Authors: Michael C. Tackenberg, Maria Luísa Jabbur, Vincent M. Cassone, Jeff R. Jones
Categories: Article from the Special Issue on Festschrift in Honor of Terry Page; Edited by Karen Gamble, Douglas McMahon, Julie Pendergast, and Carl Johnson
Source: Neurobiology of Sleep and Circadian Rhythms
Authors: Michael C. Tackenberg, Maria Luísa Jabbur, Vincent M. Cassone, Jeff R. Jones
A circadian clock enables an organism to occupy a particular temporal niche, and the evolutionary pressures associated with that temporal niche in turn shape the organization of the circadian network in which the clock is embedded. Although circadian organization has often been dichotomized into centralized and distributed systems, our modern understanding of circadian networks reveals a more complex organization that cannot be captured by this simple dichotomy. In this review, we examine how coupling between nodes of the circadian network (from cells, to tissues, to organs) gives rise to coherent circadian organization in mammals. We further highlight how comparative research on non-mammalian organisms reveals conserved and divergent strategies for circadian coupling that inform general principles of circadian network function across species.
“We can be sure that all vertebrate species would be diurnal if they could ‘get away with it’” - Gordon L. Walls (1944)
Circadian rhythms in multicellular organisms are an emergent property of a hierarchical network of autonomous circadian pacemakers at the cellular, tissue, and organ levels (Bechtel, 2024). This network rhythmically regulates a wide range of behavioral and physiological processes through the interaction between endogenous timekeeping, intercellular communication, and influence from environmental time cues. The concept of circadian organization – the precise layout of the network, including the degree to which its constituent cells respond to environmental stimuli, the strength and structure of interactions among oscillators (“coupling”), and the outputs the network controls – has been a topic of intense study (Gonze et al., 2005; Helfrich-Förster et al., 1998; Menaker, 1996). Conceptually, circadian organization across species generally falls within a spectrum between two centralized and distributed (Kumar and Sharma, 2018). A fully centralized circadian system relies on a central clock that responds to environmental time cues and transmits pacemaking information to peripheral oscillators. Conversely, a fully distributed (or “decentralized”) circadian system comprises individual cellular clocks that each receive environmental input directly, with inter-oscillator coupling ensuring synchrony.
The earliest characterizations of circadian organization centered on rodents and Drosophila, and in both cases in the context of entrainment to light/dark cycles (Helfrich-Förster et al., 1998; Menaker, 1996). These studies revealed striking differences in circadian organization between species. In the relatively centralized organization of rodents (and other mammals), light information is detected by the retina and relayed to the suprachiasmatic nucleus (SCN) of the hypothalamus, which then coordinates rhythms throughout the body. In contrast, the more distributed organization of Drosophila features clocks throughout the body that are independently photoreceptive. Of course, neither system is fully centralized nor fully distributed. In rodents, some clocks in peripheral tissues can maintain synchrony in the absence of SCN influence (Sinturel et al., 2021) and respond independently to non-photic zeitgebers (Damiola et al., 2000) and even photic cues (Welz et al., 2019). Likewise, though isolated Drosophila tissue clocks are independently photoreceptive (Plautz et al., 1997), circadian rhythms in the intact fly are regulated by clock neurons in the brain (Hermann-Luibl and Helfrich-Förster, 2015). Together, these observations suggest that centralized and distributed organizational models represent alternative strategies for using circadian coupling to overcome differences in the exposure of cellular or tissue clocks to relevant zeitgeber cues (e.g., via body opacity or burrowing behavior) which are used to various degrees and in combination by multicellular organisms to properly align with the rhythmic environment.
Circadian rhythms observed in nature are often dichotomized into active and resting phases, with the alignment of those phases and the solar day determining whether the rhythm is considered “diurnal” or “nocturnal.” When considering the relative advantages of these two temporal niches, Walls suggested that the inherent benefits in visual acuity and resource availability during the light phase were sufficient that all vertebrate organisms should tend toward diurnality (Walls, 1944). Because these benefits cut both ways, however (for instance, improved visual acuity favors both an organism and its predators) there exists evolutionary pressure toward nocturnality. Walls argues that the earliest mammals were all nocturnal, likely because this pressure led to an evolutionary bottleneck on temporal niche at the time of mammalian emergence.
The environmental challenges of nocturnality shaped the evolution of mammalian photoreception in general, including circadian photoreception. The concept of a nocturnal bottleneck, originally proposed by Walls as an explanation for mammalian eye evolution, was expanded by Menaker and colleagues in 1997 to describe the impact that nocturnal activity patterning had on the organization of the circadian system at the organismal level (Menaker et al., 1997). By becoming “secretive” (as Walls described it) during the day, mammals limited their exposure to sunlight and the time cues it provides. As a result, the circadian system of mammals is thought to have evolved a greater reliance on an indirectly-photoreceptive central clock structure (which we now know to be the SCN) to orchestrate the rhythms of peripheral clocks throughout the body. Other organisms that remained diurnal (such as plants, Drosophila, and non-mammalian vertebrates such as birds and zebrafish), and thus were more consistently exposed to the time cues of the sun, were able to maintain more distributed organizations where direct photoreception is more heavily emphasized (Plautz et al., 1997; Shimizu et al., 2015; Whitmore et al., 1998).
Walls argued that mammals were evolutionarily nocturnal because of this bottleneck, but he notes that other non-mammalian species from typically diurnal taxa have in some cases developed nocturnality. As further evidence that temporal niche shapes circadian organization, the circadian network of the nocturnal cockroach is organized similarly to that of mammals, with a centralized clock within the optic lobe that provides timekeeping regulation to the rest of the body. These similarities were leveraged by Dr. Terry Page, who demonstrated that transplantation of the cockroach central pacemaker (the optic lobe) was sufficient to reproduce the circadian period of the donor cockroach in the recipient cockroach. This experiment, performed with the tissue transplant placed in its original anatomical context, preceded analogous mammalian transplantation studies by nearly a decade (Page, 1982).
Though first proposed in 1942, the nocturnal bottleneck hypothesis has withstood modern scrutiny (Gerkema et al., 2013). As a result of that bottleneck, there is a clear correlation between the physiology of an organism, including its circadian organization, and the temporal niche it occupies. The circadian clock thus enables the occupancy of a temporal niche, and the evolutionary pressures of that niche feed back onto the system to shape its organization.
“The intelligence of the creature known as the crowd is the square root of the number of people in it.” – Terry Prachett (Pratchett, 1997)
Circadian clocks have evolved to take advantage of the rhythmic nature of the environment, and different forms of circadian organization have evolved to meet the demands of maintaining appropriate timing in complex, multi-tissued organisms. As organismal complexity increases from a single cell to heterogeneous cell populations, specialized tissues, and organ systems, the requirements for coherent circadian timing likewise become more complex. Regardless of scale, however, circadian organization ultimately depends on among cells, tissues and organs, as well as between the organism and its environment. In this section, we discuss several examples of coupling within the mammalian circadian system and how those coupling mechanisms contribute to overall circadian organization.
Because “coupling” refers to interactions across multiple biological scales with distinct meanings, a precise definition is required before considering circadian organization and its underlying mechanisms. Here, coupling refers to any mechanism that links the timing of one (circadian) oscillator to another oscillator or zeitgeber such that their phases, periods, and/or amplitudes are coordinated. In this review, we deliberately use a broad definition of coupling that spans interactions among clocks in cells, tissues, and organs, and the entrainment of internal time to environmental zeitgebers.
At a mechanistic level, coupling can be described in three complementary ways. Structural coupling refers to physical connections between oscillators, such as synaptic projections, gap junctions, or paracrine signaling (Michel and Colwell, 2001). Functional coupling refers to dynamical coordination between oscillators, such as stable phase relationships, synchrony, or resistance to damping regardless of whether the causal relationship is known (Nikhil et al., 2025). Effective coupling, in contrast, refers specifically to causal relationships inferred from perturbations of the coupled oscillators such as lesions or genetic disruptions (Tokuda et al., 2018).
Considering the importance of proper circadian coupling in maintaining appropriate relative timing throughout the network, it is no surprise that its disruption is associated with numerous dysfunctions including the development of metabolic syndrome and Type 2 diabetes (Mason et al., 2020; Maury et al., 2014; Tran et al., 2024), cancer (Fekry and Eckel-Mahan, 2022), and mood disorders (Dollish et al., 2024; Vadnie and McClung, 2017). Importantly, diseases that are precipitated, accelerated, or exacerbated by circadian disruption may induce further disruption, creating a damaging reciprocal relationship (Fishbein et al., 2021). These disruptions cause changes not just in the timing between cells, but also at the molecular circadian disruption has been shown to alter both the rhythmic transcriptome and proteome (Duong et al., 2024).
We should note, however, that there is a difference between the disruption of proper coupling and a reduction in coupling strength. Strong coupling among cells allows a network to be resistant to certain perturbations (Liu et al., 2007), but that resistance also means a relative lack of flexibility. More weakly-coupled networks are able to phase shift more quickly than more strongly-coupled networks (An et al., 2013; Yamaguchi et al., 2013), and the degree of coupling within a tissue may determine its ability to entrain to certain zeitgebers (Abraham et al., 2010). The disruption of circadian coupling is therefore not necessarily only a reduction in coupling strength, but rather a deviation from the proper degree of coupling.
In experimental and theoretical circadian neuroscience, “coupling strength” can be quantified using different metrics, such as phase-locking, phase-amplitude coupling, correlation, or coherence (Schmal et al., 2018). In general, causal inference is strongest when experimental perturbations produce predictable changes in coupling strength. In the following sections, we therefore emphasize evidence resulting from these perturbations when available and acknowledge that correlational or coherence-based observations need to be tested directly.
As the central mammalian circadian clock, much of circadian coupling involves the SCN. Circadian coupling involving the SCN can be characterized at four distinct 1) coupling within the SCN, 2) SCN outputs to other brain structures, 3) SCN outputs mediated by diffusible factors, and 4) SCN outputs to peripheral tissues. In the following subsections, we describe the mechanisms that govern circadian coupling at each of these levels.
Coupling within the SCN is essential for transforming a population of heterogeneous neuronal oscillators into a coherent and precise circadian pacemaker (Hastings et al., 2018; Herzog, 2007). Individual SCN neurons contain self-sustained molecular clocks and exhibit intrinsic daily rhythms in electrical activity, but these rhythms vary widely in intrinsic period, phase, and amplitude across cells (Colwell, 2011). In the absence of intercellular coupling, such heterogeneity leads to progressive desynchronization and damping of network-level rhythms even though many individual neurons remain rhythmic. Experimental manipulations that weaken coupling, such as pharmacological blockade of synaptic signaling or genetic disruption of critical neuropeptidergic pathways, consistently degrade phase coherence and reduce both the precision and robustness of circadian outputs (Aton et al., 2005; Harmar et al., 2002; Liu et al., 2007). Coupling therefore gives rise to emergent properties that cannot be explained by single-cell oscillators alone, including stable phase alignment, resistance to “noise,” and the ability to maintain high-amplitude rhythms over long time periods (Tokuda et al., 2018).
Mechanistically, coupling in the SCN is mediated by interacting fast synaptic and slower neuromodulatory signaling pathways that link together clock gene transcription, intracellular calcium dynamics, and neuronal firing. The predominant fast neurotransmitter within the SCN is γ-aminobutyric acid (GABA), which allows for rapid local coupling and can either excite or inhibit target neurons depending on chloride gradients, developmental stage, and circadian phase (DeWoskin et al., 2015; Klett et al., 2024). GABAergic signaling acts through both ionic and metabotropic receptors to shape moment-to-moment firing rates and contribute to local synchrony and spatial patterning within the SCN network. However, this organization presents a how can a predominantly inhibitory circuit sustain robust firing rate rhythms? This apparent contradiction is reconciled by the discovery of circadian regulation of astrocytic GABA uptake, which rhythmically modulates extracellular GABA levels and gates network excitability, permitting high firing rates during the circadian day (Patton et al., 2023).
Neuropeptides provide the SCN with slower, but more potent, synchronizing influences that are superimposed on this fast signaling. Among these, vasoactive intestinal peptide (VIP) plays a central role in maintaining network coherence by acting on VPAC2 receptors to increase intracellular cyclic AMP (cAMP) and activate cAMP response element binding protein (CREB)-dependent transcription of the core clock gene Period across SCN neurons (Ono et al., 2023; Patton et al., 2020). Other neuropeptides, including arginine vasopressin (AVP) and gastrin releasing peptide (GRP) further refine SCN coupling by stabilizing phase relationships among individual neurons (Ono et al., 2021). Importantly, these fast and slow signaling pathways operate within closed feedback loops, as activity-dependent calcium influx and firing rate changes feed back onto the molecular clock, which in turn modulates neuronal excitability and transmitter release (Enoki et al., 2017; Jones et al., 2015).
The functional consequence of this multilayered SCN coupling is a resilient network oscillator that can entrain reliably to environmental light cues while remaining resistant to weak inputs and internal physiological noise (Hafner et al., 2012). Coupling also supports the emergence of spatial and temporal structure within the SCN, including regional phase gradients and traveling “waves” of calcium activity and clock gene expression that may optimize signal integration and downstream communication (Enoki et al., 2012; Evans et al., 2011). Although the core components of SCN coupling are now relatively well characterized, several fundamental questions about how these interactions are organized and regulated at the network level remain unresolved, including how coupling strength and signal weighting differ across SCN subpopulations and how these relationships change with photoperiod, sex, aging, or disease. Disruption of intercellular coupling in the SCN also provides a parsimonious mechanism by which circadian timing can degrade without a complete loss of cellular rhythmicity.
Current evidence supports a constrained, but highly structured, model of SCN-to-brain connectivity (Starnes and Jones, 2023). Rather than broadcasting broadly throughout the brain, the SCN concentrates its strongest monosynaptic efferents within nearby hypothalamic and thalamic regions including, most prominently, the subparaventricular zone (SPZ), paraventricular nucleus of the hypothalamus (PVN), dorsomedial hypothalamus (DMH), and paraventricular thalamus (PVT) (Abrahamson and Moore, 2001). These regions function as hubs that are well-positioned to redistribute circadian timing information into relevant endocrine, autonomic, and behavioral control systems. However, this constrained projection pattern does not imply uniform output, as genetic tracing has revealed that SCN projections are organized by neuronal identity. For example, VIP-, AVP-, GRP-, and prokineticin 2 (PK2)-expressing neurons each show overlapping, but distinct projection patterns, indicating that SCN output is multiplexed across parallel information channels rather than conveyed as a single signal (Sunkin et al., 2013; Todd et al., 2020; Zhang et al., 2009). Additionally, a small number of sparse, cell-type-restricted connections have been identified, such as AVPergic projections to the central amygdala (François et al., 2025). However, for many SCN subpopulations, such as cholecystokinin (CCK)- and neuromedin S (NMS)-expressing neurons, a comprehensive, cell-type-specific output map has not yet been established.
Surprisingly, only a subset of SCN efferent pathways have been causally linked to specific circadian outputs, and even in these cases the functional architecture typically involves polysynaptic, multi-nodal circuits rather than monosynaptic SCN-to-target connections. The best-resolved examples of SCN outputs include regulation of sleep and wake timing through projections to the SPZ and DMH, feeding and locomotor rhythms through interacting SCN-DMH-arcuate nucleus (ARC) circuits, endocrine and autonomic rhythms via the PVN, body temperature rhythms through convergent SCN and ARC inputs to the median preoptic nucleus, and reproductive timing through coordinated SCN signaling to the medial preoptic area and anteroventral periventricular zone (Deurveilher and Semba, 2005; Guzmán-Ruiz et al., 2015; Teclemariam-Mesbah et al., 1997; Van Der Beek et al., 1997; Vida et al., 2010). Many SCN outputs are routed through the SPZ, which functions as a prominent integrative node (rather than an obligate intermediary), consistent with SCN output pathways being distributed in parallel across multiple downstream targets (Vujovic et al., 2015). More complex behavioral rhythms appear to depend on polysynaptic pathways originating from the SCN. Daily rhythms in aggression likely provide the clearest example of an SCN output that has been anatomically traced, via the SPZ and ventromedial hypothalamus, to a circuit governing a complex behavior (Todd et al., 2018). Circadian influences on other behaviors such as memory, motivation, and mood are supported mainly by indirect anatomical and physiological evidence, and the organization of the underlying output circuits remains unresolved.
A central unresolved question is how this sparse synaptic architecture supports coherent circadian coordination across the brain. The available evidence indicates that SCN signaling is not solely synaptic but instead combines direct wiring with several forms of non-synaptic signaling (Balsalobre et al., 2000; Buhr et al., 2010). The SCN releases diffusible signals, including neuropeptides such as VIP, AVP, and PK2, which can act locally through paracrine mechanisms, reach periventricular targets via the cerebrospinal fluid, or signal to other brain regions through specialized vascular connections (Maywood et al., 2011; Yao et al., 2021). Classic transplant studies showing that encapsulated SCN grafts can restore some, but not all circadian rhythms demonstrate that diffusible factors are sufficient for specific circadian outputs (Meyer-Bernstein et al., 1999; Silver et al., 1996). However, these results do not imply that synaptic projections are dispensable. Instead, these findings are consistent with a division of labor in which synaptic circuits contribute to target specificity and sensitivity to the functional state of the target circuits, while non-synaptic signals can act as phase-setting cues and may support the maintenance of rhythmicity in downstream circuits that vary in their dependence on synaptic SCN input. The efficacy and routing of these SCN outputs are likely further modulated by internal state, underscoring that circadian timing information from the SCN is dynamic rather than static. Collectively, this supports a model in which the SCN exerts parallel, state-dependent circadian control over multiple brain circuits, rather than solely operating as a serial broadcast signal.
As described above, SCN transplantation in mammals restores some, but not all, rhythmic properties of the organism, revealing that at least some of the pacemaking activity of the SCN is brought about through the secretion of diffusible substances. Several diffusible signals released by the SCN provide concrete examples of this non-synaptic circadian coupling. A small number of cytokines exhibit circadian patterns of expression and release and, when administered centrally, can modulate rhythmic behavior (Li et al., 2012). These include PK2, a peptide that is synthesized primarily during the subjective day (Cheng et al., 2002), transforming growth factor α (TGFα), which peaks around dawn (Kramer et al., 2001), and cardiotrophin-like cytokine (CLC), which is released during the late subjective day (Kraves and Weitz, 2006).
Receptors for these molecules are expressed in periventricular and hypothalamic regions adjacent to the SCN, and infusion of each factor into the third ventricle suppresses locomotor activity when administered during the subjective night. These findings are consistent with their proposed role as SCN-derived diffusible output signals (Li et al., 2012). Notably, these diffusible outputs are sufficient to impose circadian structure on specific behaviors but do not recapitulate the full range of SCN-regulated rhythms. This pattern is consistent with non-synaptic signaling contributing selectively within a broader, multi-modal SCN output strategy.
In addition to coordinating clocks throughout the brain, the SCN regulates peripheral organ rhythms through multisynaptic autonomic pathways, largely via hypothalamic pre-autonomic nodes such as the PVN. These pathways provide clear examples of how synaptic SCN outputs are converted into system-level circadian control, with different organs relying on distinct combinations of autonomic input, metabolic cues, and local clock autonomy.
The pineal gland represents a canonical example of strictly sympathetic circadian regulation. SCN projections to pre-autonomic neurons in the PVN initiate a multisynaptic pathway through the intermediolateral cell column of the thoracic spinal cord and the sympathetic chain to the pineal gland, where noradrenergic input stimulates the nocturnal biosynthesis and release of the indoleamine hormone melatonin (Klein et al., 1997). Disruption of this pathway at any level – from SCN ablation to PVN lesion to sympathectomy – abolishes pineal melatonin rhythmicity, underscoring the dependence of this peripheral oscillator on intact SCN-driven sympathetic input. Notably, there is no evidence for parasympathetic regulation of pineal melatonin rhythms.
Other peripheral tissues exhibit more complex coupling logic that combines local clock autonomy with SCN-mediated synchronization. The gastrointestinal system exhibits circadian rhythms in motility, secretion, absorption, and gene expression that depend on the molecular clock, as the loss of core clock genes abolishes rhythmic motility both in vivo and in vitro (Hoogerwerf, 2010). However, the amplitude and phase of these rhythms are shaped by SCN output via sympathetic pathways and feeding-related cues. For instance, SCN lesions or chemical sympathectomy reduce the rhythmic amplitude of gut motility, which can be restored through timed feeding (Malloy et al., 2012). Notably, surgical vagotomy does not disrupt intestinal rhythms, indicating minimal parasympathetic involvement in the circadian regulation of the gut (Hoogerwerf et al., 2007).
Similarly, the liver contains a self-sustained circadian clock that drives rhythms in metabolic gene expression and many biochemical processes (Asher and Schibler, 2011; Storch et al., 2002). Hepatic rhythms are entrained by timed feeding and influenced by autonomic input arising from multisynaptic SCN pathways. However, in contrast to both the pineal gland and the gastrointestinal system, parasympathetic vagal signaling plays a prominent role in synchronizing hepatic rhythms, underscoring that peripheral organs differ fundamentally in how autonomic pathways are used to convey circadian timing information (Cailotto et al., 2005, 2008; Woodie et al., 2024b).
While our understanding of coupling between the SCN and the periphery is growing, the degree and mechanisms of coupling within and among peripheral tissues remain less well-defined. Under normal conditions, cellular clocks within a peripheral tissue may be synchronized by signals from the SCN, by direct environmental input, or through intercellular communication within the tissue itself. Many peripheral tissues cultured in isolation exhibit circadian rhythms in clock gene expression, though the degree of damping is much greater in these tissues than in the SCN, demonstrating that these tissues are capable of maintaining some degree of intercellular synchrony without input from the SCN or the external environment (Yoo et al., 2004). In vivo, similar results are observed where the surgical ablation of the SCN results in a loss of rhythmic coherence across peripheral organs while maintaining partial synchronization within the liver (Sinturel et al., 2021). As described above, many peripheral oscillators are directly responsive to the timing cues provided by non-photic stimuli like feeding (Damiola et al., 2000), which may exert their influence throughout the body through metabolic signaling molecules like insulin and IGF-1 (Crosby et al., 2019).
Outside of the context of the brain, where intercellular signaling networks are anatomically defined and extensively characterized, the mechanisms of phase communication are still being revealed. Recently, TGF-β has been identified as a putative mediator of circadian information between cells in the periphery (Finger et al., 2021). The degree of local communication varies by tissue type, ranging from cultured fibroblasts, which exhibit little to no phase communication under standard conditions (Nagoshi et al., 2004), to hepatocytes which exhibit some form of local coupling (Guenthner et al., 2014). Other studies have demonstrated that mice rendered generally arrhythmic through the global deletion of Bmal1 display damped but detectable hepatic rhythmicity when Bmal1 expression is restored in that tissue (Koronowski et al., 2019). The functional importance of this partial independence is emphasized by the finding that when the liver clock period is selectively shortened via the deletion of REV-ERBα/β while the SCN period remains typical, the metabolic consequences of high-fat diet consumption are greater than when the SCN period and liver period match (Woodie et al., 2024a).
“A process which led from the amoeba to man appeared to the philosophers to be obviously a progress—though whether the amoeba would agree with this opinion is not known.” – Bertrand Russell (1915)
The functional expression of circadian rhythms is well-conserved across the tree of life (Dunlap and Loros, 2017; Golden and Canales, 2003; Kondo and Ishiura, 2000), and within metazoa, that conservation extends to homology at the level of the molecular clock (Waldridge et al., 2025). The broad distribution of circadian rhythms in nature thus provides powerful comparative examples of how coupling mechanisms are used to coordinate timing within organisms (Bell-Pedersen et al., 2005). In the following section, we discuss several non-mammalian systems that illustrate alternative strategies of circadian organization and consider how these examples inform our understanding of circadian coupling in mammals.
The earliest experimental attempts to restore circadian rhythms through tissue transplantation were performed in cockroaches, by implanting the suboesophageal ganglia of rhythmic donors into the abdomen of headless hosts (Harker, 1956). These controversial (Brady, 1967; Roberts, 1966) experiments initiated a series of transplantation studies aimed at identifying the tissues necessary and sufficient for generating circadian rhythms. Truman and Riddiford subsequently demonstrated that the transplantation of brain tissue from any one of three different silk moth species into the abdomen of a “brainless” host of one of the other species restored molting rhythms with the timing determined by the donor tissue and the behavior pattern determined by the host (Truman and Riddiford, 1970). About a decade later, Handler and Konopka showed in Drosophila that transplantation of per^S^ brain tissue into the abdomen of per^0^ hosts was capable of restoring short behavioral rhythms (Handler and Konopka, 1979). That same year, Zimmerman and Menaker extended this approach to vertebrates by transplanting pineal tissue into the eye of pinealectomized sparrows, restoring locomotor activity rhythms with a phase determined by the donor pineal rhythm (Zimmerman and Menaker, 1979).
In each of these cases, rhythmicity was restored after transplantation of tissue outside its normal anatomical context, strongly implicating diffusible, likely hormonal, mechanisms. However, the applicability of this logic to nocturnal organisms with centrally organized circadian systems was less clear. In the silk moth, the only nocturnal example assessed at the time, transplantation restored molting rhythms but not flight rhythms, indicating that some circadian outputs required intact neural connectivity (Truman, 1974). Consistent with this result, several neural connections between candidate pacemaker structures and the rest of the nervous system disrupted rhythmicity in both rats and cockroaches, suggesting that hormonal signaling alone was insufficient to convey circadian timing in these (Moore and Klein, 1974; Nishiitsutsuji-Uwo and Pittendrigh, 1968).
It was therefore pivotal when, in 1982, Terry Page demonstrated that the transplantation of cockroach optic lobes restored circadian rhythms, conferred the donor period to the host, and, critically, did so by regenerating the neural connections required to transmit timing information to the rest of the organism (Page, 1982). This experiment provided the first clear demonstration that a central circadian pacemaker could restore rhythmicity through synaptic, rather than purely diffusible, signaling. These findings directly paved the way for analogous experiments in mammals, first showing restoration of circadian period using SCN transplants (Drucker-Colín et al., 1984; Lehman et al., 1987) and later demonstrating that donor SCN tissue determines the recipient period (Ralph et al., 1990).
The circadian rhythms of Aves have numerous conceptual connections to those of Mammalia, primarily related to the role of the pineal gland and melatonin. In contrast to that of mammals, the discovery of a pacemaker in birds began with the demonstration that the pineal gland was required for overt locomotor and body temperature rhythmicity in house sparrows, Passer domesticus (Binkley et al., 1971; Gaston and Menaker, 1968). As described above, Zimmerman and Menaker later demonstrated that the transplantation of the pineal gland into the eye of arhythmic, pinealectomized house sparrows conferred both the rhythmicity and the phase of the donor, suggesting a hormonal mechanism of pineal output (Zimmerman and Menaker, 1979). That hormone, the indoleamine melatonin (Klein et al., 1997), is synthesized and secreted by the pineal gland with high levels during the night and low levels during the day, a rhythm that persists in constant darkness and can be entrained to environmental light (Binkley et al., 1978; Deguchi, 1979; Takahashi et al., 1980). In several avian species, rhythmic administration of melatonin to arrhythmic, pinealectomized individuals restores rhythmicity and entrains coherent rhythms of locomotion, feeding, vocalization, and body temperature (Chabot and Menaker, 1992; Heigl and Gwinner, 1994; Lu and Cassone, 1993; Wang et al., 2012). While not required for rhythmicity as they are in birds, melatonin and the pineal gland play a substantial role in mammalian rhythms. The SCN of several mammalian species, including humans, expresses melatonin receptors (Reppert et al., 1988, 1994). In mammalian experiments following pioneering work done in birds, free-running rats were found to be able to entrain to daily injection of melatonin (Redman et al., 1983) in a dose- and phase-dependent fashion (Cassone et al., 1986a), an effect abolished by SCN lesion (Cassone et al., 1986b). Administration of melatonin to rats inhibits glucose utilization (Cassone et al., 1988) and electrical activity (Shibata et al., 1989) in a dose- and phase-dependent fashion in parallel to its effects on behavior.
The avian suprachiasmatic complex comprises two interconnected structures in the basal the visual SCN (vSCN) and the medial suprachiasmatic nucleus (mSCN), mirroring the classical subdivision of the mammalian SCN into the ventrolateral (or “core”) SCN and dorsomedial (or “shell”) SCN. The site of melatonin action in the avian brain for circadian entrainment is almost certainly the vSCN, which, like the mammalian ventrolateral SCN, receives direct retinal input via the retinohypothalamic tract (RHT) (Cantwell and Cassone, 2006; Cassone and Moore, 1987). Also similar to the mammalian SCN, the avian vSCN receives serotonergic (5HT) and neuropeptide Y (NPY) afferents that overlap with the RHT synaptic field. Considered together, these results show that the general motif of SCN and circadian organization is similar between mammals and birds, where in birds, the pineal gland is perhaps a co-master pacemaker with the vSCN/mSCN, while in mammals, the role of the pineal gland is diminished with feedback from the pineal gland influencing the SCN.
Beyond physiological comparisons, the investigation of the evolution of avian temporal niches also reflects that of mammals. As described above, mammals are theorized to have descended from a nocturnal common ancestor, with instances of diurnality in mammals presumed to have developed later. Conversely, the large majority of bird species are diurnal, with limited examples of nocturnality (e.g., Strigiformes or owls). Interestingly, the pineal gland of several owl species have been found to be physically diminished in relation to diurnal avian orders, with the most nocturnal of the group, barn owls, having lost their pineal altogether (Quay, 1972).
Some insects and birds have the ability to use their internal circadian clock in conjunction with the position of the sun to accurately navigate, a phenomenon referred to as time-compensated sun-compass orientation (Beling, 1929; Kramer, 1950, 1952). In honey bees, that ability is used by foragers to locate food sources, return to the hive, and to transmit the location of that food source to other foragers (Wagner et al., 2013). When food sources are available only at some times of day, honey bee foragers are able to account for this rhythmic availability in what is referred to as time-memory or zeitgedächtnis (Moore, 2001). While zeitgedächtnis and time-compensated sun-compass navigation represent a remarkable moment-by-moment awareness of the time of day that (at least consciously) eludes us as humans, mammals do have some forms of time-memory. In a phenomenon termed “food anticipatory activity” or FAA, calorically-restricted rodents become active at phases in which food was previously available (Rosenwasser et al., 1984). Interestingly, while its specific timing can be influenced by the SCN (Rosenwasser et al., 1984), FAA does not require the SCN (Krieger et al., 1977; Stephan et al., 1979a, 1979b). These results have prompted the search for the anatomical basis of the food entrainable oscillator, or FEO (Davidson, 2009; Ehichioya et al., 2023), though a single responsible structure has yet to be identified. Though the response to feeding cues is often thought of as relatively distributed (Crosby et al., 2019), the existence of an SCN-independent FEO somewhere in the body suggests that a centralized organization may play a role in that response. The honey bee, which lacks an SCN yet features an even stronger version of the rodent FAA, may act as a powerful model for understanding how multiple clock structures (both within and outside of the brain) may constitute an effective FEO with a combination of centralized and distributed features. Further, the honey bee molecular clock is an excellent model for understanding how the memories associated with learned behaviors can be imprinted upon the circadian clock (Naeger et al., 2011; Roy et al., 2025).
Spiders are capable of displaying an exceptionally wide range of free-running period lengths, with three species of tangle web spiders (family Theridiidae) exhibiting periods between 19 and 30 h and one species of orb weaver (Cyclosa turbinata) having a period of roughly 19 h (Mah et al., 2020; Moore et al., 2016). These species provide an opportunity to study not only the molecular mechanisms capable of maintaining such “extreme” period lengths, but also how these atypical period lengths influence circadian organization and how evolutionary pressure influences the circadian cycle length. These studies will be important to better understand period length regulation in mammals at the molecular level, where the role of particular clock genes in influencing period length is still being actively investigated (Li et al., 2016; McManus et al., 2022, 2024; Tackenberg et al., 2025), as well as in understanding how disparate tissues can maintain stable phase relationships. The high inter-individual and inter-species variation in period length of spiders also points to a potential reduction in selection pressure for precise period length in these organisms (Mah et al., 2020), a phenomenon which may prove relevant to mammals in atypical lighting environments. One such example is the polar bear, which, like the spiders described above, exhibits a high inter-individual and intra-individual variability in period length in both constant light and constant darkness (Ware et al., 2020).
Our understanding of the circadian network has grown from a black-box neural pacemaker that drives oscillations in peripheral tissues, to an ensemble of tissue-level autonomous circadian clocks organized by the SCN, to the understanding that those tissue-level circadian clocks are themselves composed of a network of autonomous circadian clocks. As a result, the concept of circadian organization has naturally expanded to include not only the alignment of rhythmicity between tissues, but also that of rhythmicity within tissues. Moving another level deeper, we now understand that a variety of intracellular processes are governed by circadian rhythms (Chaix et al., 2016), and as such we can consider a non-traditional extension of circadian organization as the alignment of oscillations within the cell. Our knowledge of these intercellular and intracellular organization strategies has benefited from studies using cultured cells, computational modeling, and even from unicellular and syncytial microbial organisms.
Numerous advances in our understanding of circadian organization have come from the study of single mammalian cells in dispersed cultures. Observations in dispersed SCN neurons (Honma et al., 1998) and fibroblasts (Welsh et al., 2004) revealed autonomous rhythms in these cells. Later, comparisons between dispersed SCN neurons (Honma et al., 2004) or fibroblasts (Liu et al., 2007) and cultured SCN slices revealed that intercellular period regulation benefits from an intact network structure, with differential network organization strength within tissues like the SCN and lung impacting resilience to disruption. Computational models of these types of intercellular circadian relationships have provided valuable information about how coupling impacts network period length (Gu et al., 2016) and the range of entrainment (Abraham et al., 2010), how signaling within the SCN can facilitate coupling (Tokuda et al., 2018), and have provided strategies for measuring coupling (Schmal et al., 2018).
The study of the circadian clocks of microbial organisms has also offered insights into how the intracellular rhythms of mammalian cells are organized. The concept of transcriptional control of output genes (“clock-controlled genes”, or CCGs), which is the basis for a single molecular clock producing organized outputs with multiple phases, was first discovered in the filamentous fungus Neurospora crassa (Loros et al., 1989), roughly a decade prior to the first characterization of a CCG in mammals (likely Dbp, Ripperger et al., 2000). Experiments on the dinoflagellate Gonyaulax polyedra demonstrated that circadian phase was preserved in daughter cells following cell division (Homma et al., 1990), a result that was recapitulated in mammalian fibroblasts years later (Nagoshi et al., 2004).
Insights into mammalian rhythms are not limited to those from eukaryotic organisms. Despite the fact that cyanobacteria, the oldest extant organism to possess bona fide circadian clocks, have no homology with mammals in their clock genes, the characteristics of their rhythms – their period, their phase response curve, their temperature compensation – are consistent with mammalian rhythms. These similarities are remarkable, considering the stark differences in their clock eukaryotic rhythms are generated by a transcription-translation feedback loop, while cyanobacteria rhythms are generated by a (transcriptional rhythm-independent) post-translational feedback loop based on the rhythmic phosphorylation of the protein KaiC (Fang et al., 2025). Recently, however, there has been a growing appreciation for the potential roles of post-translational mechanisms in maintaining mammalian circadian rhythmicity (Crosby and Partch, 2020; Finger et al., 2020). The protein RUVBL2, a clock component that is conserved across eukaryotes, was recently shown to regulate mammalian period through very slow ATPase activity (Liao et al., 2025) – a mechanism notably similar to period regulation in cyanobacteria by the slow ATPase activity of KaiC. Given RUVBL2's phylogenetic breadth, it is possible that slow ATPases like RUVBL2 and KaiC could be primordial components of circadian clocks that may, in time, shed light upon the evolution of the organization of intracellular circadian rhythms across phylogeny.
“The complexity of things – the things within things – just seems to be endless. I mean nothing is easy, nothing is simple.” – Alice Munro (Quinn, 2001)
Multicellular organisms are complex, and the circadian organization of those organisms resists simple classification. While centralized and distributed circadian organization schemes may not exist as absolutes in nature, their interwoven combination may be the most accurate representation of how these complex networks are organized. Of course, if the schematized view of a system is just as complicated as the system itself, its value is limited. We therefore propose that circadian organization be considered in mammals (as well as in other taxa) in a zeitgeber-by-zeitgeber, node-by-node manner. In effect, this is the same approach as was taken in the earliest days of circadian experimentation, when the zeitgeber under consideration was light, and the node being considered was overt, whole-organism circadian rhythmicity like that of locomotor behavior or eclosion. Now, however, we allow for the fact that different zeitgebers (e.g., food intake or exercise) will have differential effects on other nodes of the circadian network (e.g., the SCN or the liver). A given network of entrainable cells may then be subject to one zeitgeber (Fig. 1A) or another (Fig. 1B), or both (Fig. 1C). Even in response to a single zeitgeber, there may be heterogeneity in the sub-networks within a given network may be organized differently than the rest of the network (Fig. 1D). A population of cells with a distributed circadian organization in response to a zeitgeber (Fig. 1D, black-outlined circles) may relay that timing information to a subpopulation of cells (Fig. 1D, orange-outlined circles), forming a centrally-organized subnetwork within a distributed organization.Fig. 1Schematic view of different circadian organizational strategies. A, a centralized network in which cues from Zeitgeber A are received by a central pacemaker (blue circles) and relayed to a subset of the rest of the network (blue-outlined circles). B, a distributed network in which cues from Zeitgeber B are received directly by each cell of the network (green circles). C, a hybrid network in which cues from Zeitgeber A and Zeitgeber B are received and relayed in combination. D, a subnetwork with centralized organization (orange-outlined circles) within a network with a distributed organization (black-outlined circles).Fig. 1
As an example, we can schematize that SCN response to excitatory signals from the RHT follows a centralized organization, with retinorecipient SCN neurons responding to the zeitgeber and relaying the signal to other SCN neurons. Likewise, we can visualize the effect of food intake on the peripheral organs of the body as following a distributed organization, with cues from zeitgeber being received individually by peripheral organs (Heyde and Oster, 2019). Zooming in, we can consider whether hepatocyte rhythms within the liver follow a centralized or distributed organization based on their zone, depending on which potential zeitgeber – body temperature, feeding phase, or signals from nearby cells – is being considered (Droin et al., 2021). In each case, we can examine the ways in which the interaction between the rhythmic node and zeitgeber in question has been shaped by factors like the temporal niche occupied by the organism, the accessibility of that zeitgeber to that rhythmic entity, and how that node must transmit information received from the zeitgeber to other nodes.
This reconsideration of mammalian circadian organization is only a proposed reframing of our thinking to allow for greater nuance, and the real progress will come in the form of experimentally untangling that nuance. There remain substantial gaps in our understanding of the relationships that entrainable nodes of the circadian network form with one another and with the environment. These gaps represent the areas of greatest need for further mechanisms of intercellular communication of circadian information, and mechanisms for receiving timing information. Much of what we know about circadian communication between cells is based on neurons, progress that has benefited from foundational knowledge of neural communication schemes (Colwell, 2011). That work has been expanded to include non-neuronal cells, albeit still within the brain, with investigations of the relationship between neurons and astrocytes within the SCN (Brancaccio et al., 2017, 2019; Hastings et al., 2023; Patton et al., 2023). While candidates for facilitating intercellular phase communication (Finger et al., 2021) and non-photic responses (Crosby et al., 2019) outside of the brain have been identified, more work is urgently needed to more completely understand the mechanisms employed by these cells throughout the body to receive and communicate timing information.
Michael C. Tackenberg: Writing – review & editing, Writing – original draft, Visualization, Project administration, Conceptualization. Maria Luísa Jabbur: Writing – review & editing, Writing – original draft, Visualization. Vincent M. Cassone: Writing – review & editing, Writing – original draft, Conceptualization. Jeff R. Jones: Writing – review & editing, Writing – original draft.
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.