Time-restricted feeding and the realignment of biological rhythms: translational opportunities and challenges

Evidence establishing the strong links between biological rhythms and stress response is overwhelming and, by now, very well established and accepted [14, 46]. However, the translational implications, opportunities and challenges of how to manipulate rhythms in an ICU environment are only now beginning to emerge in the scientific discourse [30]. A number of recent, and older, reviews have discussed the connections between immune function and biological rhythms [46] where the bi-directional relationship between disrupted rhythms and immune dysfunction; and its implication on the bedside have been clearly identified [36, 37]. What is even more interesting is the fact that we begin to realize that circadian dysfunction following stress may have long lasting ramifications [47] pointing to possible sources of comorbidities. In fact, it has been argued that different procedures impact post-operative circadian disruption in a differential manner, thus affecting recovery, raising the possibility of guiding operative procedures based on their capacity to minimize impact on biological rhythms [48]. Clinical studies specifically emphasized that biological night and day cycles (measured by urinary 6-sulfatoxymelatonin) were phase-delayed and normal features of sleep were lacking (REM sleep was identified only in 2 patients out of 21) in the critical care patients [49]. Studies on patients undergoing elective maxillofacial surgery showed that strengthening circadian rhythms in anticipation of disruption following surgery can be efficacious for improving the recovery phase. Patients whose circadian rhythms were adjusted pre-operatively by combined sleep/wake cycle alteration and timed food and caffeine ingestion had reduced disruption in their body temperature cycles throughout their recovery in comparison to the control group [50].

One of the most active areas of research pointing directly to circadian disruption and biological rhythm-setting interventions relate to mood disorders [51‐55]. A vast literature exists on enhancing circadian rhythms for treating depression, bipolar disorder and other related mood disorders either via pharmacological (melatonin) [55, 56] or non-pharmacological means (light) [57] aimed at boosting circadian rhythms.

Time restricted feeding (TRF) is essentially imposing rhythms on nutrient availability. Entrainment by TRF has generated significant interest due to the possibility of synchronizing peripheral clocks without clear influences on (or from) the central pacemaker (SCN) [28, 58]. It has been speculated that restricted feeding (RF) entrains rhythms in peripheral tissues (liver and lung) [6] is likely independent of the SCN. These works challenge the basic hierarchical paradigm that light entrains the SCN which subsequently entrains the peripheral clocks and emphasized the role of RF as an entraining signal. The hypothesis of independently entrained peripheral clocks has been further reinforced by the observation that even lesions in brain nuclei do not eliminate food anticipatory activity, thus pointing to likelihood of a distributed system maintaining and regulating food-anticipatory activities [59, 60]. One of the main justifications is that when food accessibility adopts specific rhythmic characteristics so will the physiology and behaviour to match nutritional resource availability [61]. It has been shown that feeding mice during the day completely reverses the phase of circadian oscillators (specifically, four clock components, Per1, Per2, Per3, Cry1; and the two circadian transcription factors DBP and Rev-erbα) in multiple peripheral cells (liver, kidney, heart and pancreas), but has little if any effect on the central oscillator in the SCN [62]. However, we must point out that RF entrains the rhythm of clock protein Per2 even in the SCN as was shown in studies that eliminated photic stimulation by keeping mice in constant darkness [63], or at constant light conditions [64], thus raising the possibility of peripheral oscillators resetting the central clock.

In a carefully designed study of a murine obesity model [53] the authors convincingly show the intimate relationship between the signalling and transcriptional components of energy metabolism and the circadian system. The study hypothesized that TRF improves diurnal rhythms; drives lipid homeostasis while preventing weight gain, hepatosteatosis and liver damage; improves adipose homeostasis and reduces inflammation. The study demonstrated that preserving natural feeding rhythms significantly dampens metabolic disruption induced by a high fat diet, including improving oscillations of the liver circadian clock components. Therefore, while the total calorie intake and food composition (high fat) remained constant, the study clearly demonstrated that an apparent lifestyle, i.e.,non-pharmacological, intervention prevented obesity, and related co-morbidities, possibly by resetting metabolic cycles. The role of food-anticipatory activity has also been explored with a focus on energy metabolism, defined by oxygen consumption [65]. Animals were allowed access to food for only few hours during either the light or the dark phases. Locomotor activity, body temperature, clock gene expression in liver and energy metabolism were recorded and their changes assessed as the time window over which food became available was changing. Continuous monitoring of energy metabolism and core body temperature indicated expected, robust diurnal rhythmic characteristics but also rapid re-entrainment and adaptation to restricted food access.

A series of publications has focused on comparing protein synthesis under a continuous and, a likely more physiologically realistic, intermittent bolus feeding regimen, delivered by orogastric tube, in neonatal pigs in the context of regulating protein synthesis [66‐68]. The analysis demonstrated that intermittent feeding (delivered every 4 hrs as a bolus feed) enhances muscle protein synthesis by imposing pulsatile patterns of amino-acid and insulin-induced translation initiation. In this very interesting series of papers it has been argued that bolus feeding promotes a more physiological surge of intestinal hormones. The studies effectively hypothesize that “[…] cyclic surge of amino acids and insulin is needed to maximally stimulate protein synthesis in skeletal muscle” and that “[…] bolus compared to continuous feeding has been advocated to promote more normal feed-fast hormonal profiles”. It has been further demonstrated that either advancing or delaying meal time in young adult mice results in reversible alterations of temperature and overall cage activities [69]. Longer time restriction (one week) alters rhythms in glucose, triglyceride and HDL levels. Food restriction results in behavioral arousal in anticipation of food presentation and induces a shift in the circadian phase of many physiological variables, likely independent of the SCN. As such, RF is expected to exert changes in organs “handling nutrients” (such as liver). As previous work had suggested RF could be associated with significant stress due to hyperphagia, In a study examining the effect of restricted feeding on stress markers, no marked changes in body weight, retroperitoneal decrease in lipid deposits and peak in glucocorticoids accompanying expectation to food access were identified [70]. Given the probable relationship between stress and metabolic alterations (in this case interest was in liver) the study explored whether an increase in acute phase proteins (APR) or pro-inflammatory state occurred after 2 weeks of 2hr food restriction. The “positive control” for APR consisting of a group injected with LPS showed a significant increase in systems APR while neither the ad libitum nor restricted feeding induced a marked increase in any of the inflammatory markers. Furthermore, a marked change in the diurnal patterns of circulating cytokines was observed as a consequence of RF. The authors advance an interesting hypothesis stating that RF may establish a distinctive state (“rheostatic response” earlier introduced in [71]) likely enabling the system to adopt a transient functional state “change in set-point”, boosting the rhythms and the overall fitness of the host.

Peripheral circadian de-synchrony may be an early indicator of metabolic disruption in shift workers due to sleep deprivation mediated disruption of circadian rhythms. By extension, strengthening the peripheral circadian rhythm, by imposing metabolic rhythms via limiting food intake during the night, may counteract comorbidities seen in human shift workers [72]. This study further implies that the manipulation of circadian rhythms need not be such that it aims at restoring the homeostatic nature of the internal clock. Rather it implies that, at least in the short term, strengthening other rhythmic frequencies may be more beneficial.

Particularly interesting is the work investigating the effect of resetting circadian clocks in peripheral tissues using non-photic signals on tumor growth rate in rats [73, 74]. Restricting the timing of meals to light time in contrast to restricted feeding during the night (active phase of rats) thereby, imposing a reversed metabolic rhythm, induced, what is referred to as, “internal desynchronization” (described as loss of phase relationship between central – light entrained – and peripheral clocks) resulted in prolonged survival and slowed down tumor growth. The authors speculate that meal timing during the light period amplifies host rhythms and assigns their peak in a time window when the tumor is most susceptible to host-mediated control and that tumor growth is hampered when the internal (metabolic) clock adopts specific rhythmic characteristics, interestingly the opposite of what would have been otherwise considered “natural”. Therefore, the emerging hypothesis is that, a radically different metabolic rhythmicity appears to be most effective at least in the short term.

It is important to draw a distinction between time-restricted feeding and caloric restriction. The former entails the delivery of a certain amount of calories albeit at specific time intervals of specific duration. Therefore subjects still receive a standard nutritional intake. Calorie restriction entails an overall reduction in caloric intake, albeit without malnutrition. While evidence for the benefits of calorie restriction in animals has been promising, the issue as it relates to humans is still debated as conducting long term studies assessing the implications of prolonged calorie restriction in a controlled manner is rather complicated [75]. Although studies have shown calorie restriction improves post-trauma outcomes [76‐78], it is likely that the long term effects of calorie restriction are related to alterations of biological mechanisms responsible for maintenance of health [79]. Recent work has indicated the possibility of caloric restriction impacting circadian clocks as well [80]. However, it is argued that this may be a secondary effect of calorie restriction resulting in time restricted feeding imposing specific rhythms on metabolic function and entraining peripheral clocks. Nevertheless, the focus of this discussion is on TRF and not on calorie restriction.

A number of fairly comprehensive clinical studies have considered the impact of temporal delivery of enteral feeding in critical patients [81‐88]. Although these studies have to be acknowledged in the context of our discussion, one should be aware of the fact that clinical studies comparing continuous vs. bolus feeding were motivated mostly by the need to address some of the key practical limitations associated with delivering nutritional support, such as interruptions of continuous feeding leading to an inability to achieve nutritional goals, gastrointestinal complications, modulation of aspiration pneumonia, stool frequency etc., rather than as an attempt to capitalize on potentially advantageous physiological and/or biochemical routes linking metabolic rhythms and immune response. Earlier studies examined various parameters influenced by delivering enteral nutrition in the form of either continuous or bolus (intermittent) delivery and the conclusions are still debated in the clinical community [84]. Studies comparing continuous to intermittent tube feeding in adult burn patients concluded that patients continuously fed had reduced stool frequency and time required to achieve nutritional goals. More recent studies, however, despite minor differences in specific goals and targets, in general do not provide evidence of significant difference in terms of patient outcome. Results show that patients intermittently fed have a higher total intake volume, are extubated earlier, and have a lower risk of aspiration pneumonia. Postoperatively, feeding at night only is more energy efficient than is feeding continuously for 24 h, but is associated with poorer nitrogen balance [82]. In one of the very few studies which complemented intermittent feeding in a clinical setting with monitoring of biomarkers, the observed decrease in urinary catecholamine secretion indicated a possible role of sympatho-adrenal mechanisms. This study provides a link between feeding patterns and putatively modulated pathways. However, no studies have been performed where time restricted feeding has been compared to either bolus or continuous feeding in the ICU.