|Home | About | Journals | Submit | Contact Us | Français|
The function of sleep is hotly contested. Two recent studies suggest that fly sleep may be required to rescale synapses in the brain.
Sleep is required to maintain cognitive function but the reason is not understood. One model proposes that sleep provides the brain a time in which incoming sensory information is minimal, so memories acquired while awake can be stabilized or consolidated within the relevant neural circuits. How memories are consolidated from a labile short-term form to a more stable long-term representation is not well understood. However, in many species, even brief periods of sleep restriction impair cognitive function and the capacity to consolidate new memories . Electrophysiological recordings in rats  have shown that activity patterns observed during wakeful experience are re-activated in the appropriate circuits during sleep, suggesting that sleep may promote off-line memory consolidation. Furthermore, reactivating a memory with a non-disruptive cue while asleep promotes memory consolidation in humans .
A second, more general model, that is not mutually exclusive to the memory consolidation hypothesis, proposes that sleep maintains cognitive function by promoting brain-wide ‘synaptic homeostasis’. Cirelli and Tononi  proposed that many synapses in the brain are strengthened, or ‘potentiated’, by normal circuit use during wake. Slow-wave neural activity that is specific to sleep globally resets the synapses, returning the brain to a baseline state. Since memories are also believed to be formed by altering synaptic connections between neurons, the mechanism must be sophisticated enough to retain the memory-relevant synaptic changes while downscaling the others. In their most recent work, Gilestro, Tononi and Cirelli  extend their synaptic homeostasis theory from humans  and rats  to the fruit fly, Drosophila melanogaster, which exhibits many of the behavioral hallmarks of mammalian sleep .
Gilestro et al.  examined the expression levels of synaptic marker proteins in the fly brain through the sleep/wake cycle and following sleep deprivation by quantifying immunoblots of brain extract and immunofluorescence signals in the intact fixed brain. They analyzed levels of the synaptic active zone protein bruchpilot (BRP), in addition to the synaptically localized discs-large (DLG), synapsin, syntaxin, and cysteine string proteins. In unperturbed flies, BRP levels were increased following six hours of spontaneous wake compared with six hours of spontaneous sleep.
To examine whether expression levels were further increased with forced waking, flies were sleep-deprived by mechanical stimulation, or by a novel method where a guest fly joined the host fly in the recording chamber. Both deprivation methods resulted in sleep loss and increased expression of each synaptic protein that was assayed. Longer periods of sleep deprivation by mechanical stimulation generally correlated with higher expression levels of each protein. Importantly, the authors verified that BRP accumulation was a result of sleep loss, and not mechanical stimulation. The flies that were most sleep-deprived following the same period of stimulation showed the highest BRP expression levels. The wake-induced increase of BRP is apparently reversed by sleep. BRP levels gradually decreased during the period the flies were asleep. The BRP decrease required sleep, and not simply passage of time, because it was prevented by keeping the flies awake.
These correlative data suggest that synaptic protein levels throughout the brain of the fly increase with wake and decrease with sleep. One might expect that specific areas of the brain with higher levels of activity during wake would be most likely to show sleep/wake-state-regulated changes in expression level. However, quantifying BRP intensity level by immunostaining and microscopy suggested a general increase throughout the brain. Gilestro et al.  quantified BRP labeling intensity in three regions of the central brain following sleep deprivation. BRP staining increased in the mushroom bodies (MB; Figure 1), an area necessary for olfactory memory  that also regulates sleep/wake amount [10,11]. Enigmatically, electrophysiological recordings of MB neurons in a restrained fly indicate that they have a very low spontaneous firing rate . BRP expression level was also elevated by wake in the ellipsoid body, a region implicated in locomotor activity control . This is not unexpected, because flies move frequently while awake. Lastly, BRP intensity increased in the antennal lobes, which contain the processes of olfactory sensory neurons, local neurons and projection neurons . Presumably, the other primary sensory regions — the visual area of the optic lobes and the gustatory area of the subesophageal ganglion — also exhibit significant changes in BRP level. It will be important to determine whether increased BRP in these neurons alters the function of the relevant synapses, or whether the excess BRP is ‘junk’ that results from general circuit use that is cleared during sleep.
Nevertheless, the finding that synaptic proteins are elevated by wake and reduced by sleep in flies provides a plausible role for fly sleep in resetting synaptic weight. It will be interesting to test whether synaptic proteins are altered in mutant flies that sleep less, such as the potassium channel mutant shaker , and in what direction. Additionally, it will be important to discern whether mutations in, or over-expression of, synaptic proteins alter sleep amount and brain function, such as memory formation and consolidation. Although the reported electrophysiological correlate for fly sleep  does not resemble mammalian slow-waves, it will also be interesting to determine whether such brain activity is required to reduce synaptic protein expression in flies.
Another question arising from the Gilestro study  is whether changes in synaptic protein levels are part of the homeostatic ‘sleep-need’ signal. The timing of sleep is thought to be regulated by the circadian clock, whereas the amount of sleep is controlled by an unknown homeostatic mechanism. The brain monitors ‘sleep-need’ similar to a thermostat that monitors temperature fluctuation and adjusts around a set-point. When deprived of sleep for one night, you sleep longer the following night to compensate. It is unclear how sleep-need is represented in the brain. If state-dependent changes in synaptic proteins constitute part of the molecular sleep-homeostat, then the challenge will be to uncover how this information is monitored and transduced to brain regions that initiate sleep/wake. A subset of circadian pacemaker neurons known as the large ventral lateral neurons (LNv) promote wake in flies [16–19]. In a companion paper, Donlea et al.  overexpressed fluorescent GFP–DLG and GFP–synaptobrevin fusion proteins in the small and large LNv. They found a modest increase in the number of GFP-labeled synapses in the optic lobe (from large LNv projections) following 48 hours of sleep deprivation. They also found an increase in GFP labeling following ‘social enrichment’. In this protocol, flies are either kept in isolation, or are enriched by housing in groups. Social enrichment increases sleep, and is thought to promote plasticity by providing a more stimulating environment for the fly. A hopeful interpretation of these data is that the LNv may monitor changes in synaptic protein expression and transduce the information to modulate sleep amount. It will be important to test whether changes in synaptic proteins within the LNv are a critical part of the sleep homeostat.
If synaptic downscaling turns out to be a critical and conserved function of sleep that allows the brain to remain plastic and/or is a critical part of the sleep homeostat, the powerful genetics of the fly will surely aid discovery of the underlying mechanisms.