precise disease etiology remains to be further investigated . Accumulating evidence suggests that therapeutically targeting the mechanisms leading to mitochondrial dysfunction may have therapeutic benefits to patients with liver disease [42–45]. The present survival analysis data showed that MS-NASH mice better tolerated CCl4 associated mortality compared to C57Bl/6 mice, which could be attributed to higher catalase activity to oxidative stress, thus, reducing oxidative DNA damage in MS-NASH mice reported by Boland et. al .
The present data also demonstrated that the degree of NASH pathology measured by both NAS and fibrosis scores from the pathologist and computer quantification appeared to be higher in MS-NASH than C57Bl/6 mice on WDF and CCl4, indicating that C57Bl/6 mice may require longer NASH induction time and have less hepatocyte ballooning degeneration, consistent with the view that MS-NASH mouse is a superior NASH model with more prominent hepatosteatosis pathology and metabolic disorders.