The circadian regulator BMAL1 programmes responses to parasitic worm infection via a dendritic cell clock

Resistance to the intestinal parasitic helminth Trichuris muris requires T-helper 2 (TH2) cellular and associated IgG1 responses, with expulsion typically taking up to 4 weeks in mice. Here, we show that the time-of-day of the initial infection affects efficiency of worm expulsion, with strong TH2 bias and early expulsion in morning-infected mice. Conversely, mice infected at the start of the night show delayed resistance to infection, and this is associated with feeding-driven metabolic cues, such that feeding restriction to the day-time in normally nocturnal-feeding mice disrupts parasitic expulsion kinetics. We deleted the circadian regulator BMAL1 in antigen-presenting dendritic cells (DCs) in vivo and found a loss of time-of-day dependency of helminth expulsion. RNAseq analyses revealed that IL-12 responses to worm antigen by circadian-synchronised DCs were dependent on BMAL1. Therefore, we find that circadian machinery in DCs contributes to the TH1/TH2 balance, and that environmental, or genetic perturbation of the DC clock results in altered parasite expulsion kinetics.

manner 13 . A significant proportion of the expressed transcriptome within tissues is circadian-regulated with reports of up to 43% of protein-coding gene transcripts oscillating in at least one organ 14 . Circadian rhythms play an important role in the regulation of immune function 15 . Both monocytes of the innate immune system and cells of the adaptive immune system such as T and B cells possess components of the circadian clock machinery, which generate autonomous oscillations in vivo [16][17][18][19][20][21][22][23][24] . Further, trafficking and migration of immune cells is also circadian regulated, with daily oscillations in lymphocytes in blood 19,[25][26][27] , T-helper-17 (T H -17) cell differentiation 28 , and lymphocyte trafficking through the lymph node 29,30 . Further, in the context of intracellular protozoan parasite infection, circadian regulation of the magnitude of Leishmania infection has recently been described with an underpinning mechanism, thought to involve circadian control of chemokine expression by macrophages and thus rhythmic infiltration of neutrophils and macrophages to the site of infection 31 .
It is not known whether the circadian clock plays a role in immunity to large, extracellular, multicellular parasites such as gastrointestinal helminths. Here, we show that mice infected with T. muris demonstrate different T H cell responses and consequently different levels of effective parasite expulsion several weeks later, dependent of time-of-day of infection. We show that antigen-presenting DCs are key components; with disruption of time-of-day responses observed in mice with a DC-conditional knock-out of the circadian regulating Bmal1 gene. Circadian-synchronised DCs in vitro showed strongly polarised responses at peak and trough of the circadian cycle, and these were disrupted in Bmal1-deficient cells. Collectively our data reveal that a DC circadian clock regulates efficient expulsion of the parasitic nematode T. muris.

T. muris expulsion kinetics depends on time-of-day of infection. Different inbred strains of mice
vary in their ability to expel a T. muris infection 32 with some strains expelling the parasite within three weeks of infection, while others require more than four weeks, and yet further strains are completely unable to clear the infection. We used the C57BL/6 mouse strain which mounts an immune response characterised by both T H 1 and T H 2 cytokines, typically leading to expulsion of parasites approximately three to four weeks later. As in all mouse strains, C57BL/6 mice become infected by the parasite with no expulsion before day 13 post infection (Schematic Fig. 1A; Establishment). Subsequently the developing adaptive immune response, characterised by both T H 1 and T H 2 cytokines, leads to expulsion of parasites approximately three to four weeks later (Resolution). If C57BL/6 mice become more biased towards a T H 1 immune response, this results in delayed-expulsion kinetics, with an elevated worm burden, as indicated by the red line, Schematic Fig. 1A; Expulsion).
The impact of time-of-day of infection on long-term parasite expulsion dynamics was measured using a dose of 200 T. muris eggs. Mice were maintained on a 12 L:12 D light-dark cycle, and eggs administered via oral gavage at lights on (ZT0; Zeitgeber Time 0) or 12 h later at lights off (ZT12). Animals were sacrificed at Day 21 or 28 post-infection, when mice would normally have started to expel or almost completely expelled their worm burden respectively 32,33 . Mice infected at ZT12 had a significantly higher T. muris worm burden at Day 21, but both groups had largely cleared worms by Day 28 (Fig. 1B). Infection of any mouse strain with T. muris eggs results in the establishment of larval stage parasites around two weeks post infection, with this infection subsequently expelled if a dominant T H 2 immune response is generated 32,34 . It is possible therefore, that the differences in expulsion kinetics observed were simply attributable to time-of-day differences in the initial extent of establishment of infection. To assess this, we quantified worm numbers in mice infected at Day 13, a time-point prior to the normal onset of expulsion. We observed no significant differences between ZT0 and ZT12 infected mice (Fig. S1A). Thus collectively this suggests that long-term programming of the adaptive immune response to parasitic worm infection is determined by the time-of-day of infection, but with manifestation of the effect occurring several weeks after infection.
IgG1 and IL-13, typical markers of the T H 2 response associated with effective worm clearance, were significantly higher following a ZT0 infection (Fig. 1C,G). Total IgE and mucosal mast cell protease-1 (MCPT-1), which also correlate with the T H 2 response associated with Trichuris infection were significantly elevated in ZT0-infected mice (Fig. 1E,F). In contrast, IgG2c and IFN-γ, markers of a worm-susceptible T H 1 immune response, were significantly raised following a ZT12 infection (Fig. 1D,H). Here, IgG2c was measured at day 28 post-infection as the optimal time point post-infection to detect differences in this antibody isotype, At day 21 post infection parasite specific IgG2c levels are usually too low in sera to use as a surrogate marker of delayed expulsion. In subsequent experiments we focussed our final autopsy on day 21 in order to monitor the delayed worm expulsion phenotype as our main readout of clock regulation. Leukocyte populations infiltrate the intestine after a T. muris infection, with DCs then migrating to the local draining lymph nodes where they initiate T cell polarisation 35 . To assess the early leukocyte recruitment to the gut we quantified two key sentinel cells, the macrophage (MHC II + , F4/80 + CD11b + ) and the DC (MHCII + , CD11c + , CD103 + Dec205 + ), in the large intestinal lamina propria by flow cytometry (Fig. 1I, on Days 0, 1, 2 and 3 post infection) and immunohistochemistry (Fig. 1J,K, on Day 0 and Day 1 post infection). These data revealed no significant differences in numbers, which might underlie the observed time-of-day of infection phenotype despite a non-significant trend for fewer macrophages in the lamina propria prior to infection at ZT12. Therefore, the early local intestinal immune response to post worm infection was not affected by time of exposure.
To establish the range of time-of-day influences over worm expulsion we further infected mice at the two intermediate time points (ZT6 and ZT18; Fig. S1B) and assessed worm burdens at day 21 post-infection. These did not differ significantly suggesting that the difference in time-of-day of worm expulsion is at its greatest when comparing ZT0 to ZT12. Analyses of antibodies and cytokines revealed a significant increase in IgG1 after a ZT6 infection and a significantly elevated IFNγ response after a ZT18 infection (Fig. S1C,F) suggesting that the circadian influence over the immune response is still apparent at these two circadian times, but insufficiently polarised to result in a delayed worm expulsion. Thus all subsequent analyses focussed on ZT0 and ZT12. Collectively these The role of metabolic cues and DC subsets in parasite expulsion. We next assessed whether metabolic feeding-associated cues might be involved in directing circadian regulation of worm expulsion. Feeding cues are now widely recognised as a powerful Zeitgeber for peripheral oscillators 36 . To address this we adopted a restricted feeding regime (RFR 37 ;), whereby mice (which usually feed mainly at night) have restricted access to food in a 6 h window centred either at mid-light or mid-dark phase. This RFR was imposed for two weeks prior to, and one week post-infection at ZT0 (Fig. 2A). Analysis of core circadian genes (per2, bmal1, dbp) in the intestine at the time of infection revealed that restricting feeding to the daytime reversed the pattern of expression of local clock genes (Fig. 2B). Analysis of the same circadian genes in the liver demonstrated similar outcomes (Fig. S2).
Mice restricted to feeding in the daytime two weeks prior to a ZT0 infection harboured significantly more worms at Day 21 post-infection compared to night-fed mice (Fig. 2C). Surprisingly unlike the time-of-day infection experiments, the effect of food reversal on expulsion kinetics was not associated with alteration in the IgG1/ IgG2c balance, with the IgG1 antibody isotype raised following daytime feeding in addition to IgG2c (Fig. 2D,E). The expression of T H 2 (il-13, Retnlb, ccl2)-associated genes in the local intestinal tissue likewise was elevated following daytime feeding (Fig. S3). The data thus suggests that a transient period of food reversal prior to and early in the infection process is sufficient to regulate the circadian machinery in the gut to impair worm expulsion even after returning to ad libitum feeding, but without perturbing the T H balance.
Infection occurs on ingestion of embryonated parasite eggs, which hatch approximately 90 minutes later upon reaching the caecum 38 . Hatching releases the first stage larva (L1), which burrows into intestinal epithelial cells, and triggers the immune system 35 . Antigen-bearing DCs leave the gut mucosa via the lymphatics, arriving in the mesenteric lymph nodes (MLN) and promote T cell activation from approximately Day 7 post-infection 35 Table S1). However the relative percentage of T cells (both CD4 + and CD8 + ) was significantly elevated in day-fed animals ( Fig. 3B, Table S1).
In order to investigate changes in cDC subsets, total cDCs were broken down into CD11b hi /CD103 − , CD11b + / CD103 + and CD11b − /CD103 + populations. This separation revealed significant increases in the CD11b hi / CD103 − and CD11b + /CD103 + populations in day-fed mice ( Fig. 3C). Overall, these data suggest that the gut circadian machinery influences the cellular composition of the mesenteric lymph nodes, in particular the proportions of DC subtypes.

Chronic infection influences body temperature and activity rhythms, but independently of time of infection. Parasitic infection can lead to widespread systemic disruption of circadian-regulated
physiology 39,40 . To assess whether infection with T. muris leads to a widespread systemic perturbation of circadian organisation and behaviour we used whole body telemetry. Mice display strong diurnal oscillations in both body temperature and activity, which continued following infection (Fig. 4A,B). Comparison of mice infected at ZT0 and ZT12 showed no significant difference in daytime body temperature or activity (Fig. S4). However, a significant increase in the mean amplitude of body temperature was recorded after infection that was independent of time-of-day of infection (Fig. 4C). Infected mice showed a significant reduction in overall activity, likely a sickness response (Fig. 4D, Fig. S4). Further, mice in light-dark cycles typically exhibit a short period of "anticipatory behaviour" to dark-onset with an increase in activity in late light phase, but by late stages of infection this did not occur (Fig. 4B). These data indicate that although infection slightly altered the phasing of activity onsets, robust rhythmic behaviour was maintained and was not significantly modified by infection.
DCs drive circadian responses. We next addressed how the DC circadian clock may be involved, by deleting the core circadian clock gene Bmal1. Mice homozygous for a floxed Bmal1 allele (Bmal1 fl/fl ) were bred with mice expressing Cre-recombinase driven by a CD11c transgene, which is expressed in DCs (CD11c-Bmal1 −/− ). All mice were maintained on a background strain carrying the PER2::Luciferase circadian reporter construct 41 . Cultured bone marrow derived DCs from Bmal1 fl/fl mice exhibited robust PER2 bioluminescent circadian oscillations (Fig. 5A). Strikingly, DCs derived from CD11c-Bmal1 −/− mice showed significantly dampened oscillations. Next, we infected CD11c-Bmal1 −/− and "floxed" (Bmal fl/fl ) littermates with T. muris at either ZT0 or ZT12. As predicted from earlier studies, ZT12-infected Bmal1 fl/fl mice harboured significantly more worms at Day 21 post-infection compared to ZT0 (Fig. 5B). In contrast, worm burdens of ZT0 and ZT12-infected, CD11c-Bmal1 −/− mice were not significantly different, suggesting a loss of circadian control over expulsion kinetics (Fig. 5B). Levels of the T H 1 marker IgG2c were low in all groups of mice at Day 21 post-infection and showed no significant differences between groups (Fig. 5C). However, in keeping with the increased worm burden in ZT12-infected Bmal1 fl/fl mice the signature T H 2 marker IgG1 was significantly reduced only in Bmal1 fl/fl mice infected at ZT12, all other groups exhibited IgG1 levels typically associated with a more efficient worm expulsion (Fig. 5D). Thus, these data suggest that Bmal1 plays a crucial role in the regulation of resistance to T. muris infection, and that loss of Bmal1 in DCs results in loss of the time-of-day variation in worm expulsion, favouring development of a protective immune response across the day.
The loss of bmal1 −/− in DCs reduces expression of T H 1 promoting cytokines. To investigate the role of Bmal1 in DCs, gene expression profiles were compared using bone marrow derived DCs as a surrogate for the tissue DC. Bone marrow derived DCs from Bmal1 fl/fl and CD11c-Bmal1 −/− mice were isolated and cultured to high purity (>94%; Fig. 6A). Following synchronisation with temperature cycles 42 cells were treated with T. muris ES antigen or vehicle at either the peak or trough of PER2::luc bioluminescence activity in Bmal fl/fl PER2::luc cells. Identical time-points were used for the non-cycling CD11c-Bmal1 −/− PER2::luc derived DCs which exhibited markedly dampened bioluminescence rhythms, as shown previously (Fig. 5A). Cells were collected for RNA-SEQ analysis 3 h post antigen challenge.
In unchallenged (naïve) Bmal fl/fl DCs 70 genes were significantly different between peak and trough. Whereas in Bmal1 −/− DCs only 3 genes exhibited significant differences between peak and trough (Fig. 6B). In ES-stimulated DCs we saw an increase in the number of genes differing at peak versus trough in the Bmal fl/fl and bmal1 −/− DCs. However, the number of genes differing in the Bmal fl/fl DC was considerably higher (Fig. 6C). Following ES-stimulation we saw differences between peak and trough in Bmal1 fl / fl DCs in the expression of genes associated with T H 1 polarisation (Table S2). This included il-12b and Tnfsf15 which were significantly down-regulated at peak (Table S2). Further, pathway analyses of the differentially expressed genes in ES stimulated Bmal1 fl/fl DCs at peak versus trough revealed major differences in pro-inflammatory and T H 1 associated  Table S1. pathways. These pathways include the IL-12 and IL-27-mediated signalling pathways, which were down-regulated at peak (Fig. 6D). In addition, changes in metabolic pathways were evident, with increases in, for example, amino acid metabolism and oxidative phosphorylation pathways in Bmal fl/fl DC at peak. Using String analysis 43 we assessed the interconnectivity of genes that differed at peak versus trough in Bmal fl/fl DC (Fig. S5A). Within this gene-set significantly more genes were seen to be interacting than would be expected by chance (P < 10 −14 ).
There were few differences between Bmal1 fl/fl and Bmal1 −/− DCs at the peak PER2 bioluminescence following antigen challenge (Fig. 6F). In marked contrast, comparison of Bmal1 fl/fl and Bmal1 −/− DCs at the trough of PER2 bioluminescence revealed differences (Fig. 6F) with significant down-regulation of T H 1 genes in the Bmal1 −/− DCs (Table S3). Both the IL-12 and IL-27 pathways were also down-regulated in cells lacking Bmal1 (Fig. 6E). In addition, loss of Bmal1 from DCs was associated with an up-regulation in metabolic pathways including amino acid metabolism and the TCA cycle. We also assessed the interconnectivity of genes that differed between genotype from cells sampled at the trough of the PER2::Luc cycle by String analysis (Fig. S5A). In this analysis, Bmal1-regulated genes emerged as a tightly connected cluster (P < 10 −14 ).
Collectively these data from cultured circadian-synchronised DCs stimulated with parasite antigen reveal that T H 1-associated genes are circadian regulated. Cytokine and signalling pathways associated with the development of T H 1 immune responses were down-regulated in the absence of Bmal1. Thus, the clock regulator BMAL1 may be critical for the establishment of DC-derived T H 1 promoting cytokines (Fig. 6H).

Discussion
The circadian clock is recognised as a key regulator of immune cell function, with molecular clock control of inflammation, sepsis and immunity to bacterial and protozoan parasite infection described 15  However, the role of the circadian clock in regulation of immunological responses to large multi-cellular parasites, that are eliminated by T H 2 immune responses, remain unexplored. To address this question, we took advantage of a gastro-intestinal nematode parasite infection model where expulsion of the pathogen is exquisitely dependent on the T H 1/T H 2 cell balance, thus enabling changes in this balance to be evidenced by altered worm expulsion kinetics and the balance of parasite-specific IgG1/IgG2c.
We now show that time of infection is a key component in tuning the efficiency of worm expulsion. Our data show that mice infected in the early light-phase were more efficient in expelling their worm burden than mice infected at the onset of the dark-phase, with cytokine signatures (increased IL-13, reduced IFN-γ), and parasite-specific antibody profiles (increased IgG1; reduced IgG2c) consistent with an immune response polarised towards T H 2. These effects were not immediate, but become manifest over several weeks, and were not associated with the altered worm establishment. Food reversal studies also strongly implicated an important role for metabolic cues during the initial stages of infection. Thus inversion of the local gut clock, by restricting feeding to the daytime, impaired parasitic immune responses. Interestingly, unlike time-of-day of infection, food reversal impaired worm expulsion in the presence of T H 2 immune responses. This was accompanied by increases in CD11b hi /CD103 − and CD11b + /CD103 + cDC with CD11b + cDC thought to be the main drivers of T H 2 immune responses in vivo 45 . One possible implication of this is that reversed-feeding leads to a state of internal de-synchrony between central and peripheral signals 46 . Thus, central signals (T H 2) defining efficient worm expulsion after a ZT0 infection may be unable to coordinate the innate T H 2-controlled effector mechanisms in intestinal tissue subjected to a reversed feeding regime. To our knowledge, this represents the first demonstration that cellular immune responses underlying worm expulsion are sensitive, either directly or indirectly, to feeding-associated metabolic cues. Interestingly, in keeping, tissue innate lymphoid cells (ILC) 2s have previously been shown to influence tissue eosinophilia via a mechanism sensitive to nutrient intake 47 . The alterations in the relative percentages of DC subsets and T-cells in the MLN observed when feeding was restricted to day-time are in keeping with the recent reports 29,30 showing strong diurnal effects on leukocyte migration and concurrent  accumulation of T cells at night compared to day under steady state conditions. Our data from RFR studies, suggests, at least for the mesenteric lymph nodes (MLN), metabolic cues may contribute to this effect. We currently do not know if the RFR regime has ablated oscillations or shifted oscillations and this warrants further investigation. Further, we suggest that populations of cDC within the MLN may also be under circadian control.
Our data identify the DC as a key cell type underpinning the BMAL1 regulated resistance to T. muris in vivo. Thus, mice with Bmal1-deficient DCs no longer showed a time-of-day dependent effect on worm expulsion, with both ZT0 and ZT12-infected mice harbouring equivalent worm burdens to ZT0-infected Bmal fl/fl littermate controls. Although our current studies cannot definitively show that the circadian clock regulates resistance to T. muris infection in vivo, we can conclude that loss of BMAL1 in DCs results in loss of the time-of-day variation in worm expulsion. To begin to explore further how BMAL1 in DCs regulates the time-of-day dependency of worm expulsion we used circadian-synchronised purified bone marrow derived DCs stimulated with T. muris antigen. These data revealed BMAL1 as an important component in determining the ability of the DC to promote polarised T cell responses with a down-regulation in T H 1-promoting cytokines and pathways in BMAL1-deficient cells. Furthermore, we also show that these differences are both DC-intrinsic and circadian-regulated. This supports a mechanism that involves dynamic changes over the day in the promotion of T H 1-promoting cytokines (e.g. IL-12) and T H 1-pathways in non-targeted (Bmal1 fl/fl ) DCs, which are circadian-regulated. Interestingly, two of the major signalling pathways down-regulated in BMAL1 deficient DCs were the IL-12 and IL-27-mediated signalling pathways, both known to be critical in enabling T H 1-mediated immune responses, and failure to expel T. muris 6,48 . Further studies are warranted to translate our in vitro driven hypothesis to in vivo, exploring the early post infection cellular events underpinning the time of day effects on worm expulsion which emerge three weeks post infection.
Our study adds the DC to the repertoire of circadian-regulated immune cells in which components of the core molecular clock-work have been targeted, with impact on either innate and/or adaptive responses 18,30,49,50 . We selected Bmal1 for genetic targeting, as uniquely loss of this transcription factor leads to suppression of rhythmic activity of the core mammalian clockwork 51,52 . Given that Bmal1 is the critical regulator of the other clock genes which comprise the molecular clock, it remains to be defined whether the effects of Bmal1 deficiency are through effects on downstream clock proteins, such as REV-ERB or CRYPTOCHROME, or are directly BMAL1-regulated.
In summary, we show that expulsion of T. muris is influenced by time-of-day of infection and involves entrainment by metabolic feeding cues. BMAL1 and clock-control of DC function now emerge as important components in the efficient expulsion of helminth parasites. Infection. The parasite was maintained as previously described 38 . 200 eggs were given per oral gavage to each mouse. For worm counts, caecum and colon were harvested, opened and the gut mucosa scraped to remove epithelia prior to screening for T. muris worms. Immunohistochemistry. Immunohistochemistry was performed on frozen colon sections. CD11c (clone N418) antibody (eBioscience, Hatfield, UK) was used to identify dendritic cells, F4/80 (clone Cl-A3-1, AbD Serotec) antibodies were used to identify macrophages. Slides were counterstained using Mayer´s haematoxylin (Sigma Aldrich Company Ltd, Dorset, UK). IgG ELISA. Serum was assayed for parasite specific IgG1 and IgG2c. 96 well plates were coated with 5 μg/ml T. muris E/S antigen overnight, incubated with serum (2 fold dilutions, 1:20-1:2560). Parasite specific antibody was measured using biotinylated IgG1 or IgG2c (BD Biosciences). The plates were read at 405 nm, with reference of 490 nm.

Culture of bone marrow derived dendritic cells. Bone marrow cells were seeded on
SCientifiC RepoRTs | (2018) 8:3782 | DOI:10.1038/s41598-018-22021-5 IgE ELISA. Serum was assayed for total IgE antibody production. 96 well plates were coated with purified anti-mouse IgE (2 ug/ml, Biolegend, Clone: RME-1) in 0.05 M carbonate/bicarbonate buffer and incubated overnight at 4 °C. Following coating, plates were washed in PBS-Tween and non-specific binding blocked with 3% BSA (Sigma-Aldrich) in PBS for 1 hour at room temperature. Plates were washed and diluted serum (1:10) added to the plate and incubated for 2 h at 37 °C. After washing HRP conjugated goat anti-mouse IgE (1 ug/ml; BioRad) was added to the plates for 1 h. Finally, plates were washed and developed with TMB substrate kit (BD Biosciences, Oxford, UK) according to the manufacturer's instructions. The reaction was stopped using 0.18 M H 2 SO 4 , when sufficient colour had developed. The plates were read by a MRX II microplate reader (DynexTechnologies, VA, USA) at 450 nm, with reference of 570 nm subtracted. qPCR. RNA was extracted after homogenising gut tissue in Trizol using a FastPrep machine with Lysing matrix D (MP Biomedicals). After genomic DNA elimination, 2 μg RNA was converted to cDNA (RNA to cDNA, Applied biosystems). qPCR was performed on a StepOne plus machine using Taqman probes and primers and qPCR MasterMix (Eurogentec). Beta actin was used as a housekeeping gene.

RNA-Seq analysis.
Strand-specific RNA-Seq libraries were prepared using the Illumina workflow.
Paired-end reads were generated and 40 to 85 M of total reads obtained from each sample. The fastq files generated by Illumina HiSeq4000 platform were analysed with FastQC and any low quality reads and contaminated barcodes were trimmed with Trimmomatic 55 . All libraries were aligned to GRCm38.p2 (mm10) assembly of mouse genome using Tophat-2.1.0 56 and only matches with the best score were reported for each read. The mapped reads were counted by genes with HTSeq 57 against gencode.vM2.annotation.gtf. Differentially expressed (DE) genes were identified by comparing between the treatment groups with DESeq2 58 . The functional analysis was carried out with a R package topGO 59 and RontoTools.

Statistics.
Where statistics are quoted, Mann-Whitney U test was used for non-parametric data comparing two groups. Kruskal-Wallis test, with Dunn's multiple comparison post hoc was used to compare three or more groups. Parametric data were assessed using unpaired T tests for pairwise comparisons or one-way ANOVA and post-hoc Tukey. A p-value of < 0.05 was classed as significant, *P < 0.05, **P < 0.01, ***P < 0.005. Statistical analysis was carried out using GraphPad Prism for windows, version 6. Data availability. The datasets generated during the current study are available from the corresponding authors on reasonable request.