Gut microbiota as a transducer of dietary cues to regulate host circadian rhythms and metabolism

Abstract

Certain members of the gut microbiota exhibit diurnal variations in relative abundance and function to serve as non-canonical drivers of host circadian rhythms and metabolism. Also known as microbial oscillators, these microorganisms entrain upon non-photic cues, primarily dietary, to modulate host metabolism by providing input to both circadian clock-dependent and clock-independent host networks. Microbial oscillators are generally promoted by plant-based, low-fat (lean) diets, and most are abolished by low-fiber, high sugar, high-fat (Western) diets. The changes in microbial oscillators under different diets then affect host metabolism by altering central and peripheral host circadian clock functions and/or by directly affecting other metabolic targets. Here, we review the unique role of the gut microbiota as a non-photic regulator of host circadian rhythms and metabolism. We describe genetic, environmental, dietary, and other host factors such as sex and gut immunity that determine the composition and behavior of microbial oscillators. The mechanisms by which these oscillators regulate host circadian gene expressions and metabolic states are further discussed. Because of the gut microbiota’s unique role as a non-photic driver of host metabolism and circadian rhythms, the development and clinical application of novel gut microbiota-related diagnostics and therapeutics hold great promise for achieving and maintaining metabolic health.

Blurb

This Review explores how the gut microbiota acts as a driver and regulator of host circadian rhythms and metabolism, highlighting its unique role in transducing dietary cues. Key determinants of microbial oscillations and insights into microbial control of chronometabolism are discussed.

Introduction

The Earth’s rotation on its axis dynamically exposes most organisms on the planet to sunlight. It creates oscillating light and dark cues that organisms need to adapt to in order to sustain their lives. The diurnal rhythms further affect environmental conditions such as temperature, water and nutrient availability to which most life forms adapt for survival and optimization of energy balance. Traditionally, photic cues have been thought to be the major, if not sole, drivers of organismal circadian rhythms, which are determined by a 24 h internal clock that controls both metabolic functions and behavior, such as sleep and wake cycles. For example, plants have circadian rhythms that are necessary for determining when and to what extent photosynthesis takes place, which is essential both for the plant’s own survival, and for the survival of obligate herbivores^1,2. These effects might then propagate through the food chain, via intricately intertwined predator–prey relationships as well as ecological cross talks based on mutualistic, commensal, or competitive interactions across various life-forms to affect their diurnal activities reciprocally³.

Interestingly, the reciprocal regulation of diurnal variations between organisms of different kingdoms, has been also observed in the interactions of the gut microbiome with its mammalian host. Within the past decade, it has become apparent that a new player, the gut microbiota, contributes to the regulation of mammalian physiology. The gut microbiota is a complex admixture of bacteria, viruses, protists, Archaea and fungi that collectively comprise a symbiont-like organ for the host⁴. While the largest number of microorganisms reside in the large intestine, they also exist in other parts of the gut with their composition varying down the cephalocaudal axis and in their luminal versus mucosal location ^5,6. This composition creates a complex of ecological niches for the gut microbiota to differentially benefit from and affect the host’s physiology. Unlike other organ systems, the gut microbial organ is acquired, selected and maintained by the host throughout life. In turn and like other organ systems, the gut microbiota supports physiological, behavioral and genetic processes that are important for development and general health of its host^7-9. The influence of the gut microbiota on host physiology, however, is not static. In addition to the composition and function of the microorganisms, their diurnal variations have a major influence on host circadian network and metabolism.

In this Review, we specifically focus on the influence of the gut microbiota–host–clock interactions on metabolism proposing a unique role of the gut microbiota in transducing dietary cues to regulate diurnal oscillations of host metabolic functions even in the absence of photic signals. We describe factors that drive microbial oscillations that feed into and affect host circadian networks and metabolism.

Host circadian clock and the gut microbiota

In mammals, the internal circadian rhythms are important for coordinating the functions of different bodily organs to align with the external environmental signals and achieve maximum efficiency¹⁰. The circadian clocks in mammals, while cell-autonomous, through genetic networks that involve transcriptional autoregulatory feedback loops, depend on external cues to regulate physiological, behavioral and metabolic networks to ensure fitness and survival¹¹. Asynchrony between the external and internal clocks can be associated with diseases such as obesity, diabetes, ulcerative colitis, cancer and Alzheimer disease^12-17. The circadian rhythms in mammals are mediated by a set of molecular components that are ubiquitously distributed inside all cells and are governed by clock genes. This intracellular machinery integrates the dynamics of external cues to create diurnal rhythms in the physiological processes of an organism. Of the many clock-controlled physiological and behavioral activities, metabolism is a primary target. The circadian genes modulate the metabolic signaling within a body, while nutrient factors can alter the expression of circadian genes¹⁸. Traditionally, the major light-mediated signals are believed to arise from the suprachiasmatic nuclei (SCN) located in the hypothalamic region of the brain^14,19. However, the nutritional state and eating behavior are suggested to be additional primary ‘zeitgebers’, which are the environmental cues that can reprogram or modulate peripheral clocks. Thus, although some metabolic processes are controlled through SCN-dependent pathways that only receive light cues, a non-photic cue such as a food-entrainable oscillator (FEO) might be another mechanism of regulating peripheral oscillations, independent of SCN²⁰. For instance, time-restricted feeding (TRF), which involves restriction of food intake to certain time intervals, for example 8-10 h of eating followed by fasting, controls nutritional cues that entrain the peripheral oscillations²¹. TRF helps maintain robust circadian rhythms and has been repeatedly shown to prevent and even reverse metabolic imbalances both in mice and humans²¹. Such dietary regimes improved health conditions of people with obesity or metabolic syndrome regardless of sexes and medication consumption^22,23. Equally important, circadian eating behavior and metabolic processes have been preserved in SCN-ablated animals maintained under a TRF cycle, which implies that TRF might alter the peripheral gene oscillations without affecting the SCN²⁴.

In the gut microbiota, the majority of studies on diurnal variations have focused on the bacterial members as far less is known about fungi and Archaea in general. In contrast to eukaryotic organisms, bacteria are generally believed to lack autonomous circadian clock gene networks or canonical clock genes. However, genetic and proteomic mechanisms of an autonomous circadian clock have been reported for the cyanobacterium, Synechococcus elongatus, which utilizes light as a primary energy source, but which does not reside in the metazoan gut^25,26. Even so, the cyanobacterial clock entrainment of light and dark signals is offset by metabolic signals²⁷. Non-photosynthetic bacteria including those in the human gut are also reported to exhibit circadian rhythms that are dependent on non-photic environmental signals⁴. For example, the circadian rhythms in the swarming motility of an intestinal commensal, Klebsiella aerogenes, was shown in vitro to be mediated by temperature cycles²⁸ and melatonin levels²⁹. Despite this evidence, knowledge on whether gut microorganisms have inherent clock mechanisms and to what extent their functions vary diurnally is still rudimentary. Instead, the studies published so far suggest that the biological rhythms of gut microorganisms might be dependent on specific environmental cues within the gut, which fluctuate according to dietary intake and host circadian rhythms and behaviors, rather than on cell-autonomous clock machinery. Reciprocally, their oscillations are integrated into the host circadian clock as another peripheral clock and can affect the circadian coordination of different host organs^30-32. Clearly, a number of intriguing questions remain to be elucidated: do various gut microorganisms have inherent clock mechanisms and can these affect the host’s circadian clock? Alternatively, is the central host clock the mastermind behind the bacterial oscillators? For example, does feeding behavior, under the control of the photosensitive central host circadian clock, alter intake of nutrients that cause bacteria to release metabolites that then serve as cues to entrain peripheral clocks and therefore metabolism? How does the local gut milieu influence the effect of feeding behavior?

What causes the gut microorganisms’ diurnal variations in microbial composition and abundance and, therefore, their influence on host metabolism has been of increasing interest. What is becoming clear is that not all members of the gut microbiota exhibit diurnal oscillations in their abundance and we and others term the small percentage that do as ‘microbial oscillators’^30,32-34. These so-called oscillators are present in both the small and large intestines of humans and mice^30,32-34. Thus, in the dark lumen of the intestines, they might not directly entrain to photic cues, but primarily to dietary signals. They then transduce the signals to modulate host circadian networks and metabolism directly or indirectly, via presumed small molecule mediators including metabolites. By doing so, the gut microbiota provides information to the host about what, when, and how much of nutrients have been consumed, enabling the host to adjust overall energy balance accordingly—they could act as other zeitgebers to the host’s circadian rhythms.

The connections between the gut microbiota, diet and metabolism have been the focus of intensive research. There is compelling evidence that the gut microbiota plays an important part in host metabolic homeostasis^9,35. An obvious corollary is that circadian modulation by the gut microbiota also influences the host’s nutritional state and feeding behavior.

Circadian clocks in mammals

In mammals, geophysical time is integrated into the body through the internal molecular clock composed of the central pacemakers, the SCN, and the auxiliary oscillators, both of which exist in the brain and the latter also exist in almost all peripheral tissues¹⁰. The canonical regulation of diurnal oscillations of physiological processes is mediated by the SCN, which integrate dynamic photic cues recognized by the retina and regulates other central and peripheral circadian networks¹⁰.

The typical mammalian circadian clock consists of two sets of genes: the primary clock genes, CLOCK (Circadian locomotor output cycles kaput), ARNTL (encoding aryl hydrocarbon receptor nuclear translocator-like protein 1, also known as BMAL1), PER (Period), CRY (Cryptochrome), which comprise the first feedback loop; and the secondary circadian effectors that include NR1D1 (encoding nuclear receptor subfamily 1 group D member 1, also known as REV-ERBα) and retinoic acid receptor–related orphan receptors such as RORA, and the downstream clock-controlled genes such as DBP (D site-binding protein) and PPARGC1A (encoding peroxisome proliferator-activated receptor gamma coactivator 1-alpha) (Fig. 1). These molecular components feedback on one another to maintain circadian rhythms in mammals (Box 1). These core circadian mediators then control many downstream genes relevant to physiological functions, including energy balance^11,14,18. The centrality of circadian rhythms in regulating myriads of biological processes is underscored by the following: almost all cells possess the primary clock genes, and the products of these genes serve as master-regulators via the E-Box of thousands of genes, many of which are themselves regulators thereby affecting a cascade of biological processes.

Fig. 1 ∣. Crosstalk between host circadian clock and the gut microbiota.

Cell-intrinsic clocks are ubiquitously distributed in host cells, including intestinal epithelial cells and immune cells, and are composed of two loops of core clock regulators. In the first loop, the heterodimeric complex CLOCK–ARNTL (also known as BMAL1) acts on E-box elements to regulate the expression of other circadian regulators, PER, CRY, REV-ERB, and ROR and other clock-controlled genes (CCGs). In turn, PER forms a heterodimer with CRY to repress the expression of ARNTL. RORs and REV-ERBs comprise the second circadian loop by competitively binding to the RORE element and regulating BMAL1 expression. The E-box elements regulate the expression of many other regulatory genes (CCGs and others not shown), to influence multiple functions. Microbial components (such as lipopolysaccharide (LPS) and flagellin) and microbial metabolites (such as short-chain fatty acids and unconjugated bile acids) also regulate the molecular clocks and the CCGs, thereby affecting a variety of physiological processes in host cells. In turn, the circadian profile of host factors such as light–dark cycle, feeding time and route, diet composition and production of mucus and antimicrobial peptides (AMPs) regulates the diurnal oscillations of the gut microbiota (upper inset). While not all gut bacteria exhibit diurnal oscillations in relative abundance, some exhibit strong or weak oscillations, which are primarily altered by host diets. (DC: dendritic cells; B, T: B and T lymphocytes; M: macrophages). The molecular clock (lower inset).

Box 1 ∣. Interlocked circadian clock loops in mammals.

Mammalian circadian clock genes are ubiquitously distributed in almost all cell types. The clocks largely consist of two interlocked loops that feedback on each other. In the first loop, the CLOCK–BMAL dimer induces the expression of PER and CRY and, in turn, the PER–CRY dimer represses CLOCK and ARNTL (also known as BMAL) expression. CLOCK–ARNTL also activate the transcription of a number of other clock-controlled genes (CCGs). Importantly, the cyclic activity of the CLOCK–ARNTL and PER–CRY dimers regulate the expression of the nuclear receptors, REV-ERBα and RORα, which form the second clock loop. RORα and REV-ERBα competitively bind to RORE (ROR/REV-ERB-response element), present in the ARNTL (encoding BMAL1) promoter to respectively increase or inhibit ARNTL transcription (Fig. 1). Equally important, these nuclear receptors are integral drivers of regulation and rhythmicity of thousands of processes, including metabolism. For example, administration of a REV-ERBα agonist increased energy expenditure and decreased fat mass in mice^71,72. Other nuclear receptors, specifically PPARs (Peroxisome proliferator-activated receptors), are also linked to the circadian control of metabolism. The expression of PPARα and PPARγ are diurnally controlled and they in turn affect the expression of circadian clock genes⁷³. High-fat diet (HFD) feeding did not induce insulin resistance in Ppara-knockout mice⁷⁴. PPARγ is required for adipocyte generation⁷⁵, whereas Ppard–knockout reduced adipose mass in mice⁷⁶. Thus, PPARs are generally related to lipid metabolism, insulin sensitivity, and glucose homeostasis and their agonists are being widely used as therapeutics in various metabolic diseases such as type 2 diabetes mellitus ^77,78.

Determinants of microbial oscillations

Microbial oscillators

Although many gut microbiome studies report the analysis of samples obtained at single, defined time points, sufficient evidence from studies involving sampling at multiple time points reveals that the composition of gut microbiota is not fixed, but dynamic throughout the day^{30,33,34,36,37}. However, only a small percentage of these microorganisms are oscillators; that is, exhibit oscillations in abundance. Up to 20% of taxa in the cecal and fecal microbiota from mice, especially Clostridiales, Lactobacillales and Bacteroidales, exhibit diurnal variation in relative abundance^30,34. The composition of the human gut microbiota was also found to vary depending on the time of defecation^34,37. Although studies interconnecting the circadian clock, motility and the microbiota are as yet sparse, it has been suggested that changes in diurnal bowel function in chronic constipation is associated with altered colonic microflora in humans³⁸. Nutrient availability and consumption fluctuate with time and are clearly contributing factors. Furthermore, metagenomic analysis indicates that certain microbial genes diurnally fluctuate, which could potentially affect host targets and functional activities in both mice and humans^34,37. The time of day when expression of oscillating microbial genes peaked also varied, possibly as a result of phasic shifts of the host environment, such as body temperature, substrate availability or the activities of the immune system. The biogeography and metabolome of the gut microbiota also changes during the course of a day. Thaiss et al. showed that different microorganisms in mice rhythmically adhere to the host gut epithelium in varying abundances ³¹. In the same study, the rhythmicity of intestinal metabolites such as ornithine and proline detected in serum was lost in antibiotic-treated and germ-free (GF) mice³¹. The time-specific exposure of the host intestinal cells to microorganisms and their products could potentially be another mechanism by which gut microorganisms influence host health.

Genetics, sex and environmental manipulation

Factors contributing to the circadian fluctuations of the gut microbiome have been identified. In addition to sex, the genetic or environmental disruption of the host’s circadian clock alters microbial composition and functional capacity in the gut^34,36,39. The deletion of one of the core circadian genes, Bmal1, altered diurnal variations of the gut microbiome at the genus level in mice. For example, both the amplitude and phase of the diurnal oscillations in relative and inferred absolute abundance of members of the Lactobacillaceae family at certain times were altered by Bmal1 deletion³⁶.The deficiency of other core circadian genes, Per1 and Per2, was further shown to abrogate compositional, functional, biogeographical and metabolomic fluctuations of the gut microbiome in mice^31,34. The influence of Bmal1 deletion on the gut microbiome was also sex-dependent; thus, although the diurnal oscillations of the relative abundance of gut (fecal) microbiota was more prominent in female than in male mice, this rhythmicity was absent in Bmal1-deleted mice. Furthermore the relative abundance of various bacteria was altered by Bmal1 deletion and this bacterial composition showed sex-related differences³⁶. Diurnal expression of host hepatic genes, including those of metabolic pathways exhibit sexual dimorphism and the sex-related differences disappear in germ-free animals⁴⁰. In terms of environmental manipulations, a mouse model of jet lag, mimicked by an 8h shift of light and dark cues, exhibited reduced number of diurnally oscillating bacteria and greater bacterial composition shift as compared to the control mice, that is, an increase in beta diversity between the two groups of mice. ³⁴. Conversely, the induction of circadian disruption by a 12h shift of light signals, reversing the day and night, altered the diurnal expression of Per2 in mice, but was not sufficient to alter the gut microbiota composition in another study³⁹. Instead, this study showed that feeding mice a high-fat, high-sugar diet in addition to the light–dark cycle reversal led to substantial microbial composition alterations decreasing the similarity between the two diet groups ³⁹. Further studies focusing on other circadian clock genes and the molecular basis for the influence of host sex and environmental disruption of the circadian clock network are needed for more comprehensive conclusions on how host clock affects microbial oscillations in the gut.

Dietary patterns and diet composition

In addition to the canonical clock genes and photic cues, other alterations in gut ecology influence the oscillations of the gut microbiome. For example, the diurnal oscillations exhibited by the mucosa-associated microbial community are mediated by the host circadian rhythms of feeding times, and production of mucus and antimicrobial peptides (AMPs)³¹. Notably, dietary patterns are shown to be the main driver of the circadian oscillations of the gut microorganisms as well as the animals’ physiological processes. The influence of feeding regimens, the circadian clock and animal physiology have been well-studied^41,42. In terms of animal physiology, rhythmic food intake, guided by SCN-driven cues, accounts for 70% of rhythmic hepatic gene expression and is largely independent of the hepatic clock⁴³. When mice fed only in the light phase were compared to mice fed in the dark phase only, the oscillatory patterns, including amplitude and peak and/or trough fluctuations, of some microorganisms such as Limosilactobacillus reuteri (formerly Lactobacillus reuteri) were similar. However, one critical difference was that the dark-phase-feeding and light-phase-feeding exhibited a 12h-difference in the relative abundance of those microorganisms. The dark-phase-only feeding recapitulated the types and rhythmic patterns of oscillating microorganisms observed in the ad libitum fed mice, reflecting that mice are nocturnal eaters. Furthermore, in Per1/2-deficient mice, TRF, either only during the light or the dark phase, restored the microorganisms’ diurnal dynamics that were vastly reduced in mice fed ad libitum³⁴. TRF-mediated alteration in gut microbiota composition was replicated in healthy male humans whose gut microbiome composition and richness differed between the TRF group (n = 56) and non-TRF group (n = 24) ⁴⁴. However, TRF did not change the gut microbiota composition and richness in 14 people with obesity⁴⁵.

As mentioned previously, the host diet strongly influences the cyclic fluctuations of the gut microbiome. It can determine what microbial oscillators are present and whether they exhibit diurnal variation. For example, members of Clostridiales, are abundant and exhibit diurnal variation in mice fed a regular chow diet, but both features are lost with high-fat diet (HFD) feeding. On the other hand, a relative increase in abundance of Ruminococcus is seen in mice with HFD without exhibiting any oscillatory activity³⁰. An increase in oscillation only under HFD, but not regular chow-feeding, is infrequently seen in some operational taxonomic units (OTUs) such as Lactocococcus representing the oscillatory changes in relative abundance of bacteria resolved to a genus-level³⁰ although how, if at all, this change influences metabolism is unknown. In another study, mice fed normal chow ad libitum exhibited circadian oscillations of microorganisms in major phyla including Firmicutes, Bacteroidetes and Verrucomicrobia. The diurnal dynamics, however, were diminished in mice fed ad libitum with a HFD³². Moreover, the oscillation of the alpha diversity , which represents the bacterial membership within the local community, peaked during the daytime in mice fed normal chow ad libitum was not observed in mice on a HFD regardless of whether the HFD feeding was time-restricted or ad libitum. By contrast, the number of cycling microorganisms that undergo oscillations in abundance also decreased in both ad libitum and the dark-phase-only HFD-fed-mice³². This finding shows that the composition of the diet can complicate the effects of TRF on microbial oscillations. Nevertheless, the subphylum types of cycling microorganisms and their oscillatory patterns differed between mice fed with a HFD ad libitum versus those fed in the dark phase only. Further beta diversity analysis also revealed that the TRF of HFD led to much more temporal fluctuations in the microorganisms than the ad libitum feeding of HFD. Interestingly, the TRF even rescued some of the changes in cyclical behavior of obesity-related microorganisms observed in the ad libitum HFD-fed mice³². The rhythmicity of the gut microbiota is further influenced by the route of feeding. For example, parenteral nutrition abrogated the effect of oral HFD feeding on the diurnal oscillations of the gut microbiota in mice³⁰. This finding underscores that the composition, route, time of feeding and sex need to be further explored as important contributors in the interplay between the host’s circadian network and the diurnal oscillations of the gut microorganisms. Although complex, such studies are easier to perform in defined animal models than in humans. What is clear in human studies is that some gut microorganisms exhibit diurnal rhythmicity and that this rhythmicity is disrupted in individuals who are obese (BMI >30)⁴⁰. In addition to species and sex differences, the difficulties of extrapolation to human studies are compounded by considerations of underlying physiological status.

Mucus and antimicrobial peptide Reg3γ

The previous studies were based on the luminal gut microbiota from fecal or cecal samples. Other investigations on the intestinal mucus layer in mice demonstrated that microbial occupation of the mucus also exhibits diurnal dynamics³¹. In addition, the mice exhibited cycling alterations in the thickness of the mucus layer³¹. The changes in the bacteria residing in the mucosa of micewere expected because the mucus and the AMP, Reg3γ, regulate the spatial segregation that reduces the exposure of host epithelium to microorganisms⁴⁶. Predictably, the diurnal rhythmicity in the mucosa-associated bacteria was not found in Reg3γ-deficient mice³¹. Although the latter study hinted at the effects of mucus and AMP on the circadian dynamics of the mucosa-associated gut microbiota, it did not explore whether these host factors also affect the diurnal oscillations of the luminal microbiota and how that compares to the changes in the mucosa-associated microbiota.

Although diurnal oscillations of gut microbiota are influenced by various factors, it has also been observed that not all bacteria fluctuate diurnally^30,33,37. Importantly, what determines oscillating and non-oscillating microbial composition in the gut and how they differentially affect host health are just beginning to be elucidated. A study published as a Preprint in 2020 investigated the role of Reg3γ in regulating the composition of microbial oscillators in the small intestine of mice³³. The study demonstrated that the membership of oscillating OTUs was greatly changed between Reg3g^+/− and Reg3g^−/− mice. In addition to an increase in the number of oscillating OTUs, the dominant oscillating microbial family was changed from Bacteriodales in Reg3g^+/− mice to Clostridiales in Reg3g^−/−mice³³. These changes were dampened by HFD and the total number of oscillating OTUs was decreased on a HFD compared to a regular chow diet. Interestingly, despite a reduction in the total number of OTUs, on a HFD, Clostridiales remains the dominant family and their OTUs were greater in Reg3g^−/− mice compared with Reg3g^+/− mice. This finding implies that in the small intestine diet could be the major driver of oscillations with Reg3γ having a supporting role³³. The influence of other host factors, including other types of AMPs, on the composition of microbial oscillators remains to be investigated.

Intestinal immune system

The close interactions between the host immune system and the gut microbiota⁴⁷ are well-documented. It is also well-established that the immune system is controlled by circadian rhythms⁴⁸ and a disruption in this circadian control has been related to many inflammatory diseases such as arthritis^49-51. It is therefore important to understand how the circadian behavior of host immune responses determines, and is mediated by, the rhythmicity of the gut microbiota. However, this field is largely understudied and lack investigations in humans. Instead, studies have highlighted the importance of the clocks in monocytes and macrophages during the responses to pathogens⁵². For example, lipopolysaccharide (LPS) treatment of human macrophages in vitro or mice in vivo led to circadian secretion of some cytokines, especiallyIL6, which was lost with deletion of the core circadian gene, Bmal1 in the mice⁵³. Jet-lagged mice exhibited hyper-response of serum cytokines such as IL6, IL-1β, and IL-12 to LPS in vivo¹². This finding shows that the response to endotoxin administration can be influenced by the time of day. The circadian regulation in immune cells has been extensively studied and reviewed elsewhere⁴⁸ (Box 2). Besides LPS, it would be interesting to study how other gut microbiota-derived-compounds interact with the circadian dynamics of the host immune system to vary the immunological health states of the host. It is possible that many inflammatory diseases could be triggered by the breakdown of diurnally homeostatic interactions between the gut immune system and the gut microbiota. Other host factors that are known to alter the composition of gut microbiota, such as secretion of bile acids, age, or disease-induced changes in gut transit time, could also contribute to gut circadian and microbial rhythmicity and need further study.

Box 2. Circadian control in the immune system.

The immune system, like many other physiological systems, is controlled by circadian clock networks in mammals. Both innate and adaptive immunities are documented to exhibit circadian oscillations in some key processes and are influenced by their cell-intrinsic clocks and by cell-extrinsic oscillations in signals from the microenvironment. Several innate immune cells including monocytes, macrophages and neutrophils have intrinsic clocks that mediate their function. The abundance and trafficking of these cells in blood and immune organs exhibit daily fluctuations depending on both their cell-intrinsic and cell-extrinsic signals. The response of innate immune cells to pathogens including bacteria, virus and parasites show a diurnal pattern. For example, the intestinal immunity response to infection by Salmonella enterica serovar Typhimurium (S. Typhimurium) showed a diurnal variation, with a higher level of infection in the light phase compared with the level in the mid-dark phase⁷⁹. The circadian expression of the antimicrobial peptide, lipocalin-2 (LCN-2), was demonstrated to regulate this rhythmic colonization of S. Typhimurium in the colon. This finding is because S. Typhimurium is relatively resistant to LCN-2, and a higher expression of LCN-2 during the day gives S. Typhimurium a competitive survival advantage⁷⁹.

The T and B cells of adaptive immunity also exhibit circadian oscillations in their clock genes, which along with cell-extrinsic signals control their development, differentiation, and trafficking. For example, the frequencies of RORγ⁺ T helper 17 (Th17) cells increased during circadian disruption in jet-lagged mice induced from light cycle perturbation⁵¹. This change during jet-lag induction was not replicated in mice lacking Rev-erbα (Nr1d1^−/− mice) because Rev-erbα normally represses the expression of Nfil3, which is an inhibitor of Rorγt expression⁵¹. The diurnal oscillations of migratory signals and expression of their receptors on T and B cells mediate their rhythmic homing and egress from lymph nodes⁴⁸. The circadian expression of pattern-recognition receptors, such as Toll-like receptors, by antigen presenting-cells (including dendritic cells and macrophages) leads to diurnal proliferation of and cytokine production by lymphocytes⁴⁸. Although not fully defined, some studies revealed that the circadian control of metabolic pathways is also present in immune cells. For example, glycolysis is regulated by the molecular clock, which further affects diurnal immune responses of dendritic cells and macrophages. Timely targeting of metabolic pathways in immune cells has been suggested to be a new treatment approach for chronic inflammatory diseases including cardiovascular disease and arthritis ⁸⁰.

The processes by which diurnal oscillations of the gut microbiota influence and are influenced by a variety of host parameters ranging from nutrient composition, food intake timing and route, to gut mucus and AMP production, and to systemic and gut-associated immune systems are just beginning to be understood. These explorations will help elucidate the interindividual and intra-individual variations in the relationship between the gut microbiota and host health. It will lead to improved design of nutritional strategies to treat metabolic diseases such as obesity and diabetes, and to help better understand the complex underpinnings of inflammatory diseases such as arthritis and inflammatory bowel disease.

Microbial control of chronometabolism

The gut microbiota regulates the circadian rhythms of host activities; lack of the gut microbiota altered both the transcriptional and the epigenetic status of the host genes³¹. RNA-seq analysis of colonic cells of mice indicated that the oscillating genes were different between control and antibiotic-treated mice³¹. However, although major transcriptional oscillations were found in both groups of mice, the average expression level for each gene did not differ greatly³¹. The rhythmicity of the host genes was also found to be altered in GF mice compared with conventionally raised control mice and, interestingly, some of these genes are involved in metabolic pathways and inflammatory processes such as fatty acid metabolism, cholesterol synthesis, and leukocyte cytolysis⁴⁰. The change in the rhythmicity of metabolic pathways was also shown in antibiotic-treated mice; for example, genes in colonic cells for pyruvate metabolism gained new rhythmicity in antibiotic-treated mice compared to wild-type nontreated control mice³¹. This finding was suggested to be a way for the host to maintain the circadian regulation of metabolism that could no longer be controlled by the rhythmic behavior of microorganisms in the control mice. Similarly, perturbations of the gut microbiota by antibiotics also altered epigenetic configurations of host genes³¹. Chromatin immunoprecipitation (CHIP)-seq analysis for histone modification showed that rhythmic promoter and enhancer activities differed between the control and antibiotic-treated mice, whereas the rhythmicity itself was evident in both groups of mice³¹.

The effects of microbiota absence on the host peripheral clock genes seem to be tissue-specific^30,54-56,40. For example, Leone et al. showed that the germ-free status and the diet of the host mice differentially affected the expression of the core clock genes, Arntl, Clock, Per2, and Cry1, in mediobasal hypothalamus as compared to the liver³⁰. Similarly, another group reported alteration in the rhythmicity of hepatic core clock genes and nuclear receptors in germ-free mice, with an accompanying alteration in the hepatic metabolome profile⁵⁷. In mice treated with a variety of specific antibiotics, Oh et al. also found altered expression of specific clock genes such as Bmal1 and Per2⁵⁸. Furthermore, rhythmicity in gene transcripts was differentially affected in duodenum and ileum compared with liver and white adipose tissue (WAT): more transcripts gained rhythmicity in the former tissues, whereas more transcripts lost rhythmicity in the latter tissues when GF mice were compared with conventionally raised control mice⁴⁰. In this study, the authors reported only subtle changes in phase and abundance of mRNA of clock genes across these tissues. In another study, when the microbiota was ablated by antibiotic treatment in mice, although the rhythmic transcription of a number of host genes were affected, there was no change in the rhythmicity of expression of the hepatic clock gene Per2³¹. Although it is clear that there are tissue-specific differences in whether the microbiota influences clock genes, the cause for the variability in findings within a tissue needs to be further examined. For example, in these studies the mechanisms of ablating the microbiota (antibiotic treatment versus GF) could influence the results as antibiotic treatment might not achieve a complete ablation while GF mice exhibit altered physiology such as enlarged cecum compared to wild-type mice⁵⁹. Another compounding factor could be the basis of analysis, quantitative RT-PCR³⁰ in which the data were normalized to a house keeping gene versus RNA-seq analysis^31,40.

The interactions between gut microbiota and host circadian network are extended to affect host metabolic phenotypes. The changes in the gut microbiota composition in HFD-fed mice reprogrammed the hepatic clocks and resulted in metabolic aberrations such as weight gain compared with mice fed a normal chow or a low-fat diet. The circadian expressions of liver core clock genes such as Arntl and Per2 remained similar between mice fed different diets, but the time-specific transcription of lipid-metabolism-related genes (such as those involved in the biosynthesis of unsaturated fatty acids and the peroxisome proliferator-activated receptor (PPAR) signaling pathway) were altered in mice fed HFD compared to those fed a control diet. Interestingly, these effects were replicated in GF mice that received a fecal transplant of the gut microbiota derived from HFD-fed mice^30,55. Furthermore, the gut microbiota also influences sex-specific differences in rhythmicity of gene expression in metabolic organs. Some of the sex-dependent rhythmic transcriptional differences in gut, liver, and WAT were abrogated in GF mice and the affected genes were related to metabolic phenotypes. For example, the rhythmic and sexually dimorphic expression of glycolysis and bile acid metabolism⁴⁰ are abrogated and the hepatic genes related to lipid and xenobiotic metabolism are altered in GF mice. Notably, the absence of the gut microbiota led to alterations in growth hormone (GH) secretion and sexual maturation in mice, leading to attenuation of sexual dimorphism, a pattern also seen in obese mice or those lacking circadian clock genes (for example, Bmal1-knockout mice). Intriguingly, secretion of ghrelin, the GH regulator, was also decreased and injection of ghrelin restored hepatic sexual dimorphism in feminized GF male mice⁴⁰.

The gut microbiota is also involved in abnormal metabolic symptoms induced by disruption of circadian rhythms. The antibiotic-treated mice were resistant to obesity and glucose intolerance exhibiting no increase in weight or body fat content and normal glucose response, even when the sleep cycle was shifted by induction of a jet lag. Additionally, the transfer of the gut microbiota from either jet-lagged mice or humans transmitted metabolic imbalances to GF mice resulting in weight gain and glucose intolerance³⁴. Despite these associations of the bidirectional interactions between the host circadian clock and the gut microbiota, the extent to which, and the molecular basis of how, the microbiota affects the host circadian clock network are unclear. Furthermore, the mechanisms underlying the tissue-specificity of the effect of the microbiota on the peripheral clocks are just being unraveled and need to be further elucidated.

Mediators of chronometabolism

Microbial metabolites

The studies on how the gut microbiota regulates host physiology through modulation of the host circadian clock network have focused on metabolism. Several of the studies showed that the diurnal oscillations of the gut microbiota-derived metabolites affect the host’s circadian control of metabolism. One study showed that hepatic circadian gene expressions are regulated by a microorganism-derived short chain fatty acid (SCFA), butyrate, whose circadian production patterns are different in the HFD-fed and low-fat-diet fed mice³⁰. The HFD feeding altered the oscillatory behavior of a few microorganisms responsible for SCFA production. Specifically, the oscillation of Lachnospiraceae, a known butyrate producer, was greatly dampened by HFD feeding as compared with low-fat-diet feeding in mice. Accordingly, the diurnal oscillations of butyrate levels disappeared upon HFD feeding³⁰. When the effect of this change on the rhythmicity of butyrate on the host was tested in vitro using hepatic organoids established from specific-pathogen free mice and in vivo using GF mice, the butyrate treatment changed both the rhythmicity and amplitude of the circadian clock genes, especially Per2 and Arntl. Acetate, another SCFA, also altered the expression of circadian clock genes, most notably that of Per2 in the hepatic organoids. Furthermore, the exposure to NaHS, which increases microbial H₂S production, blunted the circadian gene oscillations, especially those of Per2³⁰. Thus, this study showed that a variety of microbiota-derived metabolites have the potential to modulate the host circadian clock.

Another study by different investigators suggested unconjugated bile acids to be another mediator by which the gut microbiota entrains host peripheral circadian clocks⁶⁰. Bile acids are largely secreted as conjugated compounds by the liver and are then deconjugated by gut microorganisms⁶¹. In the Caco-2 intestinal epithelial cell line model in which cells were synchronized to have the same circadian rhythms, treatment with unconjugated bile acids, deoxycholic acid (DCA) and chenodeoxycholic acid (CDCA), markedly altered the amplitude and cyclic behavior of both the primary and secondary clock genes including CLOCK, ARNTL (encoding BMAL1), PER, CRY, RORA, NRID1 (encoding REV-ERBα), DBP, and E4BP4 as compared with treatment with conjugated bile acids or vehicle⁶⁰. The changes in the expression levels of the circadian genes were also observed in vivo in C57BL/6 mice. Interestingly, the direction of changes in the expression levels of some genes were the opposite in different organs. For example, unconjugated bile acids increased Cry1 expression in ileum and colon whilst decreasing it in the liver⁶⁰. Thus, again, the effects of microorganisms on the host circadian clock might be tissue-specific, although the reason for the tissue-specificity is unknown.

Another interesting metabolite group are the polyamines, produced by both the mammalian hosts and gut microbiota. It has been suggested that the diurnal rhythmicity of polyamines in serum is affected by the absence of gut microbiota. Conversely, a diet lacking in polyamines altered hepatic circadian transcriptome with the changes paralleling those observed in antibiotic-treated mice³¹. Polyamines, in turn, are known to regulate circadian oscillations by affecting the interactions between the core clock proteins, PER2 and CRY1 in mice⁶². Whether the lack of gut microbial polyamine production directly caused the observed changes in the diurnal oscillations was not tested. Nevertheless, these observations underscore the necessity to investigate the range of microbial metabolites whose effects on host metabolism have been well-documented.

Microorganism-derived components

Bacterial components such as flagellin and LPS are also thought to mediate the host circadian metabolic homeostasis, as exemplified by the work of Wang et al⁵⁶ (Fig. 2). They demonstrated that gut microorganisms are involved in diurnal dynamics of body weight in miceby regulating the amplitude of circadian rhythms of the transcription factor, NFIL3. The regulation of NFIL3 was specific for the signals of Gram-negative bacteria such as LPS and flagellin. NFIL3 regulates body fat and ablation of its gene resulted in lower body weight in mice fed either a normal diet or a HFD. The expression of Nfil3 was also higher in control mice than in antibiotic-treated or germ-free mice⁵⁶. Through the use of knockout of various genes in mouse models, the authors further demonstrated that the microbiota-induced increase in the expression of Nfil3 involved decreased expression of REV-ERBα protein (Box 1 and 2); the effect was mediated through Myd88 signaling in CD11c⁺ dendritic cells. Briefly, flagellin and LPS are recognized by TLRs on dendritic cells, which signal to group 3 innate lymphoid cells (ILC3s) via IL-23. ILC3s in turn signal via IL-22 to phosphorylate the transcription factor signal transducer and activator of transcription 3 (STAT3) in intestinal epithelial cells (IECs). Phosphorylated STAT3 decreases the expression of REV-ERBα protein. Whereas REV-ERBα mediates the circadian rhythms of NFIL3, the microbiota via STAT3 fine-tune the amplitude of these rhythms. Consequently, compared with GF mice, the microbiota-dependent increase in the epithelial NFIL3 levels affected the expression of the genes associated with lipid uptake and metabolism in the control mice. This process increased lipid absorption by IECs and lipid export to the bloodstream for fat storage in the adipose tissue⁵⁶.

Fig. 2 ∣. Lipid metabolism regulated through the control of host circadian network by the gut microbial components.

Lipopolysaccharide (LPS) and flagellin from the gut microbiota can be recognized by either intestinal dendritic cells (DCs) or intestinal epithelial cells (IECs), each relaying the microbial signals to affect host lipid metabolism and corticosterone synthesis. Upon TLR activation, DCs releases IL-23, which signals type 3 innate lymphoid cells (ILC3s) to release IL-22. The IL-22 receptor on IECs phosphorylates STAT3, which then reduces REV-ERBα expression. Lack of RORE inhibitor, REV-ERBα, leads to increased level of NFIL3, which regulates various lipid metabolism genes such as Cd36, Scd1, Cyp2e1, and Fabp4. Correspondingly, Nfil3-knockout in mice fed with a high-fat diet (HFD) limited body fat accumulation compared with control wild-typemice on a HFD.

Another study demonstrated that gut microorganisms could be involved in diurnal dynamics of murine corticosterone production, whose excessive level can lead to diabetes⁵⁴. Using a model of antibiotic-treated mice, the authors demonstrated that in the absence of the gut microbiota, as there is no Toll-like receptor (TLR) activation, the expression of PPARα is increased and not inhibited. PPARα increases the expression of Cyp11a1, the rate-limiting enzyme in corticosterone synthesis⁵⁴. PPARα activation also increases the expression of REV-ERBα, which decreases the binding of its competitive clock regulator, RORα, to RORE. As RORα normally represses Cyp11a1 via NFIL3, the net result is an overproduction of corticosterone and increased glucocorticoid receptor activity. Accordingly, blood glucose levels were increased in the antibiotic-treated mice. Altered levels ofREV-ERBα and RORα proteins and their recruitment to RORE also could affect the first clock loop genes, Bmal1, Per and Cry and reduce TLR expression. Although a promising demonstration of how circadian clock genes rhythmically transmit their signals to host metabolic genes, these observations remain to be fully substantiated. For example, Zarrinpar et al. showed a decrease in blood glucose level with antibiotic treatment in mice⁶³ and Murakami et al. showed an increase in blood glucose level in mice on a HFD, that was not seen in antibiotic-treated mice on HFD⁵⁵. Using a variety of specific antibiotic cocktails, Oh et al. did not see any effect on the expression of hepatic glucose metabolism genes in mice⁵⁸. The discrepancy in whether glucose levels are affected in microbiota-depleted mice might have arisen from use of different antibiotic combinations and doses, the composition of baseline gut microbiota, and different sampling times. The studies on microbiota effects on corticosterone are largely centered on the hypothalamic pituitary adrenal axis; while they demonstrate that the manipulation of the microbiota alters corticosterone in mice, the results are discrepant and clearly need to be further explored in mice and extended to human studies^64-66. For example, Luo et al. found that GF animals exhibited anti-anxiety and anti-depressant-like behavior with no change in serum cortisol level, but an upregulation of six glucocorticoid pathway related genes, as compared to specific pathogen free mice⁶⁴. By contrast, Neufeld et al. reported anti-anxiety behavior in GF mice, but this behavior was associated with an increase in serum corticosterone level⁶⁵. Yan et al. further showed that rats on high salt diet exhibited a reduction in levels of Bacteroides fragilis, which led to an increase in serum and intestinal corticosterone levels⁶⁶. The final changes in the metabolic states in the host caused by the influence of the altered gut microbiota on host circadian clock, therefore, should take into consideration the compounding effects of other factors that can influence the relaying pathways.

Nuclear receptors

These studies, which suggest that nuclear receptors seem to be the molecular mediators relaying microbial signals from the lumen to influence the host’s circadian control of metabolism, segue into the well-documented observations that nuclear receptors regulate the circadian clock in a variety of mammals, including humans, and thereby are important for human health⁶⁷. For example, in human and mouse hepatic cell lines in vitro, the deficiency of the core clock gene, Bmal1, leads to N⁶-methyladenosine (m6A) modification on PPARα mRNA, to increase the expression and/or alter the rhythmic expression of some lipid metabolism-associated transcripts⁶⁸. Another important nuclear receptor, in the circadian control of host metabolism and that is regulated by gut microbiota is PPARγ. HFD-induced generation of new circadian oscillations in the hepatic transcriptional network of mice was mediated by the activity of PPARγ⁶⁹. In a separate study, the absence of gut microbiota in mice was shown to change PPARγ signaling and the downstream gene transcripts it controls in IECs and liver⁵⁵. Either HFD treatment or fecal transplant of HFD microbiota led to the upregulation of liver PPARγ and its downstream genes, Cidec, Acot2, Pltp, and Pcx, all of which are involved in lipid metabolism⁵⁵. PPARγ protein in HFD microbiota-recipient mice had expression peaks at a zeitgeber time that corresponds to 6 pm and gained cyclic activities. Both treatments, however, conferred only mild changes in the circadian clock genes, including Per2 expression and the phosphorylation state of Bmal1. These effects were reversed when HFD mice were treated with antibiotics. The antibiotic-treated mice on a HFD exhibited less weight gain and had decreased serum glucose level as compared with non-treated control mice on the same HFD diet⁵⁵. More interestingly, another study revealed that REV-ERBα mediates gut-microbiota-dependent rhythmic recruitment of histone deacetylase 3 (HDAC3) to host-metabolism-associated target genes⁷⁰. The genetic and epigenetic mechanisms of how the gut microbiota controls circadian rhythms in genes and proteins associated with host metabolism warrants further investigation.

Importance of microbial chronobiology

Although the role of gut microbiota in human health has been extensively studied, its effects on diurnal rhythms are under-investigated. Only in the past 6 years, the studies on the cross-talks between host circadian rhythms and the gut microbiota have noted the importance of microorganism-mediated chronometabolism. The gut microbiota contributes to the circadian rhythms of host metabolism either by affecting the major circadian clock components (such as Arntl and Per) or by directly affecting the activity of the effectors downstream of the core circadian clock genes. Both the gut microbiota and circadian rhythm studies further hint that the diurnal oscillations of microorganisms are likely to be the result of the circadian dynamics of host factors, such as food intake, host-derived metabolites and immune system, which are known to regulate the composition, function, and biogeography of the gut microbiota. Additionally, it is important to note that ‘gut microbiota’ composition is far from uniform, and varies down the length of the gastrointestinal tract, with subtle niches within a region⁵. For example, it cannot be fully represented by stool samples, which exclude mucosa-associated microorganisms. The luminal and mucosa-associated microorganisms could affect host metabolism differently. It is further likely that metabolism mediated by the more aerotolerant microorganisms of the small intestine might exhibit very different dynamics than that elicited by the anaerobic residents of the colon. Future studies of gut microbial chronometabolism should consider these differences.

Although many questions regarding the bidirectional interactions between host circadian clock and gut microbiota remain unanswered, these interactions should be considered when designing microbiome-based interventions to improve human health. The diurnal fluctuations of the gut microbiota might be important in defining the healthy and in delivering prebiotics, probiotics, postbiotics and fecal transplantation in a time-controlled manner. Presence of microorganisms at an inopportune time can reduce the communities’ beneficial function or render them harmful to the host. Comparing microbial oscillations and circadian oscillations of host cells could also provide a clear view of the types of microorganisms required for different physiological processes in the host and whether the major drivers of a particular process are the microbial oscillations, the host clock or a combination. The intraindividual variations in the gut microbiota throughout a day will further help explain the interindividual variability in disease progression and therapeutic responses. Indeed, a study investigated human fecal samples from 1,976 individuals to predict type 2 diabetes mellitus (T2DM) and validated the results from this group by examining two additional cohorts of 1,070 individuals or 1,399 individuals. The number of rhythmic OTUs was shown to decrease in T2DM in the group of 1,976 individuals. In the validation cohorts, 13 selected arrhythmic OTUs were able to distinguish T2DM in the samples from the cohort of 1,070 indiividuals, but not in the samples from the cohort of 1,399 individuals³⁷. Although the predictive performance was not validated in the second cohort, translationally, this finding implies a potential use of gut microbial oscillators as biomarkers to predict health and disease states of humans just by collecting fecal samples. Thus, further studies that integrate the rhythmicity of the gut microbiota to predict diseases states would also advance the use of gut microbiome markers for personalized medicine.

Conclusions

This Review primarily focuses on the links between the gut microbiota and the circadian control of metabolism. However, the links with the circadian control of other physiological and behavioral events are also worth investigating. The human body is a complex collection of organs whose functions need to be regulated in an intricately timed manner. This regulation helps the body interact with the dynamic surroundings most efficiently. Thus, the influences on host health by the diurnal dynamics of the gut microbiota cannot be overemphasized. The presence and function of the microorganisms at inappropriate times can adversely affect the host circadian networks and metabolism, contributing to the breakdown of the coordination of different host body functions and the host’s adaptation to the environment. Further studies are needed to better understand the potential influence of microbial oscillators and their mechanisms of action on host metabolism and circadian rhythms. The knowledge gained will promote the development of microbiome-based interventions for maintaining human health and for countering metabolic disorders.

Key points.

• The gut microbiome has an essential role in transducing dietary cues used by central and peripheral host circadian clocks to regulate and adapt to shifts in energy balance.

• Low-fat (lean) diets promote diurnal ‘oscillations’ of certain microbial populations that are metabolically-relevant circadian drivers.

• Western diets high in fat and refined sugars, and low in fiber influence key microbial oscillators to disrupt host circadian rhythms and metabolism to promote obesity.

• The effects of microbial oscillators on host circadian networks and metabolism might involve the production of bioactive small molecules and metabolites.

• Activation of nuclear receptors by microbiome-derived mediators is one of many mechanisms to regulate host transcriptional and epigenetic pathways that influence host circadian control of energy balance.