The evolutionary ecology of circadian rhythms in infection

Biological rhythms coordinate organisms’ activities with daily rhythms in the environment. For parasites, this includes rhythms in both the external abiotic environment and the within-host biotic environment. Hosts exhibit rhythms in behaviours and physiologies, including immune responses, and parasites exhibit rhythms in traits underpinning virulence and transmission. Yet, the evolutionary and ecological drivers of rhythms in traits underpinning host defence and parasite offence are largely unknown. Here, we explore how hosts use rhythms to defend against infection, why parasites have rhythms, and whether parasites can manipulate host clocks to their own ends. Harnessing host rhythms or disrupting parasite rhythms could be exploited for clinical benefit; we propose an interdisciplinary effort to drive this emerging field forward.

Circadian rhythms have long been taken for granted by science. Indeed, the first observation of a clock-controlled behaviour (leaf opening and closing in Mimosa pudica) was not recorded until the 18^th century¹. Following the fundamental observation that organisms can adaptively anticipate daily rhythms in their environment, the field of “chronobiology” took off in the mid-20^th century with a focus on evolutionary and ecological questions. However, the advent of genetic tools a few decades later shifted the remit to determining the molecular and genetic workings of circadian clocks. Yet, despite their assumed major impact on fitness, circadian rhythms remain overlooked in evolutionary ecology^2–4. Here, we propose that the integration of chronobiology and evolutionary ecology return to its roots to tackle a topic of growing and applied interest; the role of rhythms in host-parasite interactions. Note that we use the term “parasite” to collectively refer to all agents of infection (e.g. single-celled and multicellular eukaryotes, bacteria, viruses).

One of the most fundamental ecological interactions is that between hosts and parasites. Research from diverse taxa (plants, mammals, and insects) reveals that host clocks drive daily rhythms in immune defences, disease severity and spread^5,6. Parasites display daily rhythms in traits underpinning within-host survival and between-host transmission^7,8. Rhythms in parasite activities and in host responses to infection could provide an advantage to parasites, hosts, both, or neither. To what extent parasites and hosts are in control of their own and/or each other’s rhythms is also poorly understood.

Understanding the evolution (and possibly, coevolution) of rhythms may enable vaccines and drugs to take advantage of rhythmic vulnerabilities in parasites or harness host rhythms to improve efficacy and reduce drug toxicity. For such interventions to be robust to parasite evolution, understanding how host-parasite interactions shape rhythms in hosts and parasites is necessary⁷. Key questions include how rhythms in diverse host traits contribute to defence, how parasites cope with exposure to their host’s rhythms, and whether hosts and parasites can manipulate each other’s rhythms for their own benefit. We discuss these three scenarios, identify systems to explore them, and offer ways in which this knowledge can be exploited to improve health. An evolutionary ecologist’s introduction to chronobiology is provided in Boxes 1 and 2.

Box 1. What are circadian rhythms?

Biological rhythms are deemed to be controlled by circadian clocks if they meet several criteria⁷¹. First, their duration (period) must be approximately 24 hours. Second, they must persist (free-run) in conditions without time-of-day cues, which is usually assessed by observation in constant light or dark. Third, the phase of the oscillator or outputs are set (entrained) by a time-of-day cue (Zeitgeber) which is usually light. Fourth, unlike the rate of many chemical reactions, the speed of a circadian clock varies little over a biologically realistic range of environmental temperatures (temperature compensation). Together, these criteria allow organisms to fulfil a key feature of circadian rhythms: anticipatory, rather than reactionary, behaviour. For instance, plants ready photosynthetic machinery in anticipation of sunlight^72,73 and animals exhibit food-anticipatory activity (e.g. increases in core temperature, activity, serum corticosterone, and duodenal disaccharides) prior to foraging⁷⁴. The workings of circadian clocks are sufficiently flexible to allow organisms to cope with gradual changes in photoperiod across seasons, but not flexible enough to instantly cope with changes in time zones (which is why travellers experience jet lag).

The mammalian circadian system is composed of the “central” clock in the brain (suprachiasmatic nucleus; SCN) and “peripheral clocks” in other organs and tissues (A). Clocks in nucleated cells are run by transcription-translation feedback loops (TTFL). For example, in animals the proteins CLOCK and BMAL1 act as activators and members of the PER and CRY families are repressors⁷⁵ (B). Retinal photoreceptors receive light cues which are carried through the hypothalamic optic tract and transmitted to the SCN, resulting in its synchronization/entrainment (C). Clocks in organs and tissues (peripheral clocks) can be entrained by feeding rhythms, and in taxa other than mammals, exercise, social cues, and abiotic rhythms in temperature and humidity may entrain clocks (D). Rhythms are often characterised by their period, amplitude, and markers for phase (E; grey bars illustrate night time for a rhythmic trait measured over 48 hours). They are described in relation to the time since the Zeitgeber (ZT) occurred (e.g. ZT6 refers to 6 hours after dawn) which usually differs from the actual time-of-day (Circadian Time; CT).

Box 2. Why have circadian rhythms evolved?

Circadian clocks appear so advantageous that nearly all eukaryotes have a circadian system in most cells⁷⁶. Circadian clocks may confer two kinds of fitness benefit: coordinating behaviours with rhythms in the external environment (extrinsic adaptive value), and temporally compartmentalising incompatible processes (intrinsic adaptive value)². For instance, intrinsic benefits are conferred when cell division in yeast is temporally constrained to the reductive phase of metabolism, minimising rates of genetic mutation⁷⁷. However, most studies of the fitness consequences of circadian rhythms have focussed on the benefits of synchronizing activities with rhythms in the abiotic environment: matching the period of day-night rhythms enables cyanobacteria to outcompete strains whose clocks run faster or slower⁷⁸ and enhances the survival of Arabidopsis⁷³. Rhythms in the biotic environment² matter too. For example, the sea urchin Centrostephanus coronatus avoids predatory sheephead wrasse (Pimelometopon pulchrum) by foraging at night and retreating to shelter prior to the onset of wrasse activity⁷⁹.

Despite the diversity of extrinsic rhythms that could select for the scheduling of diverse processes, there are surprisingly few demonstrations that circadian clocks actually affect fitness. For example, fitness is greater in wild-type mice than mutant mice with shortened periods⁸⁰, flies with clock mutations die more rapidly than wild types after infection with bacteria^81,82, and circadian knockout plants flower later and are less viable than wild-type plants³. However, depending on ecological context, rigidly scheduling activities according to day and night is not always the best strategy. For example, nocturnal mice boost energy efficiency by switching to diurnality when challenged with cold and hunger⁸³. Nursing honeybees, that remain in the hive are arrhythmic, because round-the-clock care is necessary for larvae; and, if needed, diurnal foraging bees can revert to arrhythmic nursing behaviour⁸⁴. Shorebirds also display considerable plasticity in activity rhythms during breeding, likely explained by predator avoidance strategies⁸⁵.

The above examples illustrate the gains to be made from integrating chronobiology with evolutionary ecology in general⁴. We propose that such an approach offers a novel advance to the study of host-parasite interactions and coevolution. Coupling the well-developed conceptual frameworks for unravelling how circadian oscillators operate, and probing the costs and benefits of phenotypically plastic traits that are relevant to infection, will explain why rhythms in immune defences and parasite traits occur.

Rhythms in host defence

The most patent defence against infection is the immune response, and a wealth of evidence reveals that circadian clocks play a role in orchestrating immune defences⁵. Circadian clock genes are expressed in many types of immune cell, and the immune and circadian systems are connected in multiple ways^9,10. For instance, the clock gene Bmal1 mediates the balance between pro- and anti-inflammatory responses¹¹. Rhythmic production of the pro-inflammatory cytokines TNF-α and IL-6 by macrophages is clock controlled¹², and mobilization of inflammatory monocytes is also regulated by the clock¹⁰. This phenomenon, termed “anticipatory inflammation”, appears uncoupled to metabolic rhythms and may defend against incoming parasites¹³. Similarly, in humans, proinflammatory cytokines peak in circulation during the day (active phase)¹⁴, whereas hematopoietic stem and progenitor cells, and most mature leukocytes, peak at night^14,15. In nocturnal mammals, an inverse rhythm is often observed, with innate defences peaking at night (active phase) and repair mechanisms peaking during the day (resting phase)⁹.

Observations of immune rhythms have given rise to the notion that organisms invest in defence during the active phase when parasite encounter is assumed most likely, and repair during the resting phase¹⁶. Temporal segregation of immune responses may thus solve problems caused by having immune defences continually tuned to maximal (e.g. collateral damage via immunopathology¹⁷). Also, energetic demands imposed by activity and metabolism may trade-off against immune defence¹⁸. Intuitively, “defence only during the active phase” suggests the host is achieving the most “bang for the buck” by ensuring activities that are energetically costly, or likely to cause collateral damage, are only performed when most useful. However, this intuition may be naïve. First, it ignores the potential for constraints imposed by the need to temporally couple (or de-couple) certain immune rhythms with other internal rhythms⁷. This includes separating the timing of metabolism from defensive actions within immune cells themselves^5,16. Second, it assumes that a parasite encounter is rhythmic and predictably occurs in the active phase. This is clearly the case for food-borne parasites, but ingestion is not the only route into a host. Rather, the immune system functions within a broad set of energetic demands in which parasite defence is just one of many requirements. For example, rhythmic stomatal opening for gas exchange during the day is a well-used route into plants by bacterial pathogens¹⁹. Consequently, Arabidopsis is better able to detect and defend against parasites in the morning than evening^20,21. Given the wealth and diversity of data (illustrated in Table 1), meta-analyses are needed to test whether the timing (phase) of rhythms in immune effectors relates to nocturnal vs diurnal lifestyles and whether they function in front-line or secondary defences, or healing.

Table 1. Impact of immune challenge during the rest and active phases of hosts.

A selection of studies identified as time-of-day immune challenges from PubMed searches for ““time of day” plus “immune and infection” and ““circadian rhythm” plus “immune and infection”. Articles were included if the study involved a time-of-day immune challenge; those without a time-of-day immune challenge were not included in the table. Time-of-day (ToD) is given as hours since lights on (ZT) for organisms in entrainment conditions, and as subjective day/night for those in constant light or dark conditions (i.e. corresponding to the light or dark portion of the cycle before experiencing constant conditions). Unless otherwise stated, entrainment conditions are 12 hour light:dark. Outcomes of challenge in the rest phase (daytime for nocturnal organisms, nighttime for diurnal organisms) are compared to challenge in the active phase in terms of virulence metrics and immune effectors measured.

Host spp.	Challenge	ToD	Outcome in rest versus active phase	Ref
Mus musculus – house mouse (nocturnal)	Salmonella typhmurium	ZT4/16	Greater inflammation and bacterial load when infected in the rest phase	22
	Leishmania major	Subjective day/night	Lower parasite burden and lower severity when infected in the rest phase	23
	Lipopolysaccharide (LPS) endotoxin	Subjective day/night	Lower concentrations of cytokines when infected in the rest phase	91
		ZT11/19	Higher mortality when challenged in the rest phase	92
		Subjective day/night	Greater inflammatory responses and lower bacterial burden when challenged/infected in the rest phase	93
	Streptococcus pneumoniae	ZT0/12
	Murid Herpesvirus 4	ZT0/10	Greater viral replication when infected in the rest phase	60
	Helicobacter pylori	ZT1/7/13	Lower lymphocyte numbers when infected in the rest phase	94
	Vesicular stomatitis virus	ZT0/12	Higher mortality when infected in the rest phase	95
Drosophila melanogaster – fruit fly (diurnal)	Pseudomonas aeruginosa	ZT1/5/9/13/17/21/1	Lowest mortality when infected in the rest phase (especially ZT21)	82
		Subjective day/night	Lowest bacterial burden when infected in the rest phase
	Streptococcus pneumoniae	ZT7/19	Slowest rate of mortality when infected in the rest phase	81
	Escherichia coli	ZT0/6/12/18	Infection at all ZT induces sleep the morning after infection and sleep was more prolonged after infection in the rest phase	96
Anopheles stephensi - Asian malaria mosquito (nocturnal)	Escherichia coli	Morning/evening	Lower bacterial growth and lower mortality when infected in the rest phase	97
Arabidopsis thaliana – thale cress (diurnal)	Pseudomonas syringae	ZT0/4/10/16	Immune defences are highest when inoculation occurs early in the active phase Note photoperiod is 9 hours light: 15 hours dark	98
	Botrytis cinerea	Dawn/dusk	Larger lesions when inoculated in the rest phase	49
		ZT0/3/6/9/12/15/18/21/24	Greater susceptibility when inoculated in the rest phase	21
	Pseudomonas syringae	Subjective day/night	Lower infiltration of bacteria when infected in the rest phase	99
		Subjective morning/evening	Greater suppression of bacterial growth at the start of the rest phase when spray-inoculated, and greater suppression of bacterial growth at the start of the active phase when syringe-infiltrated	20
	Hyaloperonospora arabidopsidis	Dawn/dusk	Highest percentage of leaves with sporangiophores when infected in the start of the rest phase	100
Danio rerio zebrafish (diurnal)	Salmonella typhimurium	ZT4/16	Lower survival when infected in the rest phase	101
Oreochromis niloticus – Nile tilapia (mostly diurnal)	LPS	ZT3/15	Greater humoral immune response when infected in the rest phase	102
Phodopus sungorus – Siberian hamster (nocturnal)	LPS	ZT1/16	Shorter febrile response and more persistent locomotor activity when infected in the rest phase. Note, photoperiod is 16 hours light: 8 hours dark	103

Infection in the active vs resting phase for diverse hosts (flies, plants, mammals) dramatically affects disease severity and mortality rates (Table 1), suggesting that the phase of immune rhythms upon infection matters. Most studies performed in plants (Table 1) point towards infection during the active phase resulting in greater resistance to infection and less damage to the plant. But the degree to which immune rhythms result in time-of-day differences in parasite control can be counter-intuitive. For example, mice mount higher clock-controlled proinflammatory responses against Salmonella enterica Typhimurium when challenged in their rest phase, but bacterial load is also higher and hosts have worse symptoms²². Furthermore, Leishmania parasites infect host neutrophils and macrophages, and the clock-controlled secretion of chemoattractants by these immune cells facilitates their infection, making parasite invasion more successful at night when immune activity is highest²³. Thus, whether immune rhythms are sufficient to entirely explain divergent outcomes of time-of-day of infection is unclear (Table 1). Studies that separate the effects of immune rhythms on preventing infection from their role in dealing with ongoing infection will reveal the extent to which immune rhythms are beneficial and when they should be overruled to deal with a major threat. Additionally, most time-of-day immune challenges have used either bacteria or chemicals, raising the question of whether a more diverse array of challenges are needed to establish general patterns.

That host circadian clocks impact on infection via traits other than immune responses has been largely overlooked. Rhythmicity in host activity may determine when hosts provide the best resources to their parasites and offer the most opportunities for onwards transmission^24–26. For example, a recent study of the intestinal helminth Trichuris muris demonstrates the role of host rhythms in foraging. Mice infected in the morning (resting phase) expel worms sooner and have a stronger T-helper 2 response than dusk-infected (active phase) mice, and this effect is reversed when mice are fed only in the day, in an immune-independent manner²⁷. Host feeding rhythms are relevant to gut microbiota, and a two-way feedback between host and microbe rhythms has been proposed²⁸. Daily rhythms in host reproductive behaviours may make hosts vulnerable to infection. For example, the crepuscular and nocturnal singing activity of the cricket Teleogryllus oceanicus allows the acoustically-orienting parasitoid fly Ormia ochracea to locate hosts, but the flies are best able to hunt when darkness is incomplete²⁹. A rhythmically expressed reproductive behaviour (singing) got the host into this mess, and it appears that natural selection has found two solutions (see Box 3).

Box 3. Case studies illustrating the role of circadian rhythms in parasite offence, host defence, and host manipulation.

Host-parasite system

Teleogryllus oceanicus (Pacific field cricket) & Ormia ochracea (parasitoid fly)

What we know

O. ochracea deposit larvae which burrow into the host and emerge 7-10 days later, resulting in host death. A flatwing morph that is physically incapable of calling has evolved to evade the risk of parasitism by acting as a silent, satellite male²⁴.

A more nuanced form of parasite evasion?

In addition to the flatwing morph, natural selection may have found another solution. Some males condense singing activity to the darkest part of the night²⁹ which may hamper the fly’s ability to use visual cues to home in on hosts. Parasite evasion (via a flatwing phenotype or phase-shifted calling) trades off against attracting females, potentially constraining selection on these strategies. Moreover, multiple activities need to be coordinated for successful reproduction (e.g. locomotion, foraging, spermatophore production). Given that many of these traits are clock-controlled, could altering the timing outputs of the clock be a streamlined way of phase-shifting all related activities and minimizing the costs of parasite evasion? [associated image = cricket_fly.png] Photo credit: Norman Lee

Host-parasite system

Carpenter ants & Ophiocordyceps spp. and Pandora spp. (fungi)

What we know

O. unilateralis s.l. induces workers of its carpenter ant host, ordinarily active during the night-time, to wander out of the ant nest during the day-time. Hosts then summit vegetation and adopt a mandibular death-grip in elevated positions. This manipulated behaviour is highly time-of-day and species-specific and occurs within a 3-hour window at dawn or in the mid-late morning, depending on the species^68,86. Clinging to vegetation, the ant dies whilst the fungus completes its life cycle by growing a spore-producing stalk out of the dorsal region of the ant’s thorax⁸⁶.

A case for coevolution and ecosystem specificity?

The jigsaw puzzle of how the fungus controls the ant is still being pieced together. Clocks may play a central role because infection alters the expression of host clock homologues period and cycle⁶⁸. Host manipulation also appears to involve altering host chemosensory abilities, potentially via rhythmic secretion of enterotoxins⁸⁷, all achieved from the fungus’s primary location in muscle tissues⁸⁸. [associated image = ant_fungi.png] Photo credit: Miles Zhang

Host-parasite system

Mammals & Plasmodium spp. (malaria parasites)

What we know

Malaria parasites synchronously burst from the host’s blood cells every 24, 48, or 72 hours depending on the parasite species⁸⁹. When out of synch with the host’s circadian rhythms, parasites incur an approximately 50 percent reduction in the densities of both asexual stages (necessary for in-host survival), and sexual stages (responsible for transmission)⁹⁰ before they become rescheduled to be in synch with host feeding rhythms^44,45.

Three worlds collide: a complex system of interactions?

Why aligning the phase of parasite rhythms with the host’s rhythms is important remains mysterious, but recent work suggests that parasites are also selected to coordinate with the time-of-day their mosquito vectors are active^54,55 (see Rund et al. 2011 for information on Anopheles circadian rhythms). If differently phased rhythms for asexual replication are required to provide the best matches to host and vector rhythms, parasites face a trade-off between maximizing in-host survival and between-host transmission. Such a tension could be exploited by novel drug treatments to coerce parasites into a loss of fitness. Further, mosquito nets have induced a shift in Anopheles gambiae biting activity, ultimately resulting in a change in host-parasite timing^8,52,53. The epidemiological consequences of this are unknown. [associated image = mosquito_malaria.png] Photo credit: Sinclair Stammers

In addition to immune responses, infected hosts often exhibit adaptive sickness behaviours consisting of endocrine, autonomic, and behavioural changes that perturb circadian rhythms^30,31. For example, wild red colobus monkeys (Procolobus rufomitratus tephrosceles) decrease energetically costly activities, and rest frequently, while shedding whipworm eggs³². Fever, another common sickness behaviour, is sufficiently advantageous to offset the 10-12.5% increase in metabolic rate required for each 1°C increase in temperature³³ and has been conserved throughout more than 600 million years of vertebrate evolution³⁴. Fever enhances an organisms chance of survival by creating a hostile environment for parasites and a more active immune response^34–37. Under normal circumstances, the so-called central (SCN) clock controls body temperature rhythms, but how the SCN and inflammation interact to control temperature is unknown. Though many behaviours altered during infection are clock-controlled during health, the extent to which organisms become too sick to maintain normal behaviour or adaptively disrupt their rhythms is unclear. Additionally, clock-control could facilitate recovery of rhythms during the return to health.

Viewing the host as a collection of traits connected by the circadian system has the potential to uncover novel strategies to resist infection and reveal the circumstance in which immune rhythms reflect constraints or adaptations. Indeed, rhythmic metabolism of xenobiotic substances (e.g. drugs and vaccines) influences efficacy and toxicity in a time-of-day dependent manner³⁸. For example, halothane (a commonly used anaesthetic) administered to mice in the daytime results in low mortality (5%), but mortality increases (76%) if administered at night³⁹ and half of the best-selling drugs in the USA for humans target the products of genes that are rhythmically expressed (in mice)⁴⁰. A better understanding of host rhythms could be harnessed to make drugs and vaccines more effective, as well as mitigating the negative effects of modern lifestyles that involve shift work and jet lag. However, for such interventions to be sustainable in the face of parasite evolution, understanding the ecology of rhythms from the perspective of parasites is also required.

Rhythms in parasite offence

Scheduling activities to take advantage of daily rhythms in transmission opportunities could be a general explanation for rhythms in parasites. The most well-known example concerns the transmission forms (microfilariae) of different species of filarial worms. They move from the host’s organs to the capillaries during the day or night, depending on whether they are transmitted by day- or night-biting insect vectors⁴¹. In addition to the activity patterns of vectors, rhythmic interactions with hosts also matter. For example, the larval stage of the blood fluke Schistosoma japonicum emerge from their invertebrate host to seek a mammalian host at different times of day. Flukes emerge in the afternoon when the preferred host is nocturnal or in the morning if seeking a diurnal host⁴². Parasites that have free-living stages are also subject to rhythms in the abiotic environments. The coccidian parasite Isospora sheds from its host in the late afternoon to minimise UV exposure and desiccation risk whilst undergoing a developmental transition necessary to infect new hosts⁴³. However, key questions remain about the adaptive nature of these rhythms. For example, why aren’t microfilariae located in the peripheral capillaries all day long? Is a cost associated with this location, which is only worth paying at times of day when vectors are active?

In contrast to the role of parasite rhythms in transmission, their role in within-host survival has received less attention. Many host rhythms (in addition to immune rhythms) present opportunities and constraints for parasites. Trypanosoma brucei (which cause sleeping sickness) display circadian clock-driven rhythms in the expression of metabolic genes⁸. These rhythms correlate with time-of-day sensitivity to oxidative damage, thereby suggesting the need to cope with redox challenges caused by rhythmic digestion of food by hosts. In contrast, rhythms in the development of asexually replicating malaria parasites capitalise on daily variation in the nutritional content of blood caused by host immune responses and feeding patterns^44,45. Whether malaria parasites cannot complete their developmental cycle until the host makes nutrients available, and/or use nutrients rhythms as a time-of-day cue to set the pace of their development, is unknown⁴⁶ (see Box 3).

Clocks in parasites or hosts could have fitness consequences for one or both parties, or neither. Fitness consequences for both hosts and parasites suggests that clocks could coevolve. Clock coevolution is suspected for the plant-pollinator system Petunia axillaris and Manduca sexta⁴⁷, in which nocturnal scent emission by P. axillaris coincides with foraging activity in the hawkmoth M. sexta. Both traits are clock-controlled, and appear so well synchronized that, even in the absence of floral scent emission, M. sexta exhibits a burst in foraging activity at the same time that floral scent emission is expected to be greatest. However, foraging behaviour also remains sensitive to the environment, as evidenced by absence of activity when the moth is subjected to light at night. If rhythms in different organisms do coevolve, then they should use the same Zeitgeber, but how robust should their timing systems be to fluctuations in the environment? If the rhythm of one party is more readily disrupted (masked) by environmental change, or faster at tracking seasonal changes in photoperiod, then the relationship may be disrupted to the gain of hosts or parasites. Exploring the degree and consequences of plasticity in rhythms is pertinent because climate change is interfering with the ability of interacting species to synchronise⁴⁸.

The situation is further complicated when interactions between both host and parasite clocks shape disease trajectories. For example, in a plant-fungus system (Arabidopsis thaliana and Botrytis cinerea, respectively), when both parties are in the same photoperiod schedule, primary plant defences peak in the morning, and the fungus produces the biggest lesions when inoculated at dusk⁴⁹. The authors were able to separate the contributions to pathogenicity by host and parasite clocks using reverse lighting schedules for fungus and plants: fungus at dusk produced more severe infections than fungus at dawn, regardless of time-of-day for recipient plants⁴⁹. Furthermore, this suggests B. cinerea anticipates and exploits weaknesses in plant defence at dusk rather than attempting to overwhelm dawn defences (see section “Rhythms in host defence”). Separately assigning the contributions of rhythms in hosts/vectors and parasites to virulence and transmission is necessary to understand whose genes control which rhythms, and hence how they can be shaped by selection.

If parasite rhythms are adaptive, then disrupting them could reduce disease severity as well as transmission. However, understanding the timing mechanisms of parasite rhythms is necessary to disrupt them⁷. Unravelling how parasite rhythms are controlled is a considerable challenge. Parasites might allow the host to inadvertently schedule their activities for them, in which case the genes encoding parasite timing mechanisms belong to hosts. Alternatively, parasites might keep time using a circadian clock (with the properties described in Box 1), as demonstrated for T. brucei and B. cinerea. Given the diversity in clock genes across taxa, searching genomes for known clock genes often yields “absence of evidence” not “evidence of absence.” Instead, round-the-clock transcriptomics or proteomics, paired with bioinformatics approaches to mine for known core clock-related functional domains and sequence patterns may find candidates. However, simpler time-keeping strategies exist, though they do not necessarily have the advantages of temperature compensation or anticipation. For example, cell division cycles are often controlled by hourglass mechanisms that rely upon threshold concentrations of substances, independently of periodic phenomena⁵⁰. Alternatively, organisms can react directly (via “tracking”) to temporal changes in the environment. Note, this differs from masking, a chronobiological phenomenon in which the expression of a clock-controlled rhythm is suppressed by a change in the environment without having a direct effect on the period or phase of the underlying rhythm⁵¹. A response that directly tracks time-of-day cues may suit parasites with multi-host lifecycles if each host type provides a different time-cue.

Given that rhythms in T. brucei metabolism and plasticity in development during the asexual cycle of Plasmodium spp. enables these parasites to tolerate drugs, there is an urgent need for proximate and ultimate explanations of their rhythms. The T. brucei clock is entrained by temperature cycles, but if other parasites use Zeitgebers to set their clocks, or respond directly to time-of-day cues, that are readily perturbed, it should be possible to reduce parasite fitness by interfering with their rhythms. Further, reports of changes to the biting time of mosquito populations that transmit malaria suggests that insecticide-treated bed nets are imposing selection on vector rhythms^8,52,53. Given that rhythms of parasites and mosquitoes each affect malaria transmission in lab experiments^54,55, what are the likely epidemiological consequences? Recent work suggests that mosquitoes are more susceptible to infection when they feed in the daytime and parasites are more infectious at night⁵⁴. Thus, day-biting could increase the prevalence, but not burden, of malaria in mosquitoes. However, in the longer term, if parasites evolve to invert their rhythm but mosquitoes do not, both prevalence and burden may increase.

Parasite manipulation of host rhythms

Rhythms in host processes offer opportunities that parasites could exploit. Could parasite fitness be increased by coercing hosts into altering their rhythms? Although many striking examples of parasite manipulation of host phenotypes (i.e. changes to host traits that benefit parasites) are known⁵⁶, the notion of “parasite manipulation of host clocks” is largely unexplored⁵⁷. A pre-requisite for parasite manipulation is that a phenotypically plastic host trait is targeted; and circadian clocks are flexible. Because clocks control much of the host’s behaviour and physiology⁵⁸ and clocks throughout a given host involve the same players in the canonical clock (the TTFL), manipulation of the host’s time-keeping may be an efficient way to simultaneously alter many aspects of the within-host environment. Alternatively, parasites interests may be served by bolstering circadian rhythms of their hosts during sickness to ensure they forage and interact with conspecifics, as usual.

As outlined in the section “Rhythms in host defence,” separating the effects of being sick per se from host defence and parasite manipulation is challenging. Recently, a combination of culture and comparison of infection models has revealed that T. brucei alters expression rhythms of clock genes in host mice⁵⁹. Specifically, infected hosts are more active in the resting phase (phase-advanced) because the clock runs faster (shorter period). Effects at organismal, cellular, and molecular levels suggests the behaviour is not just a result of sickness⁵⁹. However, it is not clear how T. brucei achieves this, and whether the parasite benefits from altering host rhythms. One target of circadian disruption by viral parasites is the gene Bmal1, a core clock gene. Herpes and influenza A virus replication and dissemination within the host is enhanced in infections where Bmal1 is knocked out⁶⁰. However, it remains unclear if virus replication is maximised by simply disturbing rhythmicity in host cell cycles or if this is a case of immune manipulation since Bmal1 appears involved in innate host defence⁶⁰. Having observed changes to host clocks, the proceeding step is to decipher the ecological context behind these effects.

The above examples lend proof-of-principle to the idea that parasites can manipulate host clocks and could be a general explanation for examples of host manipulation. Hairworms (Nematomorpha) are a well-known case of temporally linked behavioural manipulation. They infect various arthropods, notably crickets, and cause the host to wander in an erratic manner until a body of water is encountered. The host commits suicide by jumping in water, and the adult hairworm emerges. Infected hosts are found wandering only in the early part of the night⁶¹, and uninfected hosts are rarely motivated to jump into water. Infected crickets differentially express an array of proteins, some of which are linked to visual processes and circadian clocks⁶². Culturing isolated host cells with parasite products and quantifying the expression of clock genes (following Rijo-Ferreira 2018) could illuminate this case of parasite manipulation. For systems without relevant insect cells lines, or cases where manipulation is likely to be tissue/cell type specific, a transcriptomics approach may be useful⁶³. Round the clock expression data can be mined for putative core clock genes and their phase, amplitude and period assessed in control and manipulated hosts. This however, is likely to be extremely challenging for host species whose timekeeping does not rely on a canonical circadian clock.

Another putative case for clock manipulation concerns the New Zealand freshwater snail (Potamopyrgus antipodarum) infected with Microphallus trematodes⁶⁴ (Trematoda: Microphallidae). Uninfected adult snails forage primarily at night on the upper surfaces of rocks in the shallow-water margins of lakes. These snails retreat to under rocks at sunrise, which likely reduces their risk of predation by waterfowl, which are the definitive host for Microphallus. Infected snails, however, show delayed retreating, potentially making them more likely to be consumed²⁵. Crucially, the apparent manipulation only occurs when the parasite is mature. Snails infected with immature (non-transmissible) stages exhibit the same risk-averse retreating behaviour as uninfected snails²⁵. In addition, snails infected with other species of sterilizing trematodes, which are not trophically transmitted, do not exhibit the same risky behaviour as those infected with Microphallus⁶⁵, thereby eliminating the possibility that the Microphallus-induced behavioural change is a simple artefact of parasitic castration. Finally, Microphallus-infected snails spend more time foraging on the top of rocks, even when food was removed whereas uninfected snails retreated to shelter⁶⁵. Taken together, the data suggest that Microphallus induce a change in snail behaviour that increases trophic transmission, potentially via manipulation of clock-controlled activity rhythms.

There are many ways that parasites could interfere with clock-controlled host behaviours. A blunt instrument would be to alter perception/detection of the Zeitgeber that sets the time of the host’s clock, which is usually light. For example, Microphallus could interfere with photoreception to reduce the sensitivity of snails to dawn, causing their clocks to phase delay and forage at higher light intensities than un-manipulated snails. Alternatively, parasites could induce the host to ignore its clock (mask) or alter clock regulation of hormones that relay time-of-day information around the host. For example, baculoviruses appear to perturb the circadian rhythms of their caterpillar hosts by disrupting hormones that control climbing behaviour. In the baculovirus (Lymantria dispar nucleopolyhedrovirus), a single gene inactivates 20-hydroxyecdysone⁶⁶ (a host hormone regulated by a circadian oscillator), motivating the caterpillar to climb high atop their host plants. Here, they liquefy and disseminate the virus to caterpillars below, as well as infecting birds who consume the corpses⁶⁷. Similar to the manipulation of caterpillar hosts, many species of parasitic fungi (Ophiocordyceps spp. and Pandora spp.) alter the daily behavioural rhythm of a variety of ant species^68,69 (See Box 3).

Parsing out whether temporal disruption is a host response or clock manipulation is nearly, if not entirely, impossible without uncovering the mechanism of manipulation. The lack of insight into the mechanisms parasites use to interfere with their hosts has stalled progress in the field of “host manipulation by parasites”⁷⁰. This gap could be filled by harnessing the tools and conceptual framework developed in chronobiology. Many of the examples above have employed an ecological approach, yet a chronobiological approach can help elucidate both proximate and ultimate explanations.

Conclusion

Over the past few decades, the focus of chronobiology has been to elucidate the mechanistic underpinnings of biological rhythms. We propose that now is the time to integrate this knowledge into parasitology, evolutionary ecology, and immunology (see Box 2). Indeed, the role of biological rhythms in infectious disease is a growing topic that holds promise for improving human and animal health. History clearly illustrates that attempts to control parasites are usually met with counter-evolution (in the form of drug resistance, vaccine escape, and host shifts). A comprehensive understanding of how rhythms affect parasite invasion and exploitation of a host (or vector) offers novel ways to disrupt the chain of transmission and treat disease. Further, clock coevolution may occur in host-parasite-vector interactions, resulting in complex arms races best understood through the lens of chronobiology coupled with evolutionary ecology. Chronobiology supplies a myriad of tools to help elucidate rhythmic phenotypes and reveal to what extent host and parasite genes are responsible for rhythms in disease phenotypes. Adding an evolutionary ecology framework will ensure this information is generalisable and used to make interventions as evolution-proof as possible.