Circadian rhythms are a ubiquitous feature of virtually all living organisms, regulating a wide diversity of physiological systems. It has long been established that the circadian clockwork plays a key role in innate immune responses, and recent studies reveal that several aspects of adaptive immunity are also under circadian control. We discuss the latest insights into the genetic and biochemical mechanisms linking immunity to the core circadian clock of the cell and hypothesize as to why the immune system is so tightly controlled by circadian oscillations. Finally, we consider implications for human health, including vaccination strategies and the emerging field of chrono-immunotherapy.
The regular 24 h environmental cycles generated by the planet’s rotation have led to the evolution of daily circadian rhythms in almost all life forms on Earth. Circadian rhythms are driven by cell-autonomous biological clocks (Box 1), which allow organisms to adapt to and anticipate temporal changes in the environment. These biological clocks probably evolved a few billion years ago, perhaps in response to the need for single-celled eukaryotes to avoid the damaging effects of ionizing radiation and oxidative stress during processes such as cell division in which single-stranded nucleic acids become exposed1,2. In higher organisms such as mammals, virtually all aspects of physiology are regulated by internal circadian clocks, including sleep–wake cycles, behaviour and locomotor activity, body temperature cycles, cardiovascular and digestive processes, endocrine systems and metabolic and immune functions.
In the mammalian circadian system, peripheral clocks are normally maintained in phase coherence with the environment by entrainment from daily exposure to the light–dark cycle. This photic entrainment pathway operates via retinal innervation of a specialized neural circadian pacemaker of around 20,000 neurons located in a pair of hypothalamic suprachiasmatic nuclei (SCN)3. The mammalian SCN pacemaker has two unique properties: first, it is the only mammalian circadian structure that is directly entrainable to light. Second, the SCN neurons remain coupled in ectopic culture conditions, persistently generating robust rhythmic circadian cycles of electrical activity and gene expression for many months outside the body4,5. By contrast, peripheral circadian oscillators lose circadian synchrony in culture because individual cells are unable to maintain coupled phase coherence. Autonomous circadian clocks have been detected in all major organs and tissues and within many cell types of the body. Normal circadian synchrony of peripheral cells and tissues is maintained by a complex network involving neuronal signalling6, secretion of hormones (such as glucocorticoids)7 and metabolic cues8, which are driven by rhythmic feeding behaviour9,10. Although light is the key entrainment factor for the SCN, feeding-regulated metabolic cues are now recognized as being pivotal to the regulation of many peripheral clocks9,10.
Phase coherence of the organism to the normally occurring environmental light–dark cycle is known to maintain organismal fitness, as animal studies show that its disruption by abnormal lighting or feeding schedules, or due to genetic mutation of key circadian timing genes, induces pathological changes11. In support of this, human lifestyles that disrupt inherent timing systems (for example, exposure to abnormal lighting schedules in chronic shift work) are associated with an increased risk of cancer, metabolic disorders and cardiovascular and cerebrovascular disease12,13,14,15. In addition, many human inflammatory diseases exhibit rhythmicity in their pathology, including myocardial infarction, asthma and rheumatoid arthritis, as we discuss in detail later9,16,17.
Diurnal host responses to lethal infection and endotoxins were first demonstrated almost 60 years ago18, but it is only recently that studies have uncovered multiple widespread aspects of immune functions that are under the control of the circadian clockwork. These include the trafficking of immune cells, host–pathogen interactions and activation of innate and adaptive immunity, which have been discussed in recent reviews19,20,21,22. It is now apparent that circadian signals operate as a gate to control the magnitude of immune responses in a powerful time-of-day manner. In this Review, we describe the mechanisms that place the immune system under clock control. We address what is known of the molecular elements that couple the core circadian clockwork to specific output arms of the immune system and how these new understandings may lead to improved chronotherapeutic options for treatment of disease.
Core clockwork in innate immunity
Multiple types of innate immune cells have been demonstrated to possess intrinsic clocks, including monocytes23, macrophages24,25, mast cells26,27, neutrophils28, eosinophils27 and natural killer (NK) cells29 (Fig. 1a). These intrinsic timers affect the function of these cells, driving temporal gating of cell-type-specific processes (Fig. 1b). This temporal gating includes the phagocytic activity of macrophages30 and their cytokine release24,31, as well as histamine release and allergic reactions of mast cells32,33. Evidence is emerging to suggest that specific clock genes mediate different elements of immune cell physiology. Logan and colleagues showed that period circadian protein homologue 1 (PER1) is necessary for efficient function of splenic NK cells34. In the absence of Per1, these cells maintain rhythmicity but show altered rhythmic expression of IFNγ and of the cytotoxic factors perforin and granzyme B. Furthermore, rhythmicity in these factors and associated cytotoxicity can be suppressed by chronic jet lag35. In macrophages, the REV-ERB nuclear receptors are a critical nodal point between the clock and immunity. For instance, REV-ERBα (encoded by NR1D1) and REV-ERBβ (encoded by NR1D2) suppress gene expression (for example, matrix metalloproteinase 9 (Mmp9) and CX3C-chemokine receptor 1 (Cx3cr1)) in macrophages by repressing transcription from distal enhancers that are selected by macrophage-lineage determining factors36. Furthermore, REV-ERBα modulates the inflammatory functions of macrophages (including adhesion, migration and integrin activation) via direct regulation of CC-chemokine ligand 2 (Ccl2) through a REV-ERB binding motif in the promoter37 (Fig. 2a). Although not cell-specific, cryptochromes (CRYs) also directly affect inflammatory pathways through action on adenylyl cyclase38. In the absence of Cry1, basal levels of cAMP rise, leading to increased protein kinase A (PKA) activation. In turn, this induces phosphorylation of the p65 subunit of nuclear factor-κB (NF-κB) at S276, leading to NF-κB activation. Consequently, basal levels of several pro-inflammatory cytokines (including IL-6, tumour necrosis factor (TNF) and CXC-chemokine ligand 1 (CXCL1)) are increased in fibroblasts and bone-marrow-derived macrophages from Cry1−/−Cry2−/− mice38. Thus, molecular effectors of pro-inflammatory pathways are directly coupled to clock genes.
Circadian control of innate immune cell movement
The circadian clock is a key regulator of the temporal abundance of innate immune cells around the body. Studies by several laboratories have highlighted the importance of a functional clock in regulating the daily flux of inflammatory cells between bone marrow, blood and immune organs under homeostatic conditions. For instance, leukocyte numbers in blood show time-of-day-dependent variation in humans39, mice40 and hamsters41. This can be abolished by deletion of the core clock gene brain and muscle ARNT-like 1 (BMAL1; also known as ARNTL)40 or by inducing chronic SCN arrhythmia41. Leukocyte numbers also exhibit natural temporal variation in the lung42. This persists under inflammatory conditions but becomes less regular, with granulocytes rather than lymphocytes dominating as the oscillating cell type42. Gibbs and colleagues showed that neutrophil recruitment to the lung after acute inflammatory challenge is strongly circadian gated, driven by the rhythmic release of the chemokine CXCL5 from non-ciliated bronchiolar epithelial Club cells43 (Fig. 2b). Also, neutrophil recruitment to sites infected with Leishmania parasites has been shown to be circadian regulated44. Neutrophils are the most abundant leukocytes in human blood but appear to possess a weaker circadian oscillator — perhaps modified as the neutrophil develops from the myeloid progenitor28. Turnover of neutrophils is known to be under circadian control, with rhythmic clearance of aged neutrophils from the blood occurring in the bone marrow (where they are engulfed by bone-marrow-resident macrophages)45. This clearance, which occurs towards the end of the daily resting phase in mice, leads to a subsequent egress of haematopoietic precursor cells from the bone marrow in a liver X receptor (LXR)-dependent manner. Ella et al. described rhythmic daily changes in superoxide production by human neutrophils and in their ability to engulf bacteria, and this was attributed to time-of-day variations in the composition of the peripheral neutrophil pool, which in turn is a result of circadian changes in circulating chemokine levels (specifically in levels of CXCL12 and/or CXCL2)28. Thus, both trafficking and localization of neutrophils and their function are time-of-day-dependent. Similarly, IL-5 and IL-13 cytokine production by a subset of group 2 innate lymphoid cells (ILC2s) in response to hormonal cues induced by feeding influences the number of blood eosinophils and their recruitment to peripheral tissues46.
Under homeostatic conditions, the extravasation of innate immune cells from postcapillary venules to tissues is under clock control and is regulated by the rhythmic expression of pro-migratory factors (such as intercellular adhesion molecule 1 (ICAM1) and CCL2) by endothelial cells. This is dependent on local delivery of sympathetic tone and signalling through β-adrenoreceptors40. Rhythmic extravasation persists under inflammatory challenge, and in a model of septic shock in mice, neutrophil recruitment to the liver exhibited strong circadian oscillations. This influences mouse survival following administration of lipopolysaccharide (LPS; a component of the cellular wall of Gram-negative bacteria), with LPS delivery in the evening resulting in dramatically increased lethality compared with in the early morning40.
The examples discussed above describe how innate immune cell trafficking can be regulated by both intrinsic pathways and circadian signals from other cells. The trafficking of Ly6Chi inflammatory monocytes from blood to bone marrow and from blood to tissue sites of infection is under circadian control23,47. CXC-chemokine receptor 4 (CXCR4) mediates the rhythmic fluctuations in overall numbers of circulating Ly6Chi monocytes and their homing to tissue reservoirs47. During peritoneal infection with the bacterial pathogen Listeria monocytogenes, trafficking of Ly6Chi monocytes to the peritoneum is regulated by the rhythmic recruitment of the polycomb repressor complex 2 (PRC2) to the Ccl2 promoter by BMAL1–CLOCK heterodimers23. These studies show that both cell-autonomous as well as cell-extrinsic microenvironmental oscillations can govern the distribution of distinct leukocyte populations within the body.
Innate immune cell clocks in pathogen responses
At the level of the whole animal, circadian timers clearly have an impact on how the host innate immune system responds to pathogens. In part, this occurs through the regulation of inflammatory cell recruitment, where immune cell clocks are critical for maintaining effective host responses to disease23,43. Further control is exerted through the regulation of tissue-resident cells at sites of inflammation. Ingestion of the bacterial pathogen Salmonella enterica subsp. enterica serovar Typhimurium (S. Typhimurium) causes intestinal inflammation. Microarray analysis of the caecum of mice 72 h post-infection with this pathogen revealed that genes encoding antimicrobial peptides show strong circadian oscillations during infection that affect pathogen colonization of the colon48. This leads to elevated pathogen levels following infection in the morning (Zeitgeber time 4 (ZT4) (Box 2)) as compared with infection in the mid-dark phase (ZT16)48. This response is dependent on the presence of a functional copy of the circadian-regulating CLOCK (circadian locomoter output cycles protein kaput) protein as global disruption of CLOCK alters time-dependent responses to S. Typhimurium. In a Leishmania infection model, Kiessling and colleagues were able to show that parasite burden was circadian in nature. In addition, bone marrow chimaeras of Rag2–Bmal1−/− mice demonstrated that Bmal1 in non-lymphocyte immune cells (macrophages) was responsible for modulating the magnitude of Leishmania infection44. Parasites themselves can also exhibit circadian behaviour. For example, 10% of the genes of Trypanosoma brucei (parasites responsible for inducing sleeping sickness) are expressed with a circadian rhythm49. Furthermore, this circadian transcriptome affects sensitivity to treatment. The circadian clockwork also regulates cellular innate immunity against viruses. Immortalized lung fibroblasts lacking BMAL1 show altered susceptibility to viruses50. Respiratory syncytial virus (RSV) and parainfluenza virus type 3 (PIV3) (which are both non-segmented negative-strand RNA viruses) replicate at higher frequencies in the absence of BMAL1 (ref.50). Bmal1−/− mice intranasally infected with RSV had higher viral titres and increased weight loss than wild-type controls50. Along the same lines, the time of day of infection affects the host response to herpesvirus or influenza A virus, with enhanced viral titres seen when levels of Bmal1 are low or in the absence of this gene51. Therefore, the circadian clockwork can regulate cellular immunity against bacteria, parasites and viruses.
Clock control of adaptive immunity
Recently, molecular evidence has emerged to show that the adaptive immune system is also under circadian control. This evidence is perhaps surprising given that adaptive immune responses are mounted over weeks and thus were not thought to be heavily dependent on the initial time of day of challenge. However, this is in line with early reports that had initially uncovered T cell responses to be rhythmic52,53. In a similar manner to innate immune cells, circadian oscillations have been described in cells of the adaptive immune system, such as T cells54,55,56 and B cells25. In addition, dendritic cells (DCs), which are professional antigen-presenting cells (APCs) located at the interface between innate and adaptive immunity, show oscillations in core clock components25. Generally speaking, clock gene amplitudes appear to be weaker in immune cells than in other cells, such as hepatocytes55. Whether this is due to a weaker oscillating system per se, a reduced coupling capacity or increased heterogeneity in immune cell populations is currently unknown and would need to be addressed using single-cell analyses.
Role of clock genes in lymphocyte development
Circadian clock genes have been associated with the development of lymphocytes, but the extent of their involvement is not yet fully established. Deficiency of BMAL1 negatively affects mouse B cell development, with Bmal1−/− animals showing significantly reduced B cell numbers in blood and spleen57. Bmal1-deficient B cells also showed a specific deficit in production of IgG1 (but not in IgG2a, IgG2b, IgG3 or IgM) antibodies after immunization, but they exhibited no functional difference in response to LPS in vitro, suggesting that those cells that do develop are still largely functional. Adoptive transfer experiments revealed that it is not the B cell intrinsic clock but rather loss of Bmal1 in the bone marrow microenvironment that hampers B cell differentiation57. This finding is also consistent with the data of Hemmers and Rudensky, who showed that mice with a B cell-specific Bmal1 deficiency do not show altered B cell differentiation or function55. These experiments suggest that the influence of clock genes in the generation of B cells is not cell-autonomous but dependent on the microenvironment.
With respect to the influence of the circadian clock on the development of T cells, the overall T cell number appears to be unaffected by global or cell-specific Bmal1 ablation55,57. However, Yu and colleagues reported a role for the clock in regulating the differentiation of T helper 17 (TH17) cells58. TH17 cell development has been shown to be suppressed by the circadian-regulated basic leucine zipper transcription factor nuclear factor IL-3-regulated protein (NFIL3) in a cell-intrinsic manner, and Nfil3−/− mice exhibit increased TH17 cell frequencies in spleen, small intestine and colon58. NFIL3 is crucial for the development of ILCs59 and is a major regulator of TH17 cell frequency via rhythmic suppression of Rorc, which encodes RORγt, a key transcription factor necessary for TH17 cell differentiation58. Loss of ILC3s in the Nfil3−/− mice would already skew towards a higher propensity for TH17 cell differentiation as ILC3s are a significant source of IL-22, which inhibits TH17 cell differentiation. Depletion of ILCs thus enhances TH17 cell responses60. Within CD4+ T cells themselves, Nfil3 and Rorγt are rhythmically expressed in an anti-phase manner, with Nfil3 expression peaking during the dark phase (night) and Rorγt expression peaking during the light phase (day). T cells isolated during the day are therefore more prone to differentiate into TH17 cells upon stimulation than those isolated at night58. This propensity may additionally be due to the fact that during the day levels of the night-signalling hormone melatonin are low, because melatonin treatment of human T cells has been shown to inhibit TH17 cell differentiation via the NFIL3–RORγt pathway61. Instead, melatonin promotes the development of IL-10-producing type 1 regulatory (TR1) cells61. The circadian clock protein REV-ERBα directly inhibits transcription of Nfil3, and Nfil3 expression is elevated in both REV-ERBα−/− T cells and in T cells from ClockΔ19/Δ19 mice, which exhibit a lengthened circadian period and arrhythmicity in complete darkness58. ClockΔ19/Δ19 mice also displayed reduced frequencies of TH1 cells in the intestines, implying additional pathways by which the circadian clock can affect T cell polarization. The circadian regulation of T cell development may be particularly important for maintenance of an appropriate balance of cell types by time of day, as chronic circadian disruption via simulated jet lag resulted in increased proportions of TH17 cells, rendering animals more susceptible to experimental colitis58.
In addition, the circadian-regulated proteins differentially expressed in chondrocytes protein 1 (DEC1; also known as BHLHE40) and DEC2 (also known as BHLHE41) are driven by E-box mediated BMAL1–CLOCK binding. DEC1 is involved in essential CD4+ T cell effector functions, including in the regulation of a large group of CD28-dependent genes62. Antigen-specific DEC1-deficient CD4+ T cells have cell-intrinsic defects in survival and proliferation62. Meanwhile, DEC2 is critical for the development of B-1a cells, and DEC2-deficient B-1a cells show an abnormal B cell receptor (BCR) repertoire63. Taken together, the proteins of the circadian clock clearly play an important role in the function of lymphocytes.
Lymphocyte trafficking rhythms
Similarly to innate immune cells, T cells and B cells exhibit strong circadian oscillations in the blood, with peak numbers during the respective behavioural rest phase of the organism (that is, during the day in mice and at night in humans)39,54,64,65,66. These rhythms have been linked to oscillations in CXCR4 and CX3CR1 expression, regulated by glucocorticoids and catecholamines64,65, as well as hypoxia-inducible factor 1α (HIF1α) signalling67. Recent papers found that under steady-state conditions, lymphocytes within the lymph node also exhibit circadian oscillations in mice, with the highest numbers present at the beginning of the active phase (that is, at night in mice)54,66,68 (Fig. 3). These higher numbers were not due to local intranodal proliferation but depended on both the rhythmic homing of cells from blood into lymph nodes and the rhythmic egress of cells from lymph nodes into efferent lymph. A key regulatory element is the CC-chemokine receptor 7 (CCR7), which exhibited diurnal oscillations in T cells and B cells, with its expression peaking in phase with CCL21 levels on high endothelial venules in lymph nodes54. This receptor–ligand pair is critical for control of T cell migration into lymph nodes, whereas B cells also rely on additional receptors such as CXCR4 and CXCR5 (refs69,70). Peak expression levels of both CCR7 and CCL21 were observed around night onset, when the highest recruitment of CD4+ T cells, CD8+ T cells and B cells occurred54. Both the microenvironment and lymphocytes themselves contribute to this process, but T cell-specific genetic deletion of the circadian clock regulator BMAL1 was sufficient to ablate the time-dependent difference in lymph node homing and cellularity54.
Interestingly, the egress of cells into efferent lymphatics was also found to be rhythmic, with lymphocyte counts highest in lymph around ZT9 (refs54,66). This finding was dependent on rhythmic expression of sphingosine 1-phosphate receptor 1 (S1P1)54, which promotes lymphocyte egress from lymph nodes in response to the chemotactic lipid sphingosine 1-phosphate71. Cells that migrated into the lymph node at night also stayed longer within this tissue than the (fewer) cells that had immigrated during the day. Suzuki and colleagues demonstrated that this is dependent on lymphocyte expression of β2-adrenergic receptors72 and rhythmic environmental sympathetic tone66. β2-Adrenergic-receptor-deficient lymphocytes transited through lymph nodes more quickly and failed to show a diurnal pattern66. This finding indicates that enhanced sympathetic tone and engagement of the β2-adrenergic receptor at night helps maintain cells in lymph node compartments, likely by boosting lymphocyte expression of the retention factors CCR7 and CXCR4 (ref.72). This longer lymph node dwell time plays an important role in the generation of adaptive immune responses54,66. Thus, time of day both dictates expression of critical pro-migratory molecules on lymphocytes and endothelial cells and governs the trafficking behaviour of lymphocytes from blood through lymph nodes and into efferent lymphatics.
Rhythmic adaptive immune responses
Initial observations by Esquifino and colleagues showed that the adaptive immune system exhibits functional rhythmicity. They demonstrated that proliferative responses of rat lymph node cells following treatment with LPS or concanavalin A were strongly rhythmic, with significantly stronger proliferation occurring at midday than at midnight73. Subsequent studies by Fortier and colleagues confirmed these observations and further demonstrated that this was dependent on the circadian clock. Mice bearing the circadian mutation ClockΔ19/Δ19 exhibited strongly blunted rhythms of proliferative responses of T cells after direct stimulation of the T cell receptor (TCR)74. Furthermore, they found that intravenous injections of ovalbumin (OVA)-loaded DCs into recipient mice at ZT6 (daytime), as opposed to at ZT18 (night-time), generated higher numbers of OVA-specific T cells within the spleen. These data provide evidence for the importance of circadian timing in the interaction between T cells and APCs.
For an adaptive immune response to be generated, APCs must functionally interact with lymphocytes. Silver and colleagues were the first to investigate a rhythmic link between the adaptive and innate immune system. They demonstrated that the expression of Toll-like receptor 9 (TLR9) — which recognizes bacterial CpG DNA — but not of other TLRs by macrophages, DCs and B cells, exhibited oscillations that were controlled by the circadian clock gene Per2 (ref.75). In line with this, mice treated with OVA coupled to CpG at the time of enhanced TLR9 expression (ZT19 versus ZT7) showed elevated OVA-specific adaptive immune responses several weeks later75. Lymph node cultures from these mice showed higher levels of OVA-induced lymphocyte proliferation and production of IFNγ at ZT19. In addition, lymphocytes from mice lacking a functional PER2 protein showed a stronger immune response than wild-type cells75.
These data demonstrate that circadian differences in the expression of pattern-recognition receptors (PRRs) by APCs can drive rhythmic adaptive immune responses and are thus under the control of the circadian clock. However, it is currently unclear how the observed differences in the time of day of the acrophase of the response arise. Whereas Silver and Suzuki described stronger immune responses at night (ZT17 and ZT19, respectively)66,75, Fortier and Druzd observed peak responses during the day (ZT6 and ZT8, respectively)54,74. An explanation for these discrepancies might lie in the type of response elicited: injection sites (subcutaneous, intraperitoneal, intravenous or intradermal), readout tissue (lymph node versus spleen), lighting conditions, responding cell type and antigen choice all contribute to such variations, particularly whether the antigen needs to be taken up by APCs to be transported to lymph nodes or can move there directly in a soluble form.
Recent vaccination studies also showed diurnal rhythmicity in adaptive immune responses. Responses to nitrophenyl (NP), a soluble antigen that is directly transported to lymph nodes, or to myelin-oligodendrocyte glycoprotein (MOG), which is an antigenic target in certain models of experimental autoimmune encephalomyelitis (EAE), were higher at the times when more cells were present in lymph nodes54,66,76. Stimulation at high cellularity times resulted in a stronger humoral response, as demonstrated by higher NP-specific antibody titres, as well as more NP-specific IgM+ germinal centre B cells 7 days after immunization66. In addition, in the EAE model a stronger cellular immune response was observed with more IL-17-producing TH17 cells being present, more IL-2 being made, higher leukocyte infiltration into the central nervous system and ultimately a higher disease score 14 days after inoculation54. Importantly, the humoral immune response to NP became arrhythmic in mice lacking the β2-adrenergic receptor66, and animals with a myeloid-cell-specific or T cell-specific deficiency in BMAL1 became arrhythmic in the EAE model of autoimmunity54,76. Rhythmicity was not restricted solely to sterile immunization responses but was also observed following infection with pathogens, such as influenza A virus and the bacterium Helicobacter pylori54.
Current research efforts now aim to take these findings into the clinic and establish novel ways to enhance responses to vaccines. A randomized trial of different times of day of vaccination of three strains of influenza virus yielded promising results. When 276 patients (over 65 years of age) were vaccinated in the morning they had greater antibody titres 1 month later than patients vaccinated in the afternoon77. These results are encouraging, but they will need to be validated in other preclinical laboratory models and clinical trials using larger patient cohorts. Taken together, this points to a strong role of the circadian clock in the polarization of the adaptive immune response, which is strongly dependent on the initial time of day of antigen challenge.
Innate versus adaptive immune clocks
Whereas myeloid-specific ablation of the circadian gene Bmal1 leads to a general pro-inflammatory phenotype in mice that is characterized by higher levels of TNF, IL-6, IL-1β and CCL2 (refs23,31,78,79), the phenotype of T cell-specific Bmal1 deficiency appears to be more time-of-day-dependent54,55. Whether this is because the currently employed genetic targeting strategies for lymphocytes (CD4–Cre for T cells54,55) are more specific than for myeloid cells (that is, where Lyz2–Cre is currently used, thus targeting most myeloid cell populations23) is unclear. To better compare this, future studies will need to use promoters that are specific for neutrophils, monocytes or macrophages for Bmal1 ablation. Infection of mice with L. monocytogenes resulted in a lower percentage of IL-2+, IFNγ+ and TNF+ T cells in mice with a T cell-specific deficiency in Bmal1 than in control animals55. In addition, in models of EAE, mice with a T cell-specific Bmal1 deficiency had lower disease scores than controls54 whereas mice with myeloid cell-specific loss of Bmal1 showed exacerbated pathology76. This finding indicates that the outcome of Bmal1 deletion on inflammation is lineage dependent.
Microbial regulation of circadian immunity
Recent evidence demonstrates that symbiosis between microbiota and intestinal epithelial cells is critical in maintaining immune homeostasis in the gut. Intriguingly, this interaction is orchestrated by the circadian clock and cues from the microbiota that activate intestinal TLRs. Mukherji and colleagues showed that the anti-phasic binding of RORα and REV-ERBα to TLR genes acts as an activator and repressor, respectively, to mediate the circadian expression of TLR1–TLR5 and TLR9 (but not TLR6 and TLR7)80. This oscillatory promoter occupancy drives rhythmic activation of the key pro-inflammatory transcription factors activator protein 1 (AP-1) and NF-κB to exert their functions at diurnal times ZT20–ZT4 (which is the active phase in mice). Antibiotic-mediated ablation of the microbiota disrupts this rhythmicity and induces a pre-diabetic syndrome in mice, characterized by hyperglycaemia and hypoinsulinaemia80. Intriguingly, rhythmicity could be restored by administration of LPS. Additional recent findings provide a link between rhythmic Nfil3 expression in intestinal epithelial cells, microbiota and host metabolism that is mediated by immune cells81. The induction of TLR–MYD88 signalling by bacteria stimulates DCs to secrete IL-23 and activate ILC3s. This activates signal transducer and activator of transcription 3 (STAT3), leading to an inhibition of REV-ERBα, a negative regulator of NFIL3. Epithelial NFIL3 regulates lipid uptake and export in intestinal epithelial cells so that mice with an intestinal-specific deletion of Nfil3 are resistant to obesity induced by a high-fat diet81. These data indicate that the microbiota is a previously unrecognized and important general regulator of immune rhythmicity in the host (Fig. 4).
Microbial products shape circadian immunity
LPS also appears to be a critical factor in shaping the activation status of neutrophils that are present in blood. Although blood neutrophils have long been thought to be a relatively homogeneous population, heterogeneity arises owing to differences in their age and activation level. Neutrophils are short-lived and — after having been mobilized into the blood in the morning45 — age in blood over the course of the day45,82. Aged neutrophils exhibit higher levels of CXCR4 on their surface and lower levels of L-selectin45,83,84 and are present at higher concentrations during the day in mice45. This is associated with an enhanced activation status83, including an increased capacity to form neutrophil extracellular traps82. Microbiota-derived LPS drives this ageing process82. Mechanistically, this is modulated by TLR2- and TLR4-expressing haematopoietic cells, as well as by MYD88-expressing myeloid cells. Ablating the microbiota dramatically reduced the number of aged neutrophils and was also effective in alleviating the pathogenesis of sickle cell disease and endotoxin-induced septic shock82. Thus, rhythmically recurring cycles of neutrophil activation in the circulation can have dramatic consequences on the time-of-day-dependent worsening of inflammatory disorders.
Many bacterial genera within the gut microflora of mice and humans exhibit circadian oscillations themselves. For example, Lactobacillaceae show strong oscillations in general abundance over the course of the day, peaking at the onset of the mouse active phase85,86. This rhythmicity is driven by feeding rhythms and can be ablated by jet-lag protocols (repeated shifts in timing of light–dark cycles) or antibiotic treatment85,86. Chronic jet lag in mice leads to obesity, whereas in humans jet-lag protocols result in a type 2 diabetic phenotype86. This finding indicated that aberrant circadian rhythms in humans can lead to metabolic disturbances. Gut bacteria exhibit circadian rhythmicity in terms of their numbers and activity, and their rhythmic function is also important for the systemic regulation of host tissues8. This is achieved by a specific rhythmic signature of metabolites present in blood and is important in driving epigenetic changes in host tissues distant from the gut, such as the liver8. Thus, rhythmic release of metabolites and LPS from the microbiota may be key in directly driving important rhythmic immune-related phenotypes, such as neutrophil ageing and associated activation45,82. In addition, rhythms in commensal bacteria appear to help synchronize the body to the external environment and thus indirectly influence a rhythmic immune system.
Synchronization of the immune system
Microbial products and feeding rhythms can synchronize peripheral clocks in tissues87 but it is yet unclear how immune cells — most of which are migratory — are synchronized in vivo to the external environment. Initial studies focusing on circulating hormones were able to demonstrate a correlative link between glucocorticoids and adrenaline and numbers of circulating T cells in blood by modulating surface CXCR4 and CX3CR1 levels65. It has now been shown using cell-specific disruption of glucocorticoid signalling that rhythmic T cell trafficking is regulated via glucocorticoid receptor (GR)-dependent induction of the IL-7 receptor (IL-7R) and CXCR4 (ref.68). Dexamethasone88 and high transient serum concentrations (a so-called serum shock)87 are used to synchronize cells in culture, indicating that both glucocorticoids and feeding rhythms are important for the synchronization of migratory cells in vivo. With respect to the sympathetic nervous system, lack of β2-adrenergic or β3-adrenergic receptors resulted in an ablation of the rhythmic trafficking behaviour of leukocytes to tissues40,66. However, it is currently unclear whether the clockwork in immune cells and/or stromal cells was also affected in this case. Furthermore, a recent surprising study has found that the cell-intrinsic clock is sufficient for the rhythmic migration behaviour of leukocytes67. Zhao and colleagues showed that in humanized mice, human and mouse leukocyte numbers within the same organism exhibited inverted oscillations in blood, reproducing the trafficking pattern previously observed in both species. This cyclic pattern of trafficking correlated with rhythmic CXCR4 expression, which was regulated by a HIF1α–BMAL1 heterodimer. p38 mitogen-activated protein kinases (MAPKs) and MAPK kinase activated protein kinase 2 (MK2) exhibited opposite effects between mice and humans in generating intracellular reactive oxygen species, which subsequently regulated HIF1α expression67. This indicates that immune cells use reactive oxygen species as their synchronization cue. Taken together, a complex network of factors appears to be important for immune cells to be synchronized to external time.
Potential benefits of rhythmic immune responses
A key question in our field is why is the immune system so tightly controlled by circadian oscillations? The most striking examples of timing in the immune response are mice that succumb to high doses of pro-inflammatory mediators or pathogen more prominently in the afternoon18,89. However, at the same time, mice are able to fight off bacteria more effectively when administered at a more physiological dose that is more likely to be encountered in nature23. Thus, this heightened immune sensitivity would be an adaptation to potentially recurring environmental challenges, with the additional effect of reducing detrimental immune responses (such as a high activation status of neutrophils) at times when they are not needed. Thus, a timed immune response — with a refractory phase in between, possibly for opposing immune processes to occur — may ultimately increase the survival of an organism. Evidence for this lies in the fact that both immune cell-intrinsic and cell-extrinsic (such as endothelial and stromal cell) rhythms are coordinated in a highly complex fashion to balance the recruitment and activation of immune cells. An alternative explanation could be that rhythmic immune responses may have evolved as a secondary consequence of rhythmic metabolic processing within the immune cell itself. For the latter, the key benefit of circadian rhythmicity would be the ability of cells to partition metabolic flux and detrimental redox processes in time and space9. Clearly, metabolic and immune processes are intricately linked, and both affect each other also with respect to circadian timing. For example, metabolic cues are critical for the entrainment of peripheral clocks, such as those present in immune cells. In addition, disruption of clock gene function results in a host of metabolic disorders, which are often associated with immune phenotypes90, such as mice with a myeloid cell-specific deletion of Bmal1 that present with a pro-inflammatory phenotype and over time become obese23. Thus, partitioning of immune cell function to specific times of the day likely demonstrates an adaptation of the organism to pathogen exposure in addition to detrimental endogenous processes.
Numerous animal models show that genetic or environmental disruption of the clock can expedite the development of inflammatory disease or exacerbate disease pathology91,92,93,94. This translates to humans, as employment in rotating shift work has been associated with an increased risk in developing inflammatory diseases such as psoriasis95 and irritable bowel syndrome96. In keeping with this, Cuesta et al. report desynchrony in rhythmic immune parameters after simulated night-shift work97. Of course, rotating shift work affects not just biological rhythms but also other lifestyle habits (sleep quality and quantity, diet, mealtimes and exercise) that may potentially affect the development of such disorders. Nevertheless, this is a major concern for society. Together, animal studies and clinical data corroborate the notion that a healthy intact clock is necessary to regulate immune homeostasis and to help prevent the development of immune disorders. In addition to this role under homeostatic conditions, the clock continues to be critical in regulating inflammation once disease is established. It is well recognized that human chronic inflammatory diseases can show time-of-day variation in not only symptoms but also disease markers. Most often cited is the nocturnal component to asthma, whereby symptoms show exacerbations in the early morning98. Similarly, patients suffering from rheumatoid arthritis often report increased joint stiffness and pain in the early morning99. These periods of increased symptoms correlate with peaks in disease markers, indicating that these phenomena are not simply a consequence of sleep itself (for example, caused by limited movement and altered breathing patterns). There is an early morning peak in the number of pulmonary eosinophils in patients with asthma100 and in circulating cytokine levels in patients with rheumatoid arthritis101. The onset of cardiovascular events often shows time-of-day variation, with acute myocardial infarction and ventricular arrhythmias more common in the morning102. This timing is likely a consequence of circadian variation in blood pressure, the coagulation cascade and vascular function. Furthermore, there is evidence that in the days following the transition to or from daylight savings time, there is an increase in the incidence of acute coronary syndrome related to the desynchrony between the internal clock and the external environment103. Given the wide range of effects of the circadian clock on immune function, it is not surprising that in a chronic setting, there are rhythmic fluctuations in inflammatory pathways, sufficient to cause physiological effects. Our current understanding of these interactions is very limited, but it is clearly important that we gain deeper knowledge as this is likely to lead to novel treatment options.
Seasonal diseases and underlying mechanisms
There is emerging evidence that the immune system also shows seasonal fluctuations in man and animals; indeed, this may explain why some inflammatory diseases have an additional seasonal component. Within circulating immune cells, approximately 23% of the genome shows seasonal variation in expression104; furthermore, the cellular makeup of the peripheral immune system varies according to season104. It is not yet established how annual changes in the external environment modulate seasonal changes or to what extent seasonal changes affect immune function over the year. However, this new information may explain in part why some immune disorders exhibit seasonal variation in the occurrence of exacerbations. Dopico and colleagues noted a seasonal fluctuation in Bmal1 expression, with lower levels of expression during the winter months104. Herpesviruses and influenza A viruses replicate more efficiently in the absence of BMAL1, perhaps in part explaining why viral dissemination is more common in winter51. Individuals diagnosed with the debilitating lung condition chronic obstructive pulmonary disease (COPD) are also more prone to exacerbation during the winter months105. This is thought to be the result of a combination of increased prevalence of respiratory viral infections, lifestyle changes during the colder months and reduced vitamin D levels. Conversely, disease activity in patients with multiple sclerosis peaks in early spring and troughs in autumn106. This trend may be attributed in part to seasonal variation in the nocturnal pattern of secretion of pineal melatonin, with longer nocturnal durations in the autumn and winter61. These examples demonstrate that there may be important seasonal changes in the phasing and exposure of circadian-regulated signals, tuned to the environment, that can affect our physiology and health.
Our knowledge regarding circadian variation in the expression and activity of drug targets is steadily growing. A recent study has shown that a high proportion of the best-selling drugs used in the treatment of inflammatory diseases (as well as drugs for metabolic diseases, mental health disorders and cancer (Box 3)) directly target the products of genes that are themselves under clock control107. Many of these drugs typically have short half-lives (less than 6 h); consequently, timed application in line with target expression may markedly improve treatment efficacy (Fig. 5). Surprisingly, as yet there is very little use of chronotherapy for the treatment of inflammatory disorders in a clinical setting. One success involves the use of glucocorticoids in the treatment of rheumatoid arthritis. A slow-release formula of prednisone has been manufactured, designed to be taken at bedtime, resulting in release into the bloodstream 4 h later, and phased to the peak in expression of inflammatory markers, which reduced the duration of morning stiffness of the joints108. We now know that the circadian clock (through the actions of the CRY and REV-ERB proteins) regulates the response of the GR to its ligands109,110. Thus, timed application of anti-inflammatory glucocorticoids may enable maximal therapeutic benefit while at the same time reducing undesirable hepatic metabolic side effects of these heavily used drugs109.
The clock as a therapeutic target for inflammation
Our growing understanding of how components of the molecular clock interact with critical elements of inflammatory pathways offers the exciting potential for the use of pharmacological agents that target these clock proteins as anti-inflammatory agents. The challenge now is to produce high-affinity, high-efficacy molecules that enhance the activity of clock proteins. Recently, a number of tool compounds have been developed that show, in principle, the benefits of pharmacological modulation of clock proteins on the outcome of inflammatory challenge. SR9009 is a synthetic REV-ERB agonist111 that has been shown to decrease atherosclerotic plaque burden after chronic administration in mice112. Conversely, application of a synthetic REV-ERB antagonist (SR82778)113 has detrimental effects in a mouse model of viral-induced encephalitis114. In addition, the compound KL001 acts to stabilize CRY proteins by preventing ubiquitin-driven degradation115. Application of KL001 in vitro has anti-inflammatory effects on fibroblast-like synoviocytes116, and this presents as another potential anti-inflammatory target. Taken together, the future will see the development of novel tools targeting the clock in inflammatory disorders.
The immune system is heavily influenced by time-of-day cues, both under steady-state conditions and in response to inflammatory challenges. The regulation of immune responses according to the time of the day is likely optimized to provide protection at specific times. In addition, time-partitioning of metabolic and redox processes in immune cells may help minimize detrimental effects on the organism. Shifts of a few hours in the timing of an inflammatory challenge targeting the innate or the adaptive immune system have dramatic consequences on subsequent immune responses and pathological outcomes, not only in acute immune responses but also in chronic responses manifested several days or even weeks later. Despite the rapid recent progress made in understanding the mechanisms involved, the role of the circadian-controlled expression signatures of specific organs and crosstalk to other body systems remains very poorly understood. An improved understanding of these pathways is necessary in order to advance chronotherapeutic options. From this, novel drugs and treatments that could maximize therapeutic benefits at specific phases while minimizing off-target side effects may then be developed.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Manchester Biological Timing: https://www.bmh.manchester.ac.uk/research/biological-timing/
Loudon Laboratory: http://www.manchester.ac.uk/research/Andrew.loudon/personaldetails
Scheiermann Laboratory: http://scheiermannlab.de/
Gibbs Laboratory: https://www.research.manchester.ac.uk/portal/Julie.Gibbs.html
The authors thank V. Lavilla for creating the video and M. F. Loudon for providing the voice-over to it. C.S. is funded by the German Research Foundation (DFG) (Emmy-Noether grant (SCHE 1645/2-1) and SFB914 projects B09 and Z03), the European Research Council (ERC; starting grant 635872, CIRCODE), the DZHK (German Centre for Cardiovascular Research) and the BMBF (German Ministry of Education and Research). J.G. is an Arthritis Research UK Career Development Fellow (Ref. 20629). A.L. acknowledges the support of the Wellcome Trust (grant 107851/Z/15/Z).
Movie 1: Rhythmic leukocyte activity throughout the body. Migration of leukocytes from blood to tissues (and back) is regulated at multiple levels by the circadian clock. In mice, blood leukocyte content is high during the day (rest phase) and lower at night (active phase). The factors, which generate these oscillations vary between cells and tissues. For example, neutrophils in the blood express higher levels of the chemokine receptor CXCR4 during the late day. This receptor drives neutrophil homing to bone marrow, which is therefore elevated at this time point. During the night, CXCR4 expression is reduced and less homing occurs to this organ. In the lung, resident stromal cells rhythmically produce the neutrophil chemoattractant CXCL5. Inflammatory challenge by lipopolysaccharide (LPS) inhalation during the day stimulates greater production of CXCL5 than challenge at night. The differential production of chemoattractant, along with greater numbers of neutrophils in blood, therefore leads to increased neutrophil influx to the lung during the day. By contrast, cells largely home to lymph nodes at night. During the day, T cell and B cell expression of the lymph node-homing receptor CCR7 is low, and few cells migrate into the lymph node. In addition, expression of S1PR1, the receptor which mediates lymphocyte egress, is high and cells are more prone to leave the lymph node during the day. At night, the inverse occurs and T cells and B cells are retained in the lymph node for longer. Differentiation of cells is also regulated by time-of-day, as in the case of TH17 cell development in the gut. During the day, levels of the differentiation factor RORγt are high and increased differentiation is observed relative to the dark phase. At night, RORγt activity is repressed by NFIL3 and so the differentiation stimulus is reduced. In this way, the body is primed to respond differently to inflammatory challenge at different times of day, and disruption to the circadian rhythm can have severe consequences for immune function
A free-running rhythm with a period of approximately 24 h that persists in the absence of external entrainment, such as in constant darkness.
- Suprachiasmatic nuclei
(SCN). A bilateral structure in the anterior hypothalamus, home to the central pacemaker, which processes light input and conveys timing information to the rest of the body.
A pattern that occurs over the course of a day in which external entrainment (such as light–dark cycles) is used; the onset of the light cycle is defined as Zeitgeber time 0 (ZT0).
- Period circadian protein homologue 1
(PER1). PER1, PER2 and PER3 are PAS (PER–ARNT–SIM) domain-containing proteins that associate with CRY proteins to inhibit BMAL1–CLOCK-mediated gene expression.
REV-ERBα (encoded by NR1D1) and REV-ERBβ (encoded by NR1D2) are transcriptional repressors that bind to ROR response element (RORE) motifs in the BMAL1 promoter to regulate the rhythmic expression of BMAL1.
(CRYs). CRY1 and CRY2 are transcriptional repressors that associate with PER proteins to inhibit BMAL1–CLOCK-mediated gene transcription.
- Brain and muscle ARNT-like 1
(BMAL1). A basic helix–loop–helix PER–ARNT–SIM (bHLH–PAS) domain transcription factor that dimerizes with CLOCK to bind E-boxes in gene promoters to induce circadian gene expression.
(Circadian locomoter output cycles kaput). A basic helix–loop–helix PER–ARNT–SIM (bHLH–PAS) domain transcription factor that can dimerize with BMAL1 to regulate circadian gene expression.
- Zeitgeber time
Zeitgeber, literally ‘time giver’, is a cue (such as light) that entrains the circadian clock. Zeitgeber time (ZT) is the time after light onset; for example, lights on is ZT0 and lights off is ZT12 in a 12 h light-12 h dark cycle.
- Circadian time
(CT). A measure of subjective time used when organisms are isolated from Zeitgebers (for example, constant darkness). CT0 represents the start of subjective day and CT12 represents the start of subjective night.
The nuclear receptors RORα, RORβ and RORγ are transcriptional activators that bind to ROR response element (RORE) sites in target gene promoters.
About this article
A selective window after the food-intake period favors tolerance induction in mesenteric lymph nodes
Mucosal Immunology (2018)