Circadian Biology: The Supporting Cast Takes On a Starring Role

Abstract

Brain circuits are generally thought to consist solely of neurons communicating with other neurons. In Drosophila, glia-to-neuron signaling has now been shown to be critical to the function of the circadian circuit.

Glia have been regarded by many as the nervous system equivalent of extras in a movie crowd scene: they are required to set the mood but do not drive the plot. Well-documented roles for glia in vertebrates are numerous. They include metabolic and trophic support, developmental pathfinding, scavenging of neurotransmitters and cell debris, electrical insulation of axons, as well as the chemical isolation of cellular compartments. Glia have not, however, had much of a chance at the spotlight in terms of behavior. In Drosophila, glia have also been shown to carry out critical support functions, but the genetic tools available have now allowed researchers to begin to examine their roles in directly shaping nervous system output. In a recent issue of Current Biology, Ng and coworkers [1] demonstrate that calcium-dependent glia-to-neuron signaling regulates the circadian locomotor rhythm.

The idea that glia have a role in circadian rhythms has been around since the identification of the molecular clock in Drosophila. The cloning of the first circadian gene, period, allowed researchers to establish a foothold in the cellular clock. Looking for cells that expressed PERIOD in a cyclic manner, they hoped to identify a neural circuit for timekeeping. Surprisingly, the cells they found included a substantial number of glia [2,3]. This cyclic expression of clock genes in glia is not just an oddity of flies. PERIOD cycling has also been seen in astrocytes of the mammalian suprachiasmatic nucleus [4], suggesting that glia may be a component of all animal clocks. Consistent with this, mosaic studies in flies found that expression of PERIOD in ventral brain glia, without any neuronal expression, was actually sufficient to support weak rhythms [5]. For many years, however, the role of glia in rhythms remained relatively unexplored.

Glia resurfaced in the clock when Suh and Jackson [6] showed that the circadian function of the ebony gene, which regulates locomotor output downstream of the clock, resided in glia. EBONY is an N-β-alanyl-biogenic amine synthetase; it attaches β-alanine to a variety of neurotransmitter substrates. The proximity of ebony-expressing glia to dopaminergic neurons, and the known function of dopamine in arousal and locomotor activity, led the authors to speculate that N-β-alanyl-dopamine (NBAD) might regulate dopaminergic function as a gliotransmitter. Direct signaling to neurons from glia had not been previously shown in Drosophila.

Ng and colleagues [1] tested this idea that active, calcium-dependent signaling processes are required in astrocytic glia to drive circadian locomotor behavior. Using genetic tools available in Drosophila to spatially and temporally limit transgene expression, the authors showed that glial disruption of membrane potential by expression of a constitutively open sodium channel, misregulation of glial calcium by RNA interference (RNAi) knockdown of sarco/endoplasmic reticulum Ca²⁺-ATPase and blockade of vesicle trafficking by a dominant negative dynamin each can cause arrhythmicity in constant conditions (dark:dark, DD) after circadian entrainment in a normal light:dark cycle. The manipulations used were reversible and designed to limit molecular disruptions to defined times in adulthood. This is particularly important since the ability of animals to regain normal rhythms following a period of signal disruption demonstrated that there was no permanent damage to glia or the locomotor system.

Interestingly, the phase of the rhythm after restoration of normal cellular function in DD was the same as that of the pre-disruption rhythm. This implied that the central clock was still ‘ticking’ even though the animals were behaviorally arrhythmic. Consistent with this, circadian cycling of the abundance and localization of PERIOD and PDP1ε proteins in lateral clock neurons was normal on day 2 of DD while signaling was disrupted. This makes a strong case that the process that was being affected by these glial manipulations was one that was critical to the output of the circadian clock, not to the functioning of the central clock itself.

Given that alterations in glial activity did not affect molecular clock function in the canonical neuronal clock circuit, an obvious question to ask was whether cycling of clock proteins in the glia themselves was at the heart of their role in locomotor rhythmicity. To address this, the authors used glial-specific RNAi transgenes to knock down levels of PERIOD and another circadian protein, CRYPTOCHROME. Neither manipulation affected motor rhythms, indicating that having an oscillating molecular clock within glia is not necessary for locomotor rhythms.

While the transcriptional machinery that makes up the molecular clock appeared to be normal in clock neurons, one of the peptide transmitters of the ventral lateral neurons (LNvs) was significantly altered in abundance when glia were disrupted. Pigment Dispersing Factor (PDF) is essential for synchronization of the clock circuit in DD [7,8] and influences locomotor activity [9]. Previous work had shown that levels of PDF in the dorsal processes of the small LNvs cycled over the course of the day, being highest in the morning [10]. In animals with disrupted glial vesicle trafficking, PDF immunostaining in dorsal processes, but not cell bodies, was significantly reduced after two or three days in DD with apparent dampening of cycling. This was reversible; when glial function was restored, PDF levels in processes went back up, indicating that there had been no damage to LNvs. The normal high levels of PDF present in cell bodies in both control and experimental flies suggested that glial signaling promotes trafficking of PDF to processes as opposed to governing its synthesis.

These data would be consistent with high amplitude cycling of PDF levels being an important feature of its function in locomotion. However, an earlier study in which cycling of PDF was completely blocked by over-expression of a GFP-tagged neuropeptide (GFP–ANF) found no changes in locomotor rhythmicity [11]. The reason for this apparent discrepancy may lie in the difference in absolute levels of PDF in processes after the two manipulations. Blockade of glial vesicle recycling reduced the amount of PDF in the processes, while the overexpression of GFP–ANF blocked cycling but left PDF levels high. Absolute levels of PDF may be more important than cycling for locomotor output.

Another interesting outcome of this study is the clear dissociation of the downstream effects of PDF on clock proteins and locomotor rhythm. In the first few days of DD after disruption of glial vesicle trafficking, animals show locomotor arrhythmicity, but normal expression patterns of clock genes in the core clock circuit. The normal molecular cycling may indicate that even though PDF levels are low, there is some release and that this small amount of PDF is sufficient to keep the clock running. The differential sensitivity implies that the cellular targets of PDF in locomotion and clock organization are likely to be distinct. Whether the low levels of PDF and the very damped cycling are sufficient to maintain the molecular clock for longer than 2 days in DD was not tested, but it would be interesting to investigate this, since total loss of PDF does not cause immediate arrhythmicity [8].

In aggregate, these data tell us that there are signaling processes in glia that are necessary for the expression of clock circuit-driven locomotor rhythms. The involvement of intracellular calcium and vesicle trafficking suggests a straightforward model in which there is active release of some substance (perhaps an EBONY-generated amine conjugate such as NBAD?) that signals to LNvs to promote PDF trafficking and/or release. There are, however, a number of very intriguing open questions. The requirement for vesicle recycling could reflect internalization of some factor or receptor by glia [12] as opposed to release from glia. And, while there are clear effects on LNvs, there may be additional important cellular targets of glial signaling in the clock output circuit. But thanks to this novel and important study, glia get to live every bit player’s dream: they have been pulled from the chorus for a starring role.