Intermittent fasting: from calories to time restriction

Abstract

The global human population has recently experienced an increase in life expectancy with a mounting concern about the steady rise in the incidence of age-associated chronic diseases and socio-economic burden. Calorie restriction (CR), the reduction of energy intake without malnutrition, is a dietary manipulation that can increase health and longevity in most model organisms. However, the practice of CR in day-to-day life is a challenging long-term goal for human intervention. Recently, daily fasting length and periodicity have emerged as potential drivers behind CR’s beneficial health effects. Numerous strategies and eating patterns have been successfully developed to recapitulate many of CR’s benefits without its austerity. These novel feeding protocols range from shortened meal timing designed to interact with our circadian system (e.g., daily time-restricted feeding) to more extended fasting regimens known as intermittent fasting. Here, we provide a glimpse of the current status of knowledge on different strategies to reap the benefits of CR on metabolic health in murine models and in humans, without the rigor of continuous reduction in caloric intake as presented at the USU State of the Science Symposium.

Introduction

Long-term calorie restriction (CR) without malnutrition is one of the most consistent dietary interventions to improve health and extend survival across multiple species [1, 2]. Previous studies have indicated that prolonged periods of CR delay phenotypic aging and associated morbidity risks, which led to the notion that chronological age can be disconnected from biological age, the latter being the major risk factor for age-related diseases, frailty, and loss of resilience. This has led to the recognition of the “Geroscience” paradigm [3], which states that, since aging physiology exerts a critical role in chronic diseases, any strategy that delays the aging process itself would decrease age-related increase in morbidity and comorbidity risks. In the last few decades, the focus of aging research has shifted from studying individual age-related diseases in isolation toward a wider context to define mechanistically the basic biology of aging.

Although achievable in model organisms, prolonged CR implementation in humans is impractical due to its poor adherence. Efforts are now underway to tease apart the factors driving the health and survival benefits of CR and to develop effective dietary alternatives that elicit the same benefits but without its rigor. The benefits garnered from CR were initially attributed to the chronic reduction in caloric intake; however, the explanation for CR’s beneficial gains appears to be more nuanced. Manipulation of fasting time [4, 5], nutrient composition [6], circadian rhythms [5], age of interventional onset [7], and sex [2] are all variables that can impact the effectiveness and outcomes of CR. Moreover, these variables could be capitalized upon to elicit CR’s health benefits with less stringent dietary interventions.

The metabolic switch between fasting and fed states is among the variables that drive CR’s improvement in health and survival. Identification of transcriptional and metabolite remodeling by CR is reminiscent of patterns associated with prolonged fasting [8]. However, differences in feeding regimens make it difficult to ascertain whether meal frequency or the level of calorie restriction is the predominant factor [1, 7]. This is because studies examining the impact of CR in rodents have inadvertently introduced a fasting time. Animals are routinely fed once daily, and the greater the level of CR, the faster the food is consumed which is accompanied by a longer daily fasting time (16–20 h) [5, 6, 9]. Moreover, cutting the feeding window down without reducing the amount of food available increases survival, although to a lower extent than CR [6]. These findings suggest the importance of the duration of fasting in driving CR-mediated longevity benefits in rodents.

The long-term advantages of meal timing and frequency in humans are less understood. Early epidemiological studies found a worsening of risk factors associated with cardiovascular disease [10] and type 2 diabetes (T2D) [11] in response to reduced meal frequency (1–2 meals daily, primarily skipping breakfast) vs. consumption of three or more meals per day [11] or having the daily intake of calories spread over multiple meals and snacks [10]. However, no correlation between meal frequency and T2D risk was found in a 6-year follow-up study in women [12]. Prospective studies have shown that increase in snacking augments the risk of weight gain [13] and T2D [12] and that an increase in the daily frequency of meals and snacks (>3 per day) evokes a greater body mass index (BMI) gain [14]. Extended overnight fast (18–24 h) was more effective at lowering BMI than a shorter fasting period (7–11 h) [14]. Temporal meal “skipping” (early vs. late feeding) may be consequential as men who skipped breakfast had a higher risk (27%) of cardiovascular events compared with those who did not [15]; whereas overweight adults who consumed the majority of their calories in the morning showed more significant weight loss and decreases in waist circumference than those who consumed an equivalent amount of calories in the evening [16]. Continual snacking throughout the day increases caloric intake and the propensity of adults for late-night eating [17]. Meal irregularity also negatively influences metabolic parameters [18]. Taken together, maximal metabolic improvement requires a consideration of the interplay of fasting with meal duration and frequency.

Diet composition is another metabolic lynchpin, as differences in nutrient composition (ratio of protein, fat, and carbohydrates) [19, 20], food palatability [7], and physical form of food (e.g., solid vs. liquid) [21] can all contribute to the CR response or induce a CR-like state. Analysis from the National Health and Nutrition Examination Survey showed that protein restriction led to reduced mortality in those under 65 [22]. High carbohydrate, low protein diets have been implicated in increased lifespan and cardiometabolic health [20, 23]. Additionally, restriction of specific amino acids (tryptophan and methionine) can improve rodent survival and metabolic parameters and reduce tumor burden [24]. CR-mediated lifespan extension was observed regardless of whether mice were eating a semi-purified diet or a diet composed mostly of natural ingredients [6], suggesting that the benefits of CR could be independent of diet composition.

Because of the low adherence to prolonged CR for most humans [25] and the multiple variables (e.g., fasting, types of nutrients, sex, and circadian rhythm) that can impact its effect, alternative feeding methods have been developed to provide the same benefits as CR. Intermittent fasting (IF) (Fig. 1), also known as periodic energy restriction [26], encompasses eating patterns in which periods of fasting (or negligible energy intake without deprivation of essential nutrients) are followed by periods of ad libitum feeding. Common forms of IF include [1] periodic fasting (PF) and 5:2 intermittent fasting (fasting 2 days each week); [2] time-restricted feeding (TRF), where the daily feeding window is reduced; and [3] alternate-day fasting (ADF), where animals cycle between fasting on 1 day and then eating the next day [27]. Of significance, recurring IF eating patterns confer health benefits by delaying many age-related diseases while improving metabolic markers independently from weight loss. In addition to a reduction in blood glucose levels, there are improvements in insulin sensitivity, depletion of glycogen stores, elevation in circulating ketones, and a decrease in systemic inflammatory markers [28].

Fig. 1.

Various forms of dietary interventions. Daily caloric restriction (CR) and four forms of intermittent fasting (IF) are depicted: alternate-day fasting (ADF) with 3 days of feeding interspaced with 4 days of fasting, 5:2 cycle consists of 5 days of feeding followed by 2 consecutive days of fasting, time-restricted feeding where the feeding window is limited to 4 h (TRF-4) or 8 h (TRF-8) over a 24-h period, and the fasting mimicking diet (FMD) comprised of 4 days of low caloric intake followed by 10 days of ad libitum feeding. Respiratory exchange ratio (RER) in mice with unlimited food access (AL, dashed pink line) or CR (solid orange line). The RER trajectories of mice subjected to ADF, TRF-4, TRF-8, or FMD (solid green line for the low-calorie portion of the cycle “ON” and AL portion “OFF” depicted by dashed pink line) are shown. The weekly percentage of fed vs. fasted periods on the different dietary regimens, with the FMD portion spanning 2 weeks (4 days “ON” and 10 days “OFF”). Mice fed AL eat around the clock and, therefore, do not experience significant periods of fasting. Daily CR can improve markers of health span and lifespan. Similarly, different forms of IF confer health benefits and possibly improvement in long-term survival in mice, non-human primates, and humans

Metabolite remodeling in the fasted vs. fed states

Glucose, ketones, and fatty acids are metabolites driving much of the physiological responses in the fasted vs. fed states. Glucose serves as the primary energy source for most tissues, and the rise in postprandial blood glucose triggers the release of insulin from the pancreas to facilitate the uptake, utilization and/or storage of glucose in cells throughout the body. During a short period of fasting, insulin levels drop, causing the release of glucose by the liver through the breakdown of stored glycogen. Simultaneously, there is a breakdown of triglycerides from white adipose tissue and the subsequent release of non-esterified fatty acids into circulation. These fatty acid molecules undergo beta-oxidation in liver mitochondria, resulting in the production of acetyl-CoA, a key modulator of de novo glucose synthesis via hepatic gluconeogenesis. As the fasting period continues, free fatty acids are metabolized by the liver into ketone bodies as an alternative fuel source for extrahepatic tissues such as the brain and heart. There are clear differences across species, biological age, and diet composition in the adaptive physiological response to fasting. In humans, during the initial fasting period, circulating ketone levels rise (0.2–0.5 mM) and, if continued for 48 h, can reach up to 1–2 mM [29]. In mice, elevation in circulating levels of ketone bodies occurs promptly, within 4–8 h of fasting, and continues to rise into millimolar concentrations after 24 h [30].

Beyond their role as an energy source, ketone bodies help organisms cope with the fasting state and are vital in promoting cellular activation of the stress response under prolonged IF. Elevation in ketone bodies coincides with the activation of peroxisome proliferator–activated receptor γ coactivator 1α (PGC-1α) and an increase in the levels of nicotinamide adenine dinucleotide (NAD⁺), a key cofactor for a variety of longevity-promoting enzymes including sirtuins and poly(adenosine diphosphate [ADP]–ribose) polymerase 1 (PARP1) [31]. Sustained ketosis stemming from prolonged fasting results in metabolic alterations and stimulation of a coordinated cellular stress response that involves the removal of oxidatively damaged proteins and organelles, and short-term blockade of protein synthesis via downregulation of the mTOR pathway [26]. This tissue-wide activation of repair pathways enables an organism to cope better against potentially damaging insults and to delay phenotypic aging.

The ability to switch between glycolysis and ketosis (“keto-adaption”) is necessary for survival as fat-derived ketone bodies become the predominant energy source following prolonged fasting [32]. Beyond periods of IF or TRF, nutritional ketosis (low carbohydrate/high-fat diet) has been proposed as a dietary intervention to induce keto-adaptation, which has been linked to the delay of many age-related diseases, including T2D [33], cognitive impairments [34], and physical decline [35]. Unfortunately, the ability to rapidly initiate keto-adaptation declines with increasing age [34]. Studies in aged rats show a marked delay in ketosis initiation [34], which may stem from impaired metabolic switching between fasted vs. fed states or impaired glycogen dynamics. Regardless, older animals are deficient in either converting glycogen to glucose through glycogenolysis or are unable to utilize efficiently carbohydrate or fat as fuel sources. However, aged rats on TRF were refractory to this metabolic dysfunction, showing instead improved glucose and insulin sensitivity, especially when maintained on a ketogenic diet [36]. Aging is often associated with insulin resistance and elevated blood glucose, two conditions that may lead to ineffective switching between fed vs. fasted states in older individuals. By preserving or restoring this metabolic balance, CR and IF interventions are emerging as crucial tools to delay many age-related diseases that stem from metabolic dysfunction.

Impact of IF on metabolic remodeling and survival

New paradigms have been developed to recapitulate the beneficial effects of CR on metabolic remodeling with varying degrees of success. One that has gained increasing attention is IF, a series of dietary interventions characterized by short-term periods of little to no caloric intake (16–24 h in mice), followed by periods of AL feeding. IF was first reported to extend longevity in Wistar rats by 15–20% and delay tumor growth by 65–90%, with other studies showing improvement in survival in female mice [37]. Much effort has centered subsequently on understanding whether IF is a feasible and translatable alternative to CR and the extent to which it elicits the same physiological improvements.

One of the most common forms of IF is ADF, wherein organisms consume food AL over 24 h, followed by 24 h of fasting [26]. Some rodent strains subjected to ADF show a significant decrease in body weight (5–30%) [28], primarily due to reduction in visceral fat, maintenance of lean mass [38], and “browning” of white adipose tissue [39, 40]. Like CR, ADF can tilt the metabolic switch toward fatty acid oxidation and lower glucose and insulin levels [28, 36, 40, 41]. Moreover, ADF can reverse insulin resistance and improve glucose homeostasis in rodent models for T2D [28, 39, 42]. Since cancer relies heavily on glucose utilization for energy production, ADF was found to delay colorectal [43] and breast [44] tumor progression [for additional reading, please see [45–47]]. In a similar vein, ADF confers protection against myocardial infarction in mice in an autophagy-dependent manner [48] and appears to improve neurological function by reducing amyloid-beta deposition in APP/PS1 transgenic model of Alzheimer’s disease [49]. In contrast, ADF was found to exacerbate neuronal death of dopaminergic neurons in a murine model of Parkinson’s disease [50]. Late-life improvements in frailty markers have been reported in mice that were maintained on a standard laboratory diet until 20 months of age before being switched to ADF for 2.5 months compared with controls [51].

PF is another alternative feeding regimen to CR that consists of longer but less frequent fasting periods. One form of PF is the fasting mimicking diet (FMD), which combines a plant-based, very-low-caloric diet provided continuously for 4 days, twice a month. Like CR, FMD is an effective way to elicit beneficial physiological responses such as lower glucose and IGF-1 levels and increase circulating ketones, without the rigor of chronic CR. Repeated FMD cycles reduce blood and insulin levels during the fasted state and decrease visceral fat accumulation, while improving cognitive functions in C57BL/6J retired female breeders [52]. A pilot study in healthy young and middle-aged individuals also found that three cycles of FMD improved fasting blood glucose, increased serum ketone bodies, and reduced IGF-1 levels [52]. The FMD reduces the impact of many age-associated diseases and delays cancer progression in rodents undergoing chemotherapy [53] while synergistically improving outcomes to current standard of care in human breast cancer clinical trials [54]. FMD was effective in restoring beta-cell function in a mouse model of T2D [55], promoting intestinal regeneration [56], and attenuating the loss of dopaminergic neurons in a mouse model of Parkinson’s disease [57].

Though there is evidence to suggest that IF improves many metabolic parameters and delays age-associated diseases in rodents, data that supports similar improvements in humans are still relatively sparse but growing. Moreover, most IF interventions and current literature are focused on its application for weight loss and restoration of metabolic function in overweight and obese subjects [58–60] rather than in healthy individuals. IF studies in humans show large study-to-study protocol variation [59], making it nearly impossible for direct comparison between CR and IF. However, even with these limitations, initial results suggest that IF could be a practical and effective approach to improving health and combatting the burden of several common diseases in humans. The first short-term IF study conducted in obese adults showed beneficial gains in physiological parameters, including body weight loss, metabolic shift toward fat utilization, and reduction in markers of oxidative stress [61]. The two most common forms of IF are alternate-day energy restriction (24-h period of unlimited food access followed by 24 h of ~70% restriction) and 5:2 feeding regime (5 days of AL feeding followed by 2 consecutive days of low caloric intake). For more information on current IF clinical trials, please see review by de Cabo and Mattson [26]. Even with a broad range of implementation strategies, humans on IF lose fat and overall body weight [59, 60, 62–64] while also exhibiting lower circulating levels of triglycerides, glucose [65], and insulin [66], all of which contribute to the delay in the onset of age-associated diseases. Thus, it comes with little surprise that IF interventions are capable of lowering the risks associated with T2D [67], cardiovascular events [64, 68], and side effects of chemotherapy [69]. More recent work has found IF to be beneficial even in healthy adults as short-term (1 month), and long-term (6 months) interventions revealed significant improvements in physiological parameters, notably body weight loss due to lower lean and fat mass and increase in circulating ketones without decreasing resting metabolic rate or bone mineral density [70]. Interestingly, there was no significant change in insulin sensitivity, a finding attributed to the fact that healthy adults are already within the ideal range and insulin sensitivity [71]. This was a landmark study for its use of a healthy population and for finding that long-term ADF could further improve predictive markers of cardiometabolic health, including reducing cholesterol and inflammation.

The few human studies that have done direct comparisons between IF and CR have revealed that IF could be as good as and, in some cases, even better than CR. The majority of these studies had low numbers of participants and were conducted in overweight or obese adults with various forms of IF but in most instances have reported improvements similar to the CR cohort [28, 60]. Moreover, a 3-month study comparing daily CR (25% caloric deficit) vs. intermittent energy restriction (5:2 feeding regimen) in overweight women, ranging from 20–69 years old, found that, although both groups had a significant decrease in body weight, the 5:2 regimen led to greater reduction in body fat and improvement in insulin sensitivity [66]. In the context of T2D, IF and CR appear to be equally effective at lowering visceral fat, although CR may cause a more profound reduction in body weight. Both interventions were able to decrease body weight at all ages regardless of sex and in prediabetic individuals, and to improve insulin sensitivity. However, more studies conducting direct comparisons between IF and CR at different biological ages (young-, middle-, and old-age) and disease states need to be performed before IF becomes mainstream and accepted therapeutical approach.

Impact of circadian rhythms in TRF studies

TRF is a form of dietary restriction accomplished by limiting eating to less than 10 h per day, without an overall reduction in daily caloric intake [27]. Many rodent and human studies have shown that increasing the time spent in the fasted state improves metabolic homeostasis and protects against chronic metabolic disorders as well as cardiovascular, neoplastic, and neurodegenerative diseases [26, 72]. Several recent studies have shown the profound impact of circadian rhythmicity on whole-body metabolism and metabolic health, and its interplay with the timing of energy intake and/or duration of fasting [26, 28]. Restricting temporal access to food to the active phase (nighttime for rodents and daytime for humans) can prevent body weight gain and susceptibility to metabolic disorders, while eating at the “wrong time of the day” could have the opposite effect [73, 74].

The biological and physiopathological mechanisms responsible for the beneficial effects of TRF on metabolic health require an assessment of its effectiveness in terms of short-term vs. long-term regimens, sex differences, and the influence of genetics. Obesogenic diets undermine the circadian rhythmicity of many metabolic pathways through their aberrant inhibition or activation. Numerous studies have tested whether and how TRF protocols can counteract the detrimental effects of high-fat diets on metabolism [72]. Young (3 months old) male mice on a high-fat diet and subjected to TRF for 6 h each day for 8 weeks had lower weight gain, were leaner, and showed improvement in biomarkers of metabolic health that included reduction in both hepatic triglycerides and circulating leptin and cholesterol levels [75]. Likewise, 8 h of TRF each day over a period of 12 weeks has been shown to decrease body fat accumulation and improve glucose tolerance, while restoring diurnal rhythms of core clock transcribed genes in high-fat diet-fed young male mice [76]. Weight loss and better glycemic control were achieved in adult (12 months old) male mice maintained on a high-fat diet under 8 h of TRF each day for 3 months [73] and 5 months [77]. Even in young mice with a genetically ablated circadian clock, 10 h of TRF protected against high-fat diet-induced obesity, fatty liver, dyslipidemia, and glucose intolerance [78]. Lifelong meal-fed (MF given access to AL portion daily) on a normal chow diet starting at a young age resulted on some periods of fasting (up to 11 h daily) and produced a significant improvement on morbidity and mortality in C57BL/6 male mice, compared with AL–fed controls [6]. Studies on female mice are still sparse, but in a model of postmenopausal diet-induced obesity, 7 weeks of TRF at 8 h per day resulted in a rapid decrease in body weight and a metachronous reduction in insulin resistance and hepatic steatosis [79]. However, further work is necessary to assess differences in sex-specific responses.

In a recent series of clinical studies, mostly in male and overweight adults, short-term reduction of the daily eating window was an effective behavioral intervention to contain the damage of metabolic diseases. Obese adults practicing time-restricted eating (TRE) for 8 h each day for 12 weeks lost weight had a lower systolic blood pressure through a mild CR without calorie counting [80, 81]. Similarly, adults with diagnosed metabolic syndrome practicing TRE for 10 h for 12 weeks had a significant decrease in systolic and diastolic blood pressure and reduction in total and non-HDL cholesterol, without a specific attempt to change physical activity or the diet qualitative or quantitative composition [82]. Overweight or obese older adults (≥ 65 years) at risk for or with mobility impairments who underwent a 4-week TRE for 8 h each day showed a significant decrease in body weight and BMI, with small but significant improvement in gait speed [83], a critical parameter in frailty risk [84]. The practice of TRE for 6 h initiated early in the day (with dinner before 3 p.m.) for a period of 5 weeks improved whole-body insulin sensitivity and beta-cell responsiveness in prediabetic male adults independent of daily energy intake [85]. A 4-day TRE intervention for 6 h early in the day was found to improve glycemic control, accompanied by lower 24-h glucose levels, reduced glycemic excursions, and improvement in insulin signaling [86]. Noteworthily, 8 h of TRE for 5 days in overweight male adults affected the rhythmicity of serum and skeletal muscle metabolites as well as the rhythmicity of genes controlling amino acid transport, without perturbing core clock gene expression [73, 87]. Initial work suggests that reducing the feeding window below 6 h may not confer additional benefits, as obese adults implementing 4 h (3–7 p.m.) or 6 h (1–7 p.m.) of TRE for 8 weeks showed comparable reductions in body weight (~3%), caloric intake (~550 kcal/day), and oxidative stress and improvements in insulin sensitivity [88]. The combination of TRE with strength training exercises led to weight loss without impacting muscle mass in healthy young women and men [89, 90]. A pilot study has recently assessed the safety, tolerability, and overall feasibility of a 6-week TRE intervention for 8 h each day in a healthy cohort of non-obese middle-aged and older adults of both sexes. Although no weight loss was recorded and cardiovascular health was not impacted, this short-term intervention had mildly improved functional (endurance) capacity and glucose tolerance [91].

More clinical studies with higher statistical power are needed to reveal any age, sex, or genetic differences on various clinical outcomes, especially in healthy individuals. The extent to which findings from rodent models are translatable to humans has still to be determined. Nevertheless, research on large cohorts of mice of various strains and of both sexes that are initiated at different ages is key for providing further insight into a proper design of randomized clinical trials. In that regard, the Study of Longitudinal Aging in Mice (SLAM) at the National Institute on Aging (NIA/NIH) [92] is well suited to make significant contributions to our understanding of normative mouse aging.

Conclusions

The average life expectancy continues to rise around the world and with it, a likelihood of increasing age-associated burden of chronic diseases (e.g., T2D, cardiovascular disease, cancer, and neurological disorders). Arguably, one of the most influential discoveries in aging research is the recognition that simply lowering energy intake can be highly salutatory on overall health and longevity. Although prolonged periods of CR confer health benefits and increased survival, its rigor makes this dietary approach rather unappealing for the vast majority of individuals [25]. A growing body of evidence suggests that the length of fasting, dietary composition, and circadian rhythm contribute to the pleiotropic effects of CR. Hence, alternative eating patterns such as IF have sought to capitalize on these different variables to develop more palpable approaches that capture health benefits, but without its low compliance problem.

Multiple iterations of IF have been developed ranging from ADF to TRF that is designed to interact with our circadian system. Significant metabolic improvements have been reported with IF interventions such as reduction in body weight and fat mass, lower blood glucose, and improvement in insulin sensitivity both in short-term and long-term studies in rodents and in humans. Individuals with compromised metabolic health may benefit from these dietary approaches by bringing the regulated variables back within physiological range, a crucial step in staving off multiple chronic diseases. IF has also been shown to further improve metabolic and physiological markers in healthy adults. Overall, more research is needed to ascertain how IF and its various iterations can be the most favorable based on sex, age, and genetic background.

Funding

This study was supported by the Intramural Research Program of the National Institute on Aging, National Institutes of Health. L.C.D.P.W. was supported by the NIH Grant #Fi2GM123963 from the National Institute of General Medical Sciences of the National Institutes of Health.