In the present study, we sought to investigate the effects of lifelong moderate (15% initiated at 4 months of age) CR on hepatic lipid accumulation, glucose homeostasis, and physical function at timepoints that include stages representative of middle to advanced age (10-, 18-, 26, and 28-months) in mice. A more severe (30–40%) level of restriction has been the focus of most studies that aim to assess CR’s effect on metabolic function in animal models of aging. Still, few studies have investigated the metabolic impact of this restriction in advanced age. Lifelong 30% CR decreases liver fat, improves glucose clearance, and lowers circulating insulin levels in 12-month-old mice (Rusli et al. 2015). 40% CR initiated at 3 months decreases liver fat at 12- and 15-months of age (Ogrodnik et al. 2017). Similarly, 74 weeks of 40% CR (fed at 8 A.M.) decreases liver fat at 19 months of age (Kuhla et al. 2014).
We found that a modest lifelong 15% CR decreased body mass and fat mass in 10- and 18-month-old mice, but not at 26 and 28 months of age (advanced age). Our observation that CR did not further decrease body weight or fat mass compared to age-matched ad libitum fed mice is likely due to the decrease in body weight commonly seen during advanced age in ad libitum fed C57Bl/6 mice, preventing us from observing a further decrease in body weight or fat mass. Supporting these findings, Turturro and colleagues also observed an age-related decrease in body mass beginning at approximately 25 months of age, which continued to decline throughout advanced age (Turturro et al. 2002).
CR decreased liver triglyceride content in 10- and 18-month-old mice and, accordingly, improved glucose tolerance at these ages. In advanced aged (26 and 28 months) mice, liver fat was decreased compared to 10- and 18-month-old mice, and CR did not further decrease liver triglyceride, liver NEFA, or affect glucose clearance at these ages (Figs. 3A, B and 4I). Despite the key role of hepatic lipid content on metabolic flux, few studies have extended beyond 22 months of age to investigate the effect of aging on hepatic lipid concentration. Fontana and colleagues showed that liver triglyceride content was similar in 6-, 12-, and 22-month-old C57BL/6 mice fed either a low-fat chow or high fat diet (Fontana et al. 2013). Our observation that hepatic lipid accumulation decreases in advanced age is likely secondary to the age-related decrease in body weight and fat mass. The age dependent decrease in hepatic lipid accumulation in 26- and 28-month-old mice likely prevents us from observing any benefits from CR, a treatment aiming to decrease liver lipid and improve metabolic health.
Given the decrease in liver triglyceride content that we observed with both calorie restriction and aging, we performed qPCR to evaluate the mRNA expression of two crucial enzymes in the de novo lipogenesis pathway, ATP-citrate lyase (Acly or ACLY) and acetyl CoA carboxylase (Acaca or ACC). ACLY generates acetyl CoA, which is the substrate for ACC, the rate limiting and first committed step in de novo lipogenesis (Thampy and Wakil 1988). Pharmacologic and genetic inhibition of both ACLY (Wang et al. 2009) and ACC (Kim et al. 2017) decreases hepatic lipid accumulation. We did not observe robust changes in liver ACC or ACLY gene expression in response to either aging or calorie restriction, suggesting that transcriptional regulation of these genes is not robustly affecting hepatic lipid changes that result from calorie restriction or aging (Fig. 3C, D). Still, post-translational modification does robustly affect activity of acetyl CoA carboxylase. The activity of acetyl CoA carboxylase is inhibited by phosphorylation via AMP-activated protein kinase (AMPK) (Garcia et al. 2019; Lally et al. 2019) in response to a rise in the AMP:ATP ratio when cellular energy levels are low. Acetyl CoA carboxylase is regulated by the glucoregulatory hormones insulin and glucagon, encouraging lipid production when food is available and inhibiting lipid production when food is scarce. Hepatic insulin resistance in response to aging may decrease liver lipid content (Brown and Goldstein 2008).
Interestingly, we found that glucose tolerance improved with advanced age (26 and 28 months old; Fig. 4I). Yet, there was no further improvement in glucose clearance with CR in mice of advanced age. This observation is in stark contrast with what has been observed in aging humans (Shimokata et al. 1991; Ehrhardt et al. 2019). Typically, as humans age from middle to advanced age, glucose tolerance decreases. In fact, analyses from over 700 participants in the Baltimore Longitudinal Study of Aging, show that glucose tolerance declines from 60 to 92 years of age, independent of changes in body composition and activity levels (Shimokata et al. 1991). We found that OGSIS increases in advanced age in mice (Fig. 4J), possibly explaining the improved glucose clearance. Thus, aged mice secreted higher levels of insulin in order to clear blood glucose. In line with our findings, Oh and colleagues (2016), studying mice from 4 to 20 months of age, found that aging did not affect blood glucose concentrations, but improved glucose tolerance while decreasing insulin sensitivity (Oh et al. 2016). They similarly showed that aging (20 months) increased glucose-stimulated serum insulin (Oh et al. 2016). Using HOMA-IR to assess insulin resistance, we found that insulin sensitivity was similar at all ages 10–28 months of age (Fig. 2C). Importantly, 15% calorie restriction decreased HOMA-IR across mice of all ages (Fig. 2C).
Dysregulated glucose and lipid homeostasis increases the risk of developing limitations in physical function in older persons (Penninx et al. 2009). Aging causes a decline in physical function that can be delayed by caloric restriction. Similar to studies that implement a more severe level of restriction (Orenduff et al. 2022), the 15% calorie restriction we implemented improved forelimb and all limb grip strength in most age groups and improved balance and coordination in 26- and 28-month-old mice, as measured by time to fall during the Rotarod task. Grip strength measurements were normalized to body weight, and calorie restricted mice had a higher percentage of lean mass per gram body weight, thus it is reasonable that calorie restricted mice had a greater normalized grip strength than ad libitum mice. However, the increased time to fall during the Rotarod task is not corrected by body mass.
We must be judicious in raising the limitations of translating data from rodent models of caloric restriction to human aging and metabolic health. Observational studies of humans that involuntarily restrict caloric intake propose that there may be maintainable beneficial effects of modest calorie restriction. Based on six decades of archived dietary intake data, Willcox and colleagues (2007) estimated that residents of Okinawa self-impose approximately an 11% caloric restriction. This correlated with a life-long low BMI, decreased mortality from age-associated diseases, and extended mean and maximum lifespan (Willcox et al. 2007). Although promising, these correlative findings do not demonstrate direct causation between modest caloric restriction and increased lifespan in humans. Long term clinical trials are first required to assess the efficacy of moderate CR in preventing age-related disease and improving healthspan. While few controlled human trials have examined the physiological effects of long-term CR, data generated from the CALERIE trial (Comprehensive Assessment of Long-Term Effects of Reducing Intake of Energy) supports the hypothesis that there are substantial beneficial effects of modest CR. A 12% calorie restriction decreased body weight, fat mass (Das et al. 2017), and reduced multiple cardiometabolic risk factors, including LDL cholesterol, total: HDL cholesterol, and insulin sensitivity, independent of weight loss (Kraus et al. 2019). These encouraging findings from the CALERIE™ trial and our data from similarly calorie restricted (15%) mice support the need for future research aimed at understanding the metabolic impact of moderate caloric restriction in both human populations and animal models of aging.
There are some limitations of our study that must be considered when interpreting results. One potential limitation of our study is the variable fasting durations between our ad libitum and calorie restricted mice for in vivo metabolic studies and tissue collections. Ad libitum fed mice would have been imposed with a 4 h fast (likely ate very little for 8 h, since lights on), while calorie restricted mice likely fasted for 14–16 h. This limitation is hard to overcome, as feeding the daily food allotment hours prior to sacrifice would potentially have a greater impact on our measures of metabolic health. Ideally, we would have performed insulin tolerance tests on all ages of mice. However, HOMA-IR does provide a measure of insulin resistance, establishing that calorie restriction improves insulin sensitivity across all ages of mice. We recognize that the number of mice in the 28-month-old age group is low, relative to younger age groups in this study. The median lifespan of ad libitum fed C57Bl/6 mice is approximately 28 months (Turturro et al. 2002). Hence this discrepancy in the number of mice per age group was unavoidable due to mortality at this advanced age. Another limitation is that our studies are limited to male mice. Hormonal changes that occur in midlife in women are associated with dysregulation of lipid (Fan and Dwyer 2007; Derby et al. 2009; Woodard et al. 2011) and glucose (Lindheim et al. 1994; Ryan et al. 2002; Rossi et al. 2004) homeostasis. Similarly, mouse models of menopause exhibit weight gain, elevated fasting insulin, and insulin resistance (Romero-Aleshire et al. 2009). Given these metabolic consequences that occur during midlife that are unique to women, studies examining the effects of moderate caloric restriction in female mice across the lifespan is essential to understand potential sex differences in the metabolic response to caloric restriction.
Our findings indicate that a moderate, maintainable level of calorie restriction beginning at early adulthood can limit the decline in metabolic and physical (strength, balance, and coordination) function with aging in mice. In conclusion, 15% calorie restriction may cause comparable metabolic and physical benefits to the typical higher percentage CR, with the added benefit of increased likelihood of compliance in human populations. These findings support the need for future research aimed at understanding the physiological impact of modest caloric restriction in animal models of aging.