Early-onset caloric restriction alleviates ageing-associated steatohepatitis in male mice via restoring mitochondrial homeostasis

Non-alcoholic fatty liver disease is associated with ageing, and impaired mitochondrial homeostasis is the main cause for hepatic ageing. Caloric restriction (CR) is a promising therapeutic approach for fatty liver. The purpose of the present study was to investigate the possibility of early-onset CR in decelerating the progression of ageing-related steatohepatitis. The putative mechanism associated with mitochondria was further determined. C57BL/6 male mice at 8 weeks of age were randomly assigned to one of three treatments: Young-AL (AL, ad libitum), Aged-AL, or Aged-CR (60% intake of AL). Mice were sacrificed when they were 7 months old (Young) or 20 months old (Aged). Aged-AL mice displayed the greatest body weight, liver weight, and liver relative weight among treatments. Steatosis, lipid peroxidation, inflammation, and fibrosis coexisted in the aged liver. Mega mitochondria with short, randomly organized crista were noticed in the aged liver. The CR ameliorated these unfavourable outcomes. The level of hepatic ATP decreased with ageing, but this was reversed by CR. Ageing caused a decrease in mitochondrial-related protein expressions of respiratory chain complexes (NDUFB8 and SDHB) and fission (DRP1), but an increase in proteins related to mitochondrial biogenesis (TFAM), and fusion (MFN2). CR reversed the expression of these proteins in the aged liver. Both Aged-CR and Young-AL revealed a comparable pattern of protein expression. To summarize, this study demonstrated the potential of early-onset CR in preventing ageing-associated steatohepatitis, and maintaining mitochondrial functions may contribute to CR’s protection during hepatic ageing.


Introduction
Ageing liver is accompanied by physiological changes, including metabolic and hormonal alterations (Chung 2021;Gong et al. 2017;Hunt et al. 2019;Ogrodnik et al. 2017). During hepatic ageing, a reduction in fatty acid oxidation and an increase in de novo lipogenesis are noticed (Chung 2021;Ogrodnik et al. 2017). In addition, a loss of insulin sensitivity is observed. These metabolic changes contribute to lipid accumulation in the liver. A decline in growth hormone/ IGF-1 signalling and an increase in insulin and leptin resistance with age, not only deregulate lipid metabolism but also induce inflammation and fibrogenesis, thereby increasing the progression of hepatosteatosis (Chung 2021;Gong et al. 2017).
Mitochondria have critical roles in energy supply and cellular functions, such as calcium homeostasis and apoptosis in the whole body (Lopez-Lluch 2017). The integrities of quality and quantity in mitochondria (biogenesis, fission, fusion, and mitophagy) maintain healthy mitochondria. The connection between hepatic ageing and mitochondrial dysfunctions are reviewed elsewhere (Hunt et al. 2019;Lopez-Lluch 2017). Age-dependent structural changes in the cristae and inner membrane of the mitochondria are observed, and mitochondria are often enlarged in the aged cells (Hunt et al. 2019). In addition, numerous functional defects in mitochondria are noticed in the aged liver, including decreased mitochondrial biogenesis and mitophagy, dissociation of ATP synthase, increased accumulation of reactive oxygen species (ROS), and impairment of respiratory chain complexes (Lopez-Lluch 2017; Wan et al. 2020). Approaches to strengthening mitochondrial functions could be therapeutic targets for hepatic steatosis and ageing (Hernandez-Alvarez et al. 2019;Wan et al. 2020).
Dietary intervention has been applied in relief of hepatic steatosis, and caloric restriction (CR) is a promising approach to improve fatty liver via restoration of insulin sensitivity to decrease lipid accumulation and reduce inflammation in liver (Fujii et al. 2017(Fujii et al. , 2019Gong et al. 2017;Kobayashi et al. 2020). Although CR intervention has been studied extensively, the conditions of CR, such as initial timing for CR and frequency and duration of CR are varied among studies. These variations cause inconsistent effects of CR on recovery from hepatic steatosis. Moreover, whether early-onset CR ameliorates hepatic ageing, especially steatohepatitis is still unclear. The present study aimed to investigate the effect of early-onset CR on decelerating the progression of ageing-related steatohepatitis. Potential mechanisms regarding mitochondria were evaluated.

Animals and experiment diets
The Institutional Animal Care and Use Committee of the National Taiwan University College of Medicine and College of Public Health approved all animal care procedures used in this study (IACUC no: 20170321). Male C57BL/6 mice with 8 weeks old (n = 21) were obtained from the National Taiwan University College of Medicine Laboratory Animal Centre and randomly assigned into three groups, Young-AL, Aged-AL, and Aged-CR. The AL mice were fed the commercial diet (5058, LabDiet, St. Louis, MO, USA) ad libitum (3.454 g/day as determined in the ad libitum group immediately before start of the experiment), whereas the CR mice were received a limited dietary intervention (60% intake of ad libitum, 2.072 g/day). In order to achieve 60% AL in CR mice, a stepdown procedure was carried out over the course of 3 weeks. The CR group had access to food for 4 h starting from the onset of the dark active phase. The commercial diet contains: protein 21.8%, fat 9.1%, crude fibre 2.2%, nitrogenfree extract 51.8%, ash 5%, and 3.56 kcal/g of metabolic energy (Johnson-Henry et al. 2014). There was unlimited access to water, and monthly body weight was recorded. Mice were housed in an animal room with a controlled temperature of 22-24 °C, humidity of 50-55%, and a light/dark cycle of 12 h. The lights were turned off at 7:00 p.m. and back on at 7:00 a.m. each day. Young mice at 7 months old (Young) and old mice at 20 months old (Aged) were sacrificed by CO 2 , respectively. The samples for TEM and histology were in situ transcardial perfused prior to tissue collection (n = 2), whereas fresh tissue samples without prior fixation were excised, frozen in liquid nitrogen, and stored at − 80 °C until biochemical and western blotting analyses later (n = 5).

Histological analysis
Segments of liver were fixed in a 10% neutral formalin solution for a week, embedded in paraffin, and sectioned at 5 μm. The staining with hematoxylin and eosin (H&E) was used to monitor the morphology of the liver sample. The degree of hepatic fibrosis was determined by using Masson's trichrome stain (Li et al. 2019b).

ATP content
The ATP level in liver was determined using the ATP Colorimetric Assay (Biovision, Milpitas, CA, USA) following the procedure of Chien et al. (2022). Initially, ATP Assay Buffer was used to lyse the tissue, and the tissue lysate was centrifuged to collect the supernatant fraction. A 10 kDa spin column (ab93349, Abcam, Cambridge, UK) was used to deproteinize the supernate, and the deproteinized sample was collected and mixed with the ATP reaction mix. After incubation at 37 °C in darkness for 30 min, absorbance of the sample was measured at 570 nm in a spectrophotometer (CLARIOstar, BMG LABTECH, Ortenberg, Germany). A standard curve was set to calculate the ATP concentration by applying the absorbance of the ATP.
Analysis of lipid peroxidation, triacylglycerol, and cytokines Malondialdehyde (MDA), a product of lipid peroxidation, was analysed by measurement of thiobarbituric acid reactive substances (Li et al. 2019b). The MDA was isolated from the liver using RIPA lysis buffer (20-188, Millipore, Burlington, MA, USA) and centrifuged. The thiobarbituric acid solution was added to the supernatant fraction, followed by the addition of the trichloroacetic acid-HCl reagent. Following boiling, cooling, and centrifuging, the absorbance of the supernatant fraction was measured at 535 nm using microplate assays.
The hepatic triacylglycerol content was extracted using 5% NP-40 according to the manufacturer's instructions. Following three cycles of boiling and cooling to room temperature, the tissue solution was centrifuged at 14,000 × g for 10 min. The supernate was used for the triacylglycerol assay (Fortress Diagnostics, Antrim, Northern Ireland, UK).
The cytokine content of the liver sample was extracted using RIPA lysis buffer and centrifuged. The supernatant fraction was used for cytokine assay by commercial kits: TNFα (DY690B, R & D Systems, Minneapolis, NM) and IL-6 (DY686, R & D Systems). All procedures were according to the manufacturer's instructions.

Transmission electron microscopy (TEM)
The liver samples were gently washed with saline, dissected into 1mm 3 cubes, and then fixed for 24 h at 4 °C with a fixative solution of 4% paraformaldehyde and 2.5% glutaraldehyde in a 0.1 M sodium cacodylate buffer, pH 7.4. The fixed material was ultrathinly sliced (70-90 nm) and put on nickel grids. The slice was stained with saturated aqueous uranyl acetate, followed by contrast staining with Reynold's lead citrate. Organelle ultrastructure was examined using a FEI Tecnai G2 F20 S-Twin transmission electron microscope (FEI Co., Hillsboro, OR, USA).

Western blotting
Proteins of liver sample were extracted using RIPA assay buffer. Following centrifuging at 14,000×g for 10 min at 4 °C, the supernate (20 μg proteins) was applied to a sodium dodecyl sulfate/polyacrylamide gel (10 or 15% acrylamide) for electrophoresis. A PVDF membrane (Millipore) was used for protein transfer by electroblotting at 300 mA for 2 h at 4 °C. Following transfer, the proteins on the membrane was hybridised with the target protein antibodies by incubating at 4 °C ovrenight. The protein antibodies used in this study are mitochondrial transcription factor A (TFAM: gwb-22c6c2: GenWay Biotech, San Diego, CA, USA), Complex II subunit (SDHB: ab14714: Abcam, Cambridge, UK), Complex I subunit (NDUFB8: ab110242), dynamin-related protein 1 (DRP1: #5391: Cell Signalling Technology, Hertfordshire, UK), mitofusin-2 (MFN2) (#9482), and voltage-dependent anion channel (VDAC: #4866). The abundance of proteins was determined by the enhanced chemiluminescence Western blotting detection system kit (Amersham Pharmacia Biotech). The immunoblotting was visualized and quantified using the ChemiDocTM Touch Image System (Bio-Rad, Hercules, CA, USA).

Statistical analysis
The data was presented as mean±standard error (SE). The results were analysed by one-way ANOVA followed by Dunnett's multiple comparison test. Statistical analyses were performed using SAS (version 9.4; SAS Institute Inc., Cary, NC, USA). P ≤ 0.05 indicated a statistically significant difference, and P ≤ 0.1 indicated a tendency towards significance.

Results
Steatohepatitis was accompanied with ageing, and CR reduced the degree of steatohepatitis To investigate whether early-onset CR maintained the liver functions of aged mice, CR treatment was applied to male mice when they were 8 weeks old until they were 20 (Aged) months old. Aged-AL mice exhibited a greater body weight, whereas early-onset CR maintained a smaller body weight at this age (Fig. 1A). Aging caused an increase in liver weight, whereas CR reversed the liver weight of aged mice (Fig. 1B). The liver weight was adjusted with the body weight to further elucidate whether liver weight accompanied the pattern of body weight (Fig. 1C). When Compared to the Young-AL mice (7 months old), ageing caused a greater relative liver weight. An equal relative liver weight was found between Aged-CR and Young-AL mice.
Staining with H&E was applied to assess hepatic morphology (Fig. 2). Ageing caused more lipid accumulation and white blood cell infiltration in the liver than young mice, whereas CR resulted in liver morphology similar to young animals. Masson's trichrome stain was used to investigate the extent of hepatic fibrosis. An extensive degree of fibrosis was found in the liver of Aged-AL mice. In contrast, only a small area of fibrosis was displayed in the liver of Aged-CR mice.
Because morphological results implied that ageing caused lipid accumulation in the liver, the triacylglycerol content was quantified (Fig. 3A). Ageing caused almost a two-fold increase in lipid accumulation in the liver, and early-onset CR maintained similar lipid content in the aged liver as observed in the young mice. The MDA, a product of lipid peroxidation, was analysed to investigate the possibility of ageing-induced oxidative stress (Fig. 3B). Compared to the young liver, ageing tended to increase the hepatic content of MDA (p = 0.09). The MDA was suppressed in the aged CR-fed mice. Interleukin 6 and TNFα were determined as markers of inflammation in the liver (Fig. 3C, D). Aged-AL mice displayed the highest level of IL6 in the liver compared to other groups of mice. The TNFα levels in the liver of Young-AL and Aged-CR mice were lower than in the liver of Aged-AL mice. Taken together, we found that hypertrophy, lipid accumulation, inflammation and fibrosis coexisted in the aged liver, indicating that steatohepatitis accompanied hepatic ageing. The CR treatment ameliorated the severity of steatohepatitis in the aged liver.

Mitochondrial dysfunctions accompanied ageing, and CR reversed them
The ultrastructures of organelles were studied by TEM (Fig. 4A). There were numerous ultrastructural  abnormalities in the aged liver, including large lipid droplets (yellow arrows), irregular and enlarged endoplasmic reticulum (green arrows), and mega mitochondria (blue arrows). In addition, short, randomly organized cristae were observed in swollen mitochondria. The Aged-CR liver displayed fewer lipid droplets, and their mitochondria were smaller with laminar cristae than observed in the Aged-AL animals. More lysosomes (red arrows) were noticed in the Young-AL and Aged-CR livers than in the Aged-AL liver.
The primary function of mitochondria is to provide energy to cells, and abnormal mitochondrial morphology indicated an energy deficit. The content of ATP was determined to clarify the possibility of an energy deficit in the aged liver. Ageing led to a decrease in ATP levels, whereas the CR diet reversed the energy deficit in aged animals (Aged-CR) (Fig. 4B).

Mitochondrial-related protein expressions in the liver
Abnormal mitochondrial morphology in the aged liver suggested an imbalance in mitochondrial dynamics (biogenesis, fission, and fusion) during ageing, and proteins involved in mitochondrial dynamics were investigated to determine their contributions. The TFAM, as a marker of mitochondrial biogenesis, was upregulated in the Aged-AL liver, whereas . All results are expressed as mean ± SEM. n = 3-5. *, p < 0.05). TFAM mitochondrial transcription factor A, DRP1 dynaminrelated protein 1, MFN2 Mitofusin 2, NDUFB8 Complex I subunit, SDHB Complex II subunit, VDAC voltage-dependent anion channel a similar expression of hepatic TFAM was found between Young-AL and Aged-CR mice (Fig. 5A). In addition, Aged-CR mice displayed greater expressions of PGC-1α and SIRT1 in the liver than Aged-AL did (supplementary Fig. 1). The DRP1 and MFN2, were markers for mitochondrial fission and fusion, respectively. Drp1 expression decreases with age, while MFN2 expression increases. Feeding the CR diet reversed the expression of both proteins during ageing.
Because the CR diet reversed the energy deficit in the aged animals, this indicated the potential regulation of the mitochondrial oxidative phosphorylation system, and protein expressions of respiratory chain complexes were analysed. The NDUFB8 (subunit of NADH dehydrogenase) and SDHB (subunit of succinate dehydrogenase) are markers of complex I and II in the respiratory chain, respectively. Compared with the Young-AL mice, a lower expression of NDUFB8 and SDHB was found in the Aged-AL mice (Fig. 5B). An upregulation of SDHB but not NDUFB8 was observed in the Aged-CR liver.

Discussion
At present, reducing lipid accumulation, oxidative stress, inflammation, and fibrosis in the liver are the main approaches for innovative non-alcoholic steatohepatitis (NASH) therapies (Sumida and Yoneda 2018). This study established an ageing-related NASH mouse model in which hypertrophy, lipid accumulation, inflammation, oxidative stress, and fibrosis coexisted in the liver of aged mouse. Earlyonset CR successfully ameliorated the progression of ageing-related NASH, indicating its potential as a therapeutic approach.

Ageing-related NASH
The prevalence of non-alcoholic fatty liver disease (NAFLD) and NASH is frequently associated with age (Berardo et al. 2020). The major contributors to NASH during ageing could be via systemic coupled with local mechanisms (Chung 2021;Gong et al. 2017;Hunt et al. 2019;Ogrodnik et al. 2017). Adipose tissues, regulating whole-body energy metabolism via adipokine secretion and fat storage, participate in systemic regulation (Chung 2021). An increase in body weight and adiposity accompanies ageing. Redistribution of adipose tissue during ageing alters the adipokine profile and induces leptin/insulin resistance (Gong et al. 2017). Inflammatory cytokines from visceral adipose tissue further impair liver functions via the circulation. These changes in white adipose tissue target lipids to the liver, and cause a systemic effect on deterioration of normal liver functions during ageing (Chung 2021;Gong et al. 2017). The cellular senescence brought on by aging has local effects on liver function. Senescent hepatocytes have altered lipid and carbohydrate metabolism, increased lipid accumulation (elevated de novo lipogenesis and suppressed fatty acid oxidation), reduced mitochondrial functions, increased release of inflammatory cytokines, and increased ROS production (Hunt et al. 2019;Ogrodnik et al. 2017). These systemic and local changes together contribute to the progression of NAFLD and NASH during ageing.
The importance of mitochondria in regulating liver health has been extensively studied (Hunt et al. 2019;Ogrodnik et al. 2017). Defects in mitochondrial fission and fusion have the potential to cause NASH/ NAFLD and liver cancer, whereas, activating mitochondrial dynamics attenuates liver diseases (Hernandez-Alvarez et al. 2019; Li et al. 2019a;Takeichi et al. 2021). Recently, strategies to strengthen mitochondrial functions have been applied to reduce hepatosteatosis and promote liver health; benefits on the health of aged liver have been also observed (Martinez-Cisuelo et al. 2016;Shabalina et al. 2017;Wan et al. 2020). Those strategies include exercise, chemical addition to diets, and variation in diet composition and administration (Chimienti et al. 2021;Chuang et al. 2020;Martinez-Cisuelo et al. 2016;Shabalina et al. 2017).
CR reduced ageing-related NASH/ NAFLD Caloric restriction, one of the dietary interventions, is able to reverse the negative effects of obesity on inflammation, oxidative stress, and lipid content in the liver (Izquierdo et al. 2020;Talavera-Urquijo et al. 2018). It is evident that CR exhibits hepatic protection during ageing. Nevertheless, the mechanisms of CR in the aged liver are not well identified. Multiple physiological and metabolic changes induced by CR have been reported, and several signalling pathways are proposed as mediators of CR actions (Chimienti et al. 2021;Izquierdo et al. 2020).
Caloric restriction targets the metabolic remodelling of white adipose tissue by increasing fatty acid synthesis, enhancing mitochondrial biogenesis, and antioxidant enzyme activity, thus reducing ageingrelated hepatosteatosis (Fujii et al. 2017(Fujii et al. , 2019Kobayashi et al. 2020). Insulin resistance and glucose intolerance accompany ageing and are recovered by CR (Velingkaar et al. 2020). In addition to systemic regulation, CR directly regulates liver functions. A decline in mitochondrial functions and dysregulated nutrient sensing are observed in the aged liver (Hunt et al. 2019), whereas CR promotes hepatic nutrient sensing via AMPK, mTOR, and Sirtuin pathways (Lopez-Lluch 2017; Tulsian et al. 2018;Velingkaar et al. 2020). Our findings also showed that early-onset CR maintained mitochondrial homeostasis and recovered an energy deficit, thereby alleviating age-related NASH.
Nonetheless, not all CR studies show a beneficial effect on NAFLD and NASH. Perhaps these results come from the great variation in design of CR studies. Caloric restriction can be accomplished by lowering intake or energy level or by changing dietary composition (from a high-fat diet to a low-fat diet). Moreover, the differences in duration and timing of CR, and the species, sex, and age used might further affect the results of CR experiments. Wang et al. (2011) found that both a high-fat diet and a low-fat diet (the "normal diet"), as the CR source, cause the loss of body weight through discordant regulatory pathways. A low-fat diet alleviates hepatic inflammation and steatosis, whereas a high-fat diet provides greater amelioration of inflammation of adipose tissues (Wang et al. 2011). Similarly, feeding a normal diet to dietary-induced obese mice decreases body weight and hepatic inflammation (Dohmen et al. 2020). A breeding diet used in the present study had a protective effect of CR on alleviating ageing-related NASH.
Restricting eating time in the day without limiting caloric intake has been researched extensively. Time-restricted eating significantly reduces liver steatosis and improves glucose tolerance in obese mice (Chaix et al. 2021), and extends the lifespan of Drosophila and mice (Mitchell et al. 2019;Ulgherait et al. 2021). A human study also shows the benefits of time-restricted eating, including weight loss, increasing insulin sensitivity, and reducing metabolic disorders (Swiatkiewicz et al. 2021). Our study fed the CR mice at a 4-h interval during the dark active phase, which is comparable to time-restricted eating but limited the caloric intake. The hepatic protection provided by CR in this study could be attributed to the combination of restricted eating time and calorie intake, particularly over the long term.
Four-weeks of CR treatment in mice with NASH caused a reduction in body weight and alleviated inflammation but had no effect on the steatosis/ NASH score (Poekes et al. 2019). The CR for more than 3 months leads to a decrease in lipid accumulation in the liver (Ogrodnik et al. 2017), and CR ameliorates lipid accumulation even when applied in middle age (Fernandez et al. 2019;Ogrodnik et al. 2017). Our study extended the CR duration to 18 months (20 months old) and showed a significant decline in NASH. However, there seems to be an age-related boundary of CR efficacy because extending CR to 32 months in rats did not reduce NASH (Chimienti et al. 2021).
The ageing effect is sex-dependent; male mice exhibit a metabolic memory of early-life CR, thus own a better ability to control lipid and carbohydrate metabolism in later life, whereas female mice did not display the metabolic memory (Cameron et al. 2012;Selman and Hempenstall 2012). According to Astafev et al. (2017), the IGF-1 signalling pathway and circadian rhythms, responsible for energy metabolism in the liver, have sexual dimorphism when treated with CR treatment. Shortterm CR fails to overcome sexual dimorphism in mitochondrial functionalities and antioxidant capacity of liver (Valle et al. 2007). In the liver of mice with a 10-month CR, a difference in metabolomic profiles is also observed between the sexes (Gibbs et al. 2018). Long-term CR perturbs the reproductive system in females but not in males; furthermore, CR elevates the plasma levels of testosterone and corticosterone in female mice but not in male mice (Martin et al. 2007). These changes in endocrines caused by CR treatment might further alter liver metabolism. This study demonstrated the protective effects of CR on male liver during ageing, however, the exact effect of CR on female mice during hepatic ageing is unknown, and requires additional research.
The potential mechanisms involved in CR regulation Noticeably, this study provided novel mediators for CR protection on easing ageing-related steatohepatitis. Our findings suggest that during hepatic ageing, some compensatory mechanisms of mitochondrial functions were activated, including increased biogenesis (TFAM) and fusion (MFN2), and reduced fission (DRP1); however, these actions were insufficient to rescue the mitochondrial dysfunctions, and ended with an energy deficit. Early-onset CR maintained the mitochondrial dynamics in the aged liver as observed in the young liver.
Both Chimienti et al. (2021) and Picca et al. (2013) investigated the protective effects of early-onset CR on hepatic ageing using experimental designs similar to our study. They found that CR elevates mitochondrial biogenesis and suppresses oxidative damage to mitochondria. Interestingly, their data show a decline in the expression of TFAM, DRP1 and MFN2 during ageing, and CR reversed the decline in expressions of these proteins. These patterns regulated by ageing and CR are different from ours; they used rats at 18 and 28 months of age, whereas we used mice at 7 and 20 months of age.
To summarize, our study established an ageingrelated NASH mouse model. Early-onset CR reversed the negative effects of ageing associated NASH, including oxidative stress, inflammation, steatosis, and fibrosis. The CR treatment had recovered the energy deficit in the aged liver, presumably via maintenance of mitochondrial homeostasis.

Implications of this study
60% AL is a common CR intervention in animal models due to its promising health benefits and lifespan extension (Mitchell et al. 2016). This work utilized the same intervention and commenced early-onset CR due to the increased lifespan effect in early life (Lee et al. 2021). However, such a long-term intervention is difficult to adhere to when translating for clinical application, which limits the scope of this study. Nevertheless, the molecular mechanisms of CR's protection found in this study could be an alternative target for the development of CR mimetics in the future. Moreover, the combination of CR with other treatments such as exercise (Izquierdo et al. 2020), surgery (sleeve gastrectomy) (Talavera-Urquijo et al. 2018), and chemicals/ mimetics (Gibbs et al. 2018;Poekes et al. 2019) display a more effective effect on hepatosteatosis and lifespan, making it a more practical option when translating to clinical application.