Ontogeny of the circadian system: a multiscale process throughout development

1 Ontogeny of the circadian system: A multiscale process throughout development 1 Maria Comas1, Davide De Pietri Tonelli2, Luca Berdondini3 and Mariana Astiz1,4,5* 2 3 1Circadian Physiology of Neurons and Glia Laboratory, Achucarro Basque Center for 4 Neuroscience, 48940 Leioa, Basque Country, Spain. 5 2Neurobiology of miRNA, Fondazione Istituto Italiano di Tecnologia (IIT), 16163 Genova, Italy. 6 3Microtechnology for Neuroelectronics, Fondazione Istituto Italiano di Tecnologia (IIT), 16163 7 Genova, Italy. 8 4IKERBASQUE, Basque Foundation for Science, Bilbao, Spain. 9 5Institute of Neurobiology, University of Lübeck, 23562 Lübeck, Germany. 10 11 *Correspondence: Mariana Astiz, PhD. Circadian Physiology of Neurons and Glia Laboratory, 12 Achucarro Basque Center for Neuroscience, 48940 Leioa, Basque Country, Spain. Email: 13 [email protected]. Tel: +34 946018160 14 Keywords: circadian clock, mouse, humans, suprachiasmatic nuclei, astrocytes, neurons 15 Abstract: 16 The 24 h (circadian) timing system develops in mammals during the perinatal period. It carries 17 out the essential task of anticipating daily recurring environmental changes to identify the 18 best time of day for each molecular, cellular and systemic process. While significant knowledge 19 has been acquired about the organization and function of the adult circadian system, relatively 20 little is known about its ontogeny. During the perinatal period, the circadian system 21 progressively gains functionality under the influence of the early environment. This review 22 explores current evidence on the development of the circadian clock in mammals, highlighting 23 the multi-level complexity of the process and the importance of gaining better understanding 24 of its underlying biology. 25 26 2 The emergence of the mammalian circadian system 27 Life on Earth has evolved under the influence of geophysical cycles that generate recurrent 28 environmental changes with stable periods. The diurnal oscillations (see glossary) of light 29 intensity and temperature, as a result of the Earth rotating around its axis, have a period of 24 30 hours (h) and influence every organism’s physiology. These oscillations are efficiently 31 anticipated by the circadian system, which is necessary for organizing sleep-wake cycles and 32 most physiological rhythms to match the 24 h environmental period (the term circadian 33 derives from the Latin words circa – approximately, and dies – day) (reviewed in [1]). The 34 circadian system can be conceptualized as a three-component system: it requires input 35 pathways providing temporal information to an “entrainable” (but self-sustained) oscillator 36 which is, in turn, able to produce the rhythmic output necessary to synchronize different 37 processes. In multicellular organisms this three-component system can be found at three 38 levels: molecular, tissue/circuit and systemic [2]. 39 Compared with the knowledge that has been acquired on the adult circadian system, its 40 ontogeny is not well understood. The mammalian circadian system gains functionality 41 progressively under the dynamic influence of the perinatal environment (reviewed in [3]). In 42 both rodents and humans, the perinatal period is considered the critical window when the 43 development and maturation of the system takes place. While in humans most hypothalamic 44 nuclei, including the master clock in the suprachiasmatic nuclei (SCN), are mature by the end 45 of gestation, in rodents this process continues after birth (reviewed in [4]). 46 With a brief introduction on how the circadian system is organized in adults, this review 47 focuses on current evidence on the ontogeny of the master clock in mammals at the 48 molecular, circuit and systemic level. We highlight some of the less explored aspects of this 49 complex process and discuss the need for better understanding the influence of an adverse 50 perinatal environment later in life. 51 52 How is the adult circadian system organized? 53 Circadian rhythms in mammals are coordinated at systemic level by the master pacemaker in 54 the SCN, located on each side of the third ventricle (3V) and above the optic chiasm (OC). Light 55 3 signals are received by the retina and transmitted through neuronal pathways to the SCN, 56 thus, entraining the oscillator to the 24 h light-dark cycle [5,6]. In absence of light input, the 57 oscillator maintains self-sustained rhythms, with a period that approximates 24 h (τ) [7–9]. 58 The ability of the SCN to synchronize rhythms systemically, relies on extensive neuronal 59 projections and paracrine communication. These pathways reach different brain areas 60 including the medial hypothalamus coordinating hormone release and the tone of the 61 autonomic nervous system [10]. Two well-known examples of hormonal outputs are the 62 secretions of melatonin (during the dark phase) and glucocorticoids (GCs, before the active 63 phase) [11–13]. The SCN has been identified as the master circadian pacemaker by lesion 64 experiments. SCN-lesioned rats showed arrhythmic behavior, evidenced by an altered pattern 65 of locomotor activity, disrupted sleep-wake and feeding-fasting cycles and loss of circadian 66 hormonal levels [7,14]. The systemic rhythmicity can be restored by transplanting a graft 67 containing SCN in the lesion site, even if the graft is wrapped in a mesh blocking efferent 68 outgrowth, evidencing the relevance of paracrine output signals [15,16]. These key 69 discoveries, were followed by the characterization of neuronal populations, coupling 70 mechanisms between neighboring cells within the circuit, and the synaptic and paracrine 71 signalling pathways that keep the time. 72 The master clock circuit in adult rodents is formed by few thousands GABAergic neurons 73 highly interconnected, exhibiting precise and high amplitude circadian cycles of gene 74 expression, metabolic and electrical activity that persist autonomously in absence of light and 75 when cultured in vitro [17]. Topologically, the SCN is divided into a ventral core and a dorsal 76 shell. Neurons from the core receive glutamatergic innervation from the retina and propagate 77 their activation to the surrounding shell by releasing mainly gamma-aminobutyric acid (GABA) 78 and vasoactive intestinal polypeptide (Vip). Shell neurons, which in absence of light show 79 autonomous time-keeper activity with a period ~24 h, are then entrained and transfer their 80 synchrony to downstream clocks through efferent connections releasing vasopressin (Avp), 81 GABA and other signals (reviewed in [18]). Paracrine communication within the SCN circuit is 82 mediated by an heterogenous array of hundreds of neuropeptides (e.g. gastrin-releasing 83 peptide (Grp), neuromedin-S (Nms), among others) which are believed to confer the network84 level properties necessary to maintain highly robust oscillations even in absence of 85 environmental input [19]. This traditional neuron-centric view of the circuit has changed over 86 4 time, with demonstrations in rodents that astrocytes are competent circadian oscillators, 87 being able to modulate the clockwork of other cell types ([20–22], reviewed in [23]). In mice, 88 the astrocytes’ clock is sufficient and partially necessary to drive neuronal circadian rhythms 89 in the SCN, very likely contributing to the highly robust oscillations of the circuit [24–28]. In 90 light of this evidence, the current view on the adult SCN assumes a tight neuron-astrocyte 91 interaction, modulating paracrine and synaptic communication to maintain the circuit-level 92 properties of the oscillator and long-range organization of the output within and outside the 93 SCN (reviewed in [29]). As in the SCN, most other tissues/circuits show coupled oscillations 94 between neighboring cells, although they tend to lose synchrony in absence of timing cues. 95 The weaker coupling of extra-SCN tissues and the input-dependent synchrony seems to be 96 essential for the top-to-bottom coordination of tissue-specific circadian functions (reviewed 97 in [30]). 98 The molecular clock, which is present in virtually all the cells, is the third level of organization 99 of the circadian system. A set of core clock proteins (e.g., BMAL1, CLOCK, PER1-3, CRY1-2, 100 RORα, REV-ERBα/β) generate self-sustained oscillations that can be entrained by input signals. 101 As a result, interlocked auto-regulatory transcription-translation feedback loops (TTFLs) 102 produce molecular rhythms of about 24 h (reviewed in [31]). The TTFL regulates the rhythmic 103 expression of the so called “clock-controlled genes” (CCG) which represent, depending on the 104 cell type, between 10-40% of the transcriptome (reviewed in [32]). In addition, post105 transcriptional, translational and post-translational processes (including splicing, 106 polyadenylation, RNA binding proteins, microRNAs and other epigenetic and 107 epitranscriptomic mechanisms) were found to play a role in shaping the rhythmicity of 108 mRNAs, and in the accumulation of clock proteins [33–35]. The current knowledge on these 109 mechanisms is more limited. 110 111 What defines a functional circadian clock? 112 The mature circadian system depends on its multilevel nature (i.e.: molecular, cellular and 113 systemic) and on the integration of the three components: input, oscillator and output (Figure 114 1). During development, the mammalian clock starts gaining function in utero and it matures 115 completely after birth (reviewed in [36–40]). Understanding this process in its complexity is 116 5 imperative, especially because during the last decade, studies have revealed an association 117 between circadian disruption during pregnancy/early postnatal life and poor 118 neurodevelopmental outcomes in animal models and humans ([41-46], reviewed in [47]). To 119 understand the ontogeny of the circadian system, two critical notions to consider are: that the 120 functional clock is defined as “entrainable” and as able to maintain self-sustained rhythms. 121 This implies that the temporal input pathways have to respond to time cues and synchronize 122 or “entrain” the oscillator with a 24 h period (T). Furthermore, in the absence of 123 environmental cues, the oscillator must have the ability to generate self-sustained rhythms 124 that approximate a 24 h period (τ). Lastly, the multiple output pathways must produce signals 125 with a stable phase relationship, or phase angle, between the input and the oscillator to 126 ensure proper systemic synchronization. 127 Circadian rhythms are therefore innate. What remains unclear is when, during the process of 128 ontogeny, the circadian system starts to tick in an integrated and multiscale manner and what 129 is the influence of the early environment shaping the gain of function. The main experimental 130 difficulty has been to dissect whether an oscillation is driven by an entrainment synchronizing 131 signal (e.g. maternal rhythmic hormones reaching fetal/newborn tissues) or whether it is self132 sustained and autonomously generated (e.g. by the fetal/newborn clock). Moreover, little is 133 known about how the maturation process is coordinated during the perinatal period and how 134 milestone/sequential events lead to the gradual gain of functionality. Current evidence in 135 rodents and humans shows that the environment to which the system is exposed to during 136 development is critical for defining its functionality later on. Thereby, the functional 137 maturation of the circadian system results from a complex and likely dynamic interaction 138 between exogenous (i.e.: maternal/early environment signals) and endogenous factors (i.e.: 139 intrinsic molecular programs) (reviewed in [38], [47]). In the following sections we explore the 140 current knowledge on the maturation of the circadian system at all three levels, and discuss 141 current knowledge gaps. 142 143 The maturation of the molecular clock 144 During mammalian embryonic development progenitor cells undergo precise fate decisions 145 to become functionally mature. Circadian oscillations in the expression of the clockwork genes 146 6 appear to be absent in pluripotent stem cells (either embryonic or induced) [48]. Rhythmic 147 clock gene expression arises progressively and gains amplitude as the cell fate becomes more 148 defined, suggesting a tight coupling between the development of a functional molecular clock 149 and the cellular differentiation states ([49], reviewed in [50]). Post-transcriptional suppression 150 of CLOCK protein has been identified as one of the mechanisms setting the time for the 151 molecular clock to start ticking [51]. However, cell differentiation depends on genetic and 152 epigenetic determinants modulating whole regulatory networks that specifically define the 153 fate of each cell. Therefore, it seems plausible that multiple gene-modulatory mechanisms 154 (posttranscriptional [52,53], translational [54] and posttranslational [34,35,55] that are just 155 starting to be explored, play a role in this complex developmental process (reviewed in [56]). 156 As an example of the intricacy of these mechanisms, we highlight a link between genomic 157 imprinting, microRNAs, and light-experience during the critical period of the visual system 158 development in mice [57]. Interestingly, the majority of the imprinted miRNAs found, several 159 of which are known to control neural stem cell fate [58,59], were predominantly clustered into 160 the Dlk1-Dio3 locus, a genomic region that is associated with neuronal plasticity and several 161 neurodevelopmental disorders (reviewed in [60]). Hence, post-transcriptional/translational 162 mechanisms in the context of circadian ontogeny are of interest for future investigations. In 163 mammals, however, these mechanisms were mostly studied in peripheral organs, such as liver 164 and intestine [61], and the contribution of these mechanisms in different regions of the brain, 165 or in neural cell subpopulations, remains largely unknown. 166 In the mouse embryo, the rhythmic expression of clock genes has been detected quite early, 167 at about two-thirds of the in-utero development (13 days of gestation) [62–65]. In the 168 hypothalamus, this developmental stage is comparable to the end of the third trimester in 169 humans [4]. However, it is possible that these rhythms are not self-sustained and only induced 170 by maternal entrainment signals in vivo or by the culture conditions in vitro. Several maternal 171 signals such as melatonin [66], dopamine [67,68] and glucocorticoids [41,69] are essential for 172 promoting fetal growth and tissue maturation as well as for communicating time. All three are 173 produced in a circadian manner and are able to cross the placenta and reach fetal tissues ([70], 174 reviewed in [38,71]). Interestingly, maternal SCN lesions in rats or experiments in mice done 175 with clock deficient dams (Per1/Per2 double knockouts) have shown that during pregnancy, 176 the absence of maternal clock reduces the synchrony of fetal rhythms likely due to low 177 7 entrainment by maternal rhythmic signals [72,73]. However, the development of circadian 178 rhythms in pups is not impaired, suggesting a certain degree of fetal/newborn clock autonomy 179 during development [74]. Indeed, it has been demonstrated in mice that the molecular clock 180 in the fetus, already at 15 days of gestation, is able to gate the sensitivity of the fetal 181 hypothalamus to maternal glucocorticoids (GCs) at different times of day [41]. The modulation 182 observed in the activity of the GCs receptor in a circadian manner could be also relevant for 183 other maternal signals such as dopamine or melatonin [67,68,75]. These data suggest that 184 autonomic molecular clock mechanisms have an early role during fetal development, shaping 185 the influence of exogenous rhythmic signals. Similar mechanisms, that still need to be 186 explored, could be responsible for driving self-sustained molecular oscillations (i.e.: 187 functionality) of the developing clock. 188 189 The intercellular/circuit level maturation of the clock 190 Due to its complex structure, highly diverse cell composition and anatomical location, the 191 study of hypothalamic development has been slower than for other brain regions (reviewed 192 in [76]). The maturation of the SCN as a circuit result from a dynamic interplay between time 193 cues produced by the mother and intrinsic molecular programs, with variable influence along 194 the perinatal developmental window. During this period, the SCN circuit progresses from a 195 collection of undifferentiated individual cells to a highly differentiated, interconnected and 196 synchronised multicellular network. In mice, around the time of neurogenesis (between 197 gestational day (GD) 10-15), SCN cells express low amplitude clock genes oscillations with low 198 intercellular synchrony [77]. By the time when the afferent retinal projections reach the SCN 199 and the eyes open (around postnatal day (PND) 12) [78,79], the circuit shows high amplitude 200 clock gene oscillations and high intercellular synchrony (Figure 2). In mice and other rodents, 201 neurons of the core and the shell of the SCN seem to have different ontogeny timing [80]. 202 Cells born around GD12 are mainly confined to the core, whereas cells produced later, around 203 GD 13-14, form the shell and distribute to the posterior and anterior areas of the SCN [80]. To 204 the best of our knowledge, no functional link of this differential ontogenic timing has been 205 described so far. The importance of paracrine signaling promoting the network level 206 properties of the SCN circuit was demonstrated in foundational studies showing that in SCN207 lesioned rodents the systemic rhythmicity can be restored by transplanting a graft containing 208 8 SCN [16], even if the graft is wrapped in a mesh blocking efferent outgrowth [15]. Interestingly, 209 the capacity of the graft to rescue the rhythmicity strongly depends on the age of the donor 210 (grafts of hamster pups older than PND7 lose the ability to rescue circadian rhythmicity) [81]. 211 Furthermore, experiments done with organotypic cultures have shown that the way in which 212 rhythms are coordinated changes along development depending on Vip and Avp signaling, but 213 independently of the presence of key members of the molecular clock machinery [82,83]. This 214 was demonstrated by culturing neonatal SCN tissue slices obtained from both wild-type and 215 Cry1/2 double knockout mice, showing the ability of both to entrain rhythmicity in an 216 arrhythmic adult SCN from Cry1/2 double knockout [82]. Overall, these data suggest that the 217 contribution of paracrine and synaptic signaling might change along SCN development. 218 Moreover, GABA-ergic circuits, such as the SCN, might be impacted by the GABA excitatory219 inhibitory switch, as the circuit matures. In rodents, the switch relies on changes in the 220 chloride (Cl-) flow dependent on the activation of Cl--permeable GABAa receptors [84]. During 221 development, the Na-K-2Cl cotransporter isoform 1 (NKCC1, promoting Claccumulation) 222 downregulates, while the K-Cl cotransporter isoform 2 (KCC2, promoting Clextrusion) 223 upregulates, thus, switching from an excitatory to an inhibitory circuit [84]. Perturbations in 224 the timing when this switch takes place have been proposed as a potential cause of circuit 225 misfunction in the context of conditions such as Down syndrome [85,86], epilepsy [87], autism 226 [88], Rett syndrome [89–91], fragile X syndrome [92,93], 22q11.2 microdeletion syndrome 227 [94], schizophrenia [95], Huntington’s disease [96,97] and behavioral disorders associated 228 with an early adverse environment [98,99]. However, to our knowledge, this possibility has 229 not been explored in the context of SCN development. In the adult mouse SCN, the GABAergic 230 excitation/inhibition ratio (E/I ratio) increases in SCN neurons exposed to long day 231 photoperiod compared to short day photoperiod, indicating that a certain degree of plasticity 232 remains even when the circuit is fully mature [100,101]. 233 Furthermore, the fact that the function of the adult circadian circuit in the mouse depends on 234 neuron-astrocyte interaction suggests that the ontology might be influenced by (or involve) 235 proliferation and functional maturation of astrocytes. Astrocytes are organized in structurally 236 non-overlapping domains [102], being able to integrate information at multiple scales within 237 a neuronal circuit [103] and even to respond to signals derived from the environment 238 (reviewed in [104]). Gap junctions-mediated communication ([24,105,106], reviewed in 239 9 [107]), glutamate release [25] and GABA uptake [28,108] have been found to be essential for 240 the time-keeping role of astrocytes in the adult SCN. Astrocytes can actively support long241 range molecular clock synchronization of segregated neuronal populations, for which gap 242 junctions-mediated communication is required [103]. Moreover, the local proliferation and 243 expansion of astrocytes occurring during perinatal development is regarded as the major 244 developmental origin of astrocytes in the mammalian cerebral cortex and might happen also 245 in the hypothalamus [109,110]. The current view is that astrocytes are modulators of neuronal 246 communication and close partners of neurons for regulating complex circuit level functions 247 and behavior such as sleep ([111,112], reviewed by [113]). Therefore, it is likely that astrocytes 248 are as heterogeneous as neurons in phenotype and function ([114], reviewed in [115]). Since 249 astrocytes and the circadian system develop simultaneously during the perinatal period in 250 mice, it is possible that astrocytes contribute at various levels to the gain of robustness of the 251 circadian circuit [116,117]. Indeed, mathematical models suggest a positive correlation 252 between the density of astrocyte-neuron connectivity and the coordination and amplitude of 253 rhythms [118]. Much work remains to be done in order to understand the interrelationships 254 between the co-occurring events of SCN cell ontogeny, the variable contribution of paracrine 255 signals, the GABA switch, and the proliferation of astrocytes. 256 257 The maturation of SCN neuronal afferent and efferent connections 258 The integration of time-giving signals and the coordination of circadian rhythms at the 259 systemic level, requires the afferent connections to the SCN and efferent connections from it 260 to become mature. In mice, the maturation of the neuronal connection between the retina 261 and the SCN through the retino-hypothalamic tract (RHT), might be one of the main 262 milestones of the master clock’s development. The melanopsin containing intrinsic 263 photosensitive retinal ganglion cells (ipRGCs) convey photic input to the SCN. Interestingly, in 264 mice, these neurons are born and mature throughout a much wider window than SCN 265 neurons, i.e., during GDs 11-18. IpRGCs are glutamatergic neurons that are already able to 266 activate SCN neurons by PNDs 0-1, and complete their innervation around PND7 ([78, 79], 267 reviewed in [39]). These events, are taking place before eyes open (which happens at about 268 PND12 in the mouse) [79]. Interestingly, it has been suggested that the postnatal development 269 of SCN astrocytes might be promoted by the RHT innervation [116]. Indeed, in hamsters, 270 16 behavior in adulthood. Sci Rep, 12, 7366 463 43. Salazar E.R. et al. (2018) Gestational chronodisruption leads to persistent changes in the 464 rat fetal and adult adrenal clock and function. J. Physiol. 596, 5839–5857 465 44. Smarr, B.L. et al. (2017) Maternal and Early-Life Circadian Disruption Have Long-Lasting 466 Negative Consequences on Offspring Development and Adult Behavior in Mice. Sci. Rep. 7, 467 3326 468 45. Varcoe, T.L. et al. (2016) The impact of prenatal circadian rhythm disruption on 469 pregnancy outcomes and long-term metabolic health of mice progeny. Chronobiol. Int. 33, 470 1171–1181 471 46. Jørgensen, J.T. et al. (2021) Shift work and incidence of psychiatric disorders: The Danish 472 Nurse Cohort study. J. Psychiatr. Res. 139, 132–138 473 47. Logan, R.W. and McClung, C.A. (2019) Rhythms of life: circadian disruption and brain 474 disorders across the lifespan. Nat. Rev. Neurosci. 20, 49–65 475 48. Yagita, K. et al. (2010) Development of the circadian oscillator during differentiation of 476 mouse embryonic stem cells in vitro. Proc. Natl. Acad. Sci. U. S. A. 107, 3846–3851 477 49. Dolatshad, H. et al. (2010) Differential expression of the circadian clock in maternal and 478 embryonic tissues of mice. PLoS One 5, 9855 479 50. Umemura, Y. and Yagita, K. et al. (2020) Development of the Circadian Core Machinery 480 in Mammals. J. Mol. Biol. 432, 3611–3617 481 51. Umemura, Y. et al. (2017) Involvement of posttranscriptional regulation of Clock in the 482 emergence of circadian clock oscillation during mouse development. Proc. Natl. Acad. Sci. U. 483 S. A. 114, E7479–E7488 484 52. Fustin, J.M. et al. (2013) RNA-methylation-dependent RNA processing controls the speed 485 of the circadian clock. Cell, 155, 793–806 486 53. Kojima, S. et al. (2012) Circadian control of mRNA polyadenylation dynamics regulates 487 rhythmic protein expression. Genes Dev. 26, 2724–2736 488 54. Umemura, Y. et al. (2014) Transcriptional program of Kpna2/Importin-α2 regulates 489 cellular differentiation-coupled circadian clock development in mammalian cells. Proc. Natl. 490 Acad. Sci. U. S. A. 111, E5039-5048 491 55. Lee, H. et al. (2009) Essential roles of CKIdelta and CKIepsilon in the mammalian 492 circadian clock. Proc. Natl. Acad. Sci. U. S. A. 106, 21359–21364 493 56. Anna, G. and Kannan, N.N. (2021) Post-transcriptional modulators and mediators of the 494 circadian clock. Chronobiol. Int. 38, 1244–1261 495 57. Hsu, C.L. et al. (2018) Analysis of experience-regulated transcriptome and imprintome 496 during critical periods of mouse visual system development reveals spatiotemporal 497 dynamics. Hum. Mol. Genet. 27, 1039–1054 498 17 58. Pons-Espinal, M. et al. (2017) Synergic Functions of miRNAs Determine Neuronal Fate of 499 Adult Neural Stem Cells. Stem Cell Rep. 8, 1046–1061 500 59. Pons-Espinal, M. et al. (2019) MiR-135a-5p Is Critical for Exercise-Induced Adult 501 Neurogenesis. Stem Cell Rep. 12, 1298–1312 502 60. Laufer, B.I. and Singh, S.M. (2012) A macro role for imprinted clusters of microRNAs in 503 the brain. Microrna 1, 59–64 504 61. Freschi, A. et al. (2018) Tissue-specific and mosaic imprinting defects underlie opposite 505 congenital growth disorders in mice. PLoS Genet.14, e1007243 506 62. Landgraf, D. et al. (2015) Embryonic development and maternal regulation of murine 507 circadian clock function. Chronobiol. Int. 32, 416–427 508 63. Čečmanová, V. et al. (2019) Development and Entrainment of the Fetal Clock in the 509 Suprachiasmatic Nuclei: The Role of Glucocorticoids. J. Biol. Rhythms 34, 307–322 510 64. Wreschnig, D. et al. (2014) Embryonic development of circadian oscillations in the 511 mouse hypothalamus. J. Biol. Rhythms 29, 299–310 512 65. Shimomura, H. et al. (2001) Differential daily expression of Per1 and Per2 mRNA in the 513 suprachiasmatic nucleus of fetal and early postnatal mice. Eur. J. Neurosci. 13, 687–693 514 66. Davis, F.C. and Mannion, J. (1988) Entrainment of hamster pup circadian rhythms by 515 prenatal melatonin injections to the mother. Am. J. Physiol. 255, 439-448 516 67. Viswanathan, N. and Davis, F.C. (1997) Single prenatal injections of melatonin or the D1517 dopamine receptor agonist SKF 38393 to pregnant hamsters sets the offsprings’ circadian 518 rhythms to phases 180 degrees apart. J. Comp. Physiol. A 180, 339–346 519 68. Viswanathan, N. et al. (1994) Entrainment of the fetal hamster circadian pacemaker by 520 prenatal injections of the dopamine agonist SKF 38393. J. Neurosci. 14, 5393–5398 521 69. Busada, J.T. and Cidlowski J.A. (2017) Mechanisms of Glucocorticoid Action During 522 Development. Curr. Top Dev. Biol. 125, 147–170 523 70. Lorelli, C. J. et al. (2011) Resilience of circadian pacemaker development in hamsters. J. 524 Biol. Rhythms 26, 221–229 525 71. Bates, K. and Herzog, E.D. (2020) Maternal-Fetal Circadian Communication During 526 Pregnancy. Front. Endocrinol. 11, 198 527 72. Reppert, S.M. and Schwartz, W. J. (1986) Maternal suprachiasmatic nuclei are necessary 528 for maternal coordination of the developing circadian system. J. Neurosci. 6, 2724–2729 529 73. Jud, C. and Albrecht, U. (2006) Circadian rhythms in murine pups develop in absence of a 530 functional maternal circadian clock. J. Biol. Rhythms 21, 149–154 531 74. Davis, F.C. and Gorski, R. A. (1988) Development of hamster circadian rhythms: role of 532 the maternal suprachiasmatic nucleus. J. Comp. Physiol. A, 162, 601–610 533 18 75. Mendez, N. et al. (2022) Maternal melatonin treatment rescues endocrine, 534 inflammatory, and transcriptional deregulation in the adult rat female offspring from 535 gestational chronodisruption. Front. Neurosci. 16, 1039977 536 76. Bedont, J.L. et al. (2015) Patterning, specification, and differentiation in the developing 537 hypothalamus. Wiley Interdiscip. Rev. Dev. Biol. 4, 445–468 538 77. Carmona-Alcocer, V. et al. (2018) Ontogeny of Circadian Rhythms and Synchrony in the 539 Suprachiasmatic Nucleus. J. Neurosci. 38, 1326–1334 540 78. Sekaran, S. et al. (2005) Melanopsin-dependent photoreception provides earliest light 541 detection in the mammalian retina. Curr. Biol. 15, 1099–1107 542 79. Hoy, J. L. and Niell, C.M. (2015) Layer-specific refinement of visual cortex function after 543 eye opening in the awake mouse. J. Neurosci. 35, 3370–3383. 544 80. Kabrita, C.S. and Davis, F.C. (2008) Development of the mouse suprachiasmatic nucleus: 545 determination of time of cell origin and spatial arrangements within the nucleus. Brain Res. 546 1195, 20–27 547 81. Romero, M.T. et al. (1993) Age of donor influences ability of suprachiasmatic nucleus 548 grafts to restore circadian rhythmicity. Brain Res. Dev. Brain Res. 71, 45–52 549 82. Ono, D. et al. (2013) Cryptochromes are critical for the development of coherent 550 circadian rhythms in the mouse suprachiasmatic nucleus. Nat. Commun. 4, 1666 551 83. Ono, D. et al. (2016) Differential roles of AVP and VIP signaling in the postnatal changes 552 of neural networks for coherent circadian rhythms in the SCN. Sci. Adv. 2, e1600960 553 84. Ganguly, K. et al. (2001) GABA itself promotes the developmental switch of neuronal 554 GABAergic responses from excitation to inhibition. Cell, 105, 521–532 555 85. Deidda, G. et al. (2015) Reversing excitatory GABAAR signaling restores synaptic 556 plasticity and memory in a mouse model of Down syndrome. Nat. Med. 21, 318–326 557 86. Lysenko, L.V. et al. (2018) Developmental excitatory-to-inhibitory GABA polarity switch is 558 delayed in Ts65Dn mice, a genetic model of Down syndrome. Neurobiol. Dis. 115, 1–8 559 87. Kahle, K.T. et al. (2014) Genetically encoded impairment of neuronal KCC2 cotransporter 560 function in human idiopathic generalized epilepsy. EMBO Rep. 15, 766–774 561 88. Tyzio, R. et al. (2014) Oxytocin-mediated GABA inhibition during delivery attenuates 562 autism pathogenesis in rodent offspring. Science 343, 675–679 563 89. Lozovaya, N. et al. (2019) Early alterations in a mouse model of Rett syndrome: the 564 GABA developmental shift is abolished at birth. Sci. Rep. 9, 9276 565 90. Tang, X. et al. (2016) KCC2 rescues functional deficits in human neurons derived from 566 patients with Rett syndrome. Proc. Natl. Acad. Sci. U. S. A. 113, 751–756 567 91. Banerjee, A. et al. (2016) Jointly reduced inhibition and excitation underlies circuit-wide 568 changes in cortical processing in Rett syndrome. Proc. Natl. Acad. Sci. U. S. A. 113, E7287– 569 19 E7296 570 92. He, Q. et al. (2019) Critical period inhibition of NKCC1 rectifies synapse plasticity in the 571 somatosensory cortex and restores adult tactile response maps in fragile X mice. mol. 572 Psychiatry 24, 1732–1747 573 93. He, Q. et al. (2014) The developmental switch in GABA polarity is delayed in fragile X 574 mice. J. Neurosci. 34, 446–450 575 94. Amin, H. et al. (2017) Developmental excitatory-to-inhibitory GABA-polarity switch is 576 disrupted in 22q11.2 deletion syndrome: a potential target for clinical therapeutics. Sci. Rep. 577 7, 15752 578 95. Kim, H.R. et al. (2021). Depolarizing GABAA current in the prefrontal cortex is linked 579 with cognitive impairment in a mouse model relevant for schizophrenia. Sci. Adv. 7, 5032 580 96. Dargaei, Z. et al. (2018) Restoring GABAergic inhibition rescues memory deficits in a 581 Huntington’s disease mouse model. Proc. Natl. Acad. Sci. U. S. A. 115, E1618–E1626 582 97. Hsu, Y.T. et al. (2019) Enhanced Na+ -K+ -2Clcotransporter 1 underlies motor 583 dysfunction in huntington’s disease. Mov. Disord. 34, 845–857 584 98. Karst, H. et al. (2023). Acceleration of GABA-switch after early life stress changes mouse 585 prefrontal glutamatergic transmission. Neuropharmacology 234, 109543 586 99. Kang, E. et al. (2019) Interplay between a Mental Disorder Risk Gene and Developmental 587 Polarity Switch of GABA Action Leads to Excitation-Inhibition Imbalance. Cell Rep. 28, 1419588 1428 589 100. Farajnia, S. et al. (2014) Seasonal induction of GABAergic excitation in the central 590 mammalian clock. Proc. Natl. Acad. Sci. U. S. A. 111, 9627–9632 591 101. Eick A. K. et al. (2022) The opposing chloride cotransporters KCC and NKCC control 592 locomotor activity in constant light and during long days. Curr. Biol. 32, 1420-1428 593 102. Ogata, K. and Kosaka, T. (2002) Structural and quantitative analysis of astrocytes in the 594 mouse hippocampus. Neuroscience 113, 221–233 595 103. Giantomasi, L. et al. (2023) Astrocytes actively support long-range molecular clock 596 synchronization of segregated neuronal populations. Sci. Rep. 13, 4815 597 104. Araque, A. et al. (2014) Gliotransmitters travel in time and space. Neuron 81, 728–739 598 105. Shinohara, K. et al. (2000). Effects of gap junction blocker on vasopressin and 599 vasoactive intestinal polypeptide rhythms in the rat suprachiasmatic nucleus in vitro. 600 Neurosci. Res. 38, 43–47 601 106. Clasadonte, J. et al. (2017) Connexin 43-Mediated Astroglial Metabolic Networks 602 Contribute to the Regulation of the Sleep-Wake Cycle. Neuron, 95, 1365-1380 603 107. Bennett, M.V.L. et al. (2003) New roles for astrocytes: gap junction hemichannels have 604 something to communicate. Trends Neurosci. 26, 610–617 605 20 108. Patton, A.P. et al. (2023) Astrocytic control of extracellular GABA drives circadian 606 timekeeping in the suprachiasmatic nucleus. Proc. Natl. Acad. Sci. U. S. A. 120, e2301330120 607 109. Bandeira, F. et al (2009) Changing numbers of neuronal and non-neuronal cells underlie 608 postnatal brain growth in the rat. Proc. Natl. Acad. Sci. U. S. A. 106, 14108–14113 609 110. Ge, W.P. et al. (2012) Local generation of glia is a major astrocyte source in postnatal 610 cortex. Nature 484, 376–380 611 111. Bojarskaite, L. et al. (2020) Astrocytic Ca2+ signaling is reduced during sleep and is 612 involved in the regulation of slow wave sleep. Nat. Commun. 11, 3240 613 112. Ingiosi, A.M. et al. (2020) A Role for Astroglial Calcium in Mammalian Sleep and Sleep 614 Regulation. Curr. Biol. 30, 4373-4383. 615 113. Oliveira, J.F. et al. (2015). Do stars govern our actions? Astrocyte involvement in rodent 616 behavior. Trends Neurosci. 38, 535–549. 617 114. Götz, S. et al. (2021). Heterogeneity of astrocytes: Electrophysiological properties of 618 juxtavascular astrocytes before and after brain injury. Glia 69, 346–361 619 115. Ben Haim, L. and Rowitch, D.H. (2017) Functional diversity of astrocytes in neural 620 circuit regulation. Nat. Rev. Neurosci. 18, 31–41 621 116. Munekawa, K. et al. (2000) Development of astroglial elements in the suprachiasmatic 622 nucleus of the rat: with special reference to the involvement of the optic nerve. Exp. Neurol. 623 166, 44–51 624 117. Lavialle, M. et al. (2001) Modifications of retinal afferent activity induce changes in 625 astroglial plasticity in the hamster circadian clock. Glia 34, 88–100 626 118. Sueviriyapan, N. et al. (2020) Astrocytic Modulation of Neuronal Activity in the 627 Suprachiasmatic Nucleus: Insights from Mathematical Modeling. J. Biol. Rhythms, 35, 287– 628 301 629 119. Tosini, G. and Menaker, M. (1996) Circadian rhythms in cultured mammalian retina. 630 Science 272, 419–421 10.1126/science.272.5260.419. 631 120. Ruan, G.X. et al. (2006). Circadian organization of the mammalian retina. Proc. Natl. 632 Acad. Sci. U. S. A. 103, 9703–9708 633 121. Sakamoto, K. et al. (2000). Two circadian oscillatory mechanisms in the mammalian 634 retina. Neuroreport. 11, 3995–3997 635 122. Riccitelli, S. et al. (2022) Glial Bmal1 role in mammalian retina daily changes. Sci. Rep. 636 12, 21561 637 123. Jidigam, V.K. et al. (2022) Neuronal Bmal1 regulates retinal angiogenesis and 638 neovascularization in mice. Commun. Biol. 5, 792 639 124. Ferguson, S.A. and Kennaway, D. J. (2000). The ontogeny of induction of c-fos in the rat 640 SCN by a 5-HT(2A/2C) agonist. Brain Res. Dev. Brain Res. 121, 229–231 641 21 125. Ueda, S. et al. (1996) Alteration of serotonergic innervation in the suprachiasmatic 642 nucleus of the rat following removal of input fibers from retina and lateral geniculate 643 nucleus. Neurosci. Lett. 211, 97–100 644 126. Botchkina, G.I. and Morin, L. P. (1995) Specialized neuronal and glial contributions to 645 development of the hamster lateral geniculate complex and circadian visual system. J. 646 Neurosci. 15, 190–201 647 127. Ferguson, S.A. et al. (2000). Prenatal exposure to the dopamine agonist SKF-38393 648 disrupts the timing of the initial response of the suprachiasmatic nucleus to light. Brain Res. 649 858, 284–289 650 128. Buijs, R. M. et al. (1993) Projections of the suprachiasmatic nucleus to stress-related 651 areas in the rat hypothalamus: a light and electron microscopic study. J. Comp. Neurol. 335, 652 42–54 653 129. Buijs, F.N. et al. (2017) Suprachiasmatic Nucleus Interaction with the Arcuate Nucleus; 654 Essential for Organizing Physiological Rhythms. eNeuro 4, ENEURO.0028-17.2017 655 130. Acosta-Galvan, G. et al. (2011) Interaction between hypothalamic dorsomedial nucleus 656 and the suprachiasmatic nucleus determines intensity of food anticipatory behavior. Proc. 657 Natl. Acad. Sci. U. S. A. 108, 5813–5818 658 131. Deurveilher, S. et al. (2002). Indirect projections from the suprachiasmatic nucleus to 659 the ventrolateral preoptic nucleus: a dual tract-tracing study in rat. Eur. J. Neurosci. 16, 660 1195–1213 661 132. Deurveilher, S. and Semba, K. (2003) Indirect projections from the suprachiasmatic 662 nucleus to the median preoptic nucleus in rat. Brain Res. 987, 100–106 663 133. Ma, M.A. and Morrison, E.H. (2023) Neuroanatomy, Nucleus Suprachiasmatic in 664 StatPearls, Treasure Island (FL): StatPearls Publishing 665 134. Yao, Y. et al. (2023) Vasculature of the Suprachiasmatic Nucleus: Pathways for 666 Diffusible Output Signals. J. Biol. Rhythms 38, 571–585 667 135. Yao, Y. et al. (2021) Identification of the suprachiasmatic nucleus venous portal system 668 in the mammalian brain. Nat. Commun.12, 5643 669 136. Mendez, N. et al. (2016) Gestational Chronodisruption Impairs Circadian Physiology in 670 Rat Male Offspring, Increasing the Risk of Chronic Disease. Endocrinology 157, 4654–4668 671 137. An, K. et al. (2020) A circadian rhythm-gated subcortical pathway for nighttime-light672 induced depressive-like behaviors in mice. Nat. Neurosci. 23, 869–880. 673 138. Kaur, S. et al. (2020) Circadian rhythm and its association with birth and infant 674 outcomes: research protocol of a prospective cohort study. BMC Pregnancy Childbirth, 20, 675 96 676 677 22 Figure Legends 678 Figure 1: Multiscale organization of the circadian system. The figure represents the three 679 levels of organization of the circadian system (molecular, tissue/circuit and systemic), and 680 examples of the three components: input, oscillator and output that are essential for receiving 681 temporal cues, integrating them and synchronizing downstream physiological processes. 682 Neurons are represented as yellow, green and blue circles and astrocytes as orange stars. 683 Abbreviations: Avp: vasopressin, BMAL1: brain and muscle aryl hydrocarbon receptor nuclear 684 translocator-like 1, CLOCK: circadian locomotor output cycles kaput, CRY1/2: Cryptochrome 685 1/2, GABA: gamma-aminobutyric acid, GCs: glucocorticoids, Glu: glutamate, MEL: melatonin, 686 Nms: Neuromedin S, OC: optic chiasm, Pacap: pituitary adenylate-cyclase-activating 687 polypeptide, PER1/2: Period 1/2, RHT: retinohypothalamic tract, SCN: suprachiasmatic nuclei, 688 T (°C): Temperature, Vip: Vasoactive intestinal peptide, 3V: third ventricle. 689 690 Figure 2: Timeline of the master clock ontogeny in mice. The figure represents the timeline 691 of the ontogeny of the hypothalamic SCN in mice during the perinatal period. The circuit 692 progresses from a collection of undifferentiated individual cells (grey circles), to a highly 693 differentiated, interconnected and synchronised multicellular network (neurons as green 694 circles and astrocytes as orange stars). In mice, around the time of neurogenesis (between 695 gestational day (GD) 10-15), SCN cells express clock genes with low amplitude oscillations and 696 low intercellular synchrony. By the time when the afferent retinal projections reach the SCN 697 and the eyes open (around postnatal day (PND) 12), the circuit express clock genes with high 698 amplitude oscillations and high intercellular synchrony. The maturation of the SCN as a circuit 699 occurs under the influence of maternal rhythmic hormonal signals during both the preand 700 post-natal periods (i.e.: signals that reach fetal/newborn tissues through placenta/breast 701 milk). Abbreviations: GD: Gestational day, Opn4: Melanopsin, PND: Postnatal day, RHT: retino 702 hypothalamic tract. 703 704 3V Vip Gaba Photic Glu Pacap RHT Core Shell Avp Nms Gaba Input Oscillator Output OC Circuit Systemic Glu Pacap RHT Photic Food T (°C) Nonphotic SCN Extra-SCN Peripheral clocks Avp Gaba MEL GCs Molecular Hormones MEL GCs Hormones 24 h BMAL1 CLOCK PER1/2 CRY1/2 Autonomic nervous system Clockcontrolled genes (CCG) GD10-15 Neurogenesis GD13 Clock gene expression PND12 RHT maturation GD15 Opn4 expression GD0 GD10 GD16 BIRTH PND10 PND30 Neurogenesis Gliogenesis RHT Maternal signals LIGHT GD17-PND0 Neuro-gliogenic switch GD16 Rhythmic gene expression