Sleep may reverse age-related decline in brain support cells' cleanup function

Sleep is a universal behavior that is critical for diverse aspects of brain function [1]. Chronic sleep disturbances are associated with numerous health consequences including neurodegenerative disease and cognitive decline [2,3]. Neurite damage due to apoptosis, trauma, or genetic factors is a common feature of aging, and the clearance of damaged neurons is thought to be critical for maintenance of proper brain function [4,5]. In the central and peripheral nervous systems, damaged neurites are cleared by Wallerian degeneration, a process where microglia or macrophages and Schwan cells are activated and engulf damaged neurites [6]. Despite the critical role of neurite clearance in maintenance of brain function and protection against neurodegenerative disease, surprisingly little is known about how life-history traits and environment modulate neurite clearance [7].

In flies and mammals, axotomy triggers Wallerian degeneration, where specialized glia engulf damaged neurites [8]. The genetic accessibility of the Drosophila olfactory system, combined with powerful genetic tools that allow for cell-type specific manipulation of gene expression, have made Drosophila a leading model to study injury-induced glial engulfment [9]. In Drosophila, olfactory ensheathing glia surrounds the antennal lobe then engulf the damaged olfactory neurons. Genetic screens in this system have identified numerous novel genetic factors and intercellular signaling pathways required for glial engulfment including Stat92E, Draper, and Insulin Receptor [10–14]. Further, numerous factors have been identified that function within neurons to signal neural injury including loss of the NAD+ synthase Nmat1 and Sarm [9,12,15]. Modeling of Wallerian degeneration in other systems including neurons in the Drosophila wing, larval peripheral neurons, and mouse peripheral nerves confirm the mechanisms regulating Wallerian degeneration in the olfactory system are conserved in other Drosophila and mammalian neurons [16–18].

Glia are critical modulators of sleep duration, sleep-mediated neuronal homeostasis, and clearance of toxic substances during sleep [16–20]. Multiple mechanisms link glia to neurodegenerative disease including a role for clearance of Amyloid β (Aβ) by the glymphatic system and pruning of synapses by microglia in Alzheimer’s disease [21–23]. Recent findings suggest microglia function is inhibited by sleep deprivation in mouse models, raising the possibility that sleep directly impacts neuronal pruning and clearance within the brain [24]. Despite these bidirectional interactions between glia and sleep, surprisingly little is known about the mechanisms through which sleep deprivation impacts glial function, and how this contributes to neurodegenerative disease.

Growing evidence suggests sleep loss contributes to many of the functional deficits associated with aging [25,26]. Sleep loss shortens lifespan and accelerates many factors associated with aging, while quality and duration are reduced in old individuals [26,27]. In Drosophila aging, high calorie diets that accelerate aging, as well as acute sleep loss, all inhibit draper expression, resulting in reduced glial plasticity and clearance of injured neurites [19,28–30]. While the effects of age and sleep-related factors have largely been studied independently, both result in disrupted sleep, raising the possibility that reduced sleep quality contributes to age-related decline in glial function [31–33]. Examining the role of sleep on aging-associated loss of glial engulfment has potential to guide experimental approaches and therapeutic treatments for age-related decline in brain function.

Here, we examine the relationship between sleep and age-dependent changes on glial elimination of damaged neurites. Both pharmacological and genetic induction of sleep restore glial plasticity and Draper induction in aged flies. Further, the sleep-inducing effects of neural injury are absent in aged flies, revealing bidirectional interactions between sleep and aging. Together, these findings suggest sleep loss plays a critical role in age-dependent loss of glial cell function.

Sleep duration and quality are reduced in aged flies [27,33–36]. To identify the specific timepoints when sleep duration and quality are reduced and to replicate previous findings that link aging to sleep loss, we compared sleep in flies aged 5, 25 and 40 days. There was a significant reduction in sleep in flies aged 25 days, and a greater loss at 40 days (Fig 1A and 1B). These changes in sleep duration are associated with reduced bout length and increased bout number, indicating reduced sleep quality (S1A-S1F Fig). Therefore, these timepoints were used to investigate the effects of age-related sleep loss on Wallerian degeneration.

Previous findings suggest that glial plasticity and Draper induction are impaired in aged flies [29]. To further examine the effects of aging on glial function we quantified Draper induction, neurite clearance and glial membrane infiltration in young and old flies. We ablated the antennae of flies aged 5, 25, and 40 days, expressing mCD8:GFP under control of the pan-glial driver Repo (S1G Fig) [9]. Glial membranes labeled with GFP robustly extend into the antennal lobes 24 hours following antennal ablation in young 5-day-old flies (S1H Fig). In contrast, glial membrane infiltration of the antennal lobes is diminished in flies injured at 25 days old, with an additional reduction observed in 40-day-old injured flies (S1I Fig). To confirm that the phenotypes observed were due to the recruitment of olfactory ensheathing glia, we selectively labeled these glia with mz0709-Gal4 and quantified antennal lobe infiltration following antennal ablation (Fig 1C) [37]. Similar to results obtained with Repo-Gal4, the extension of glial processes into the antennal lobe following injury was reduced at 25 days and further diminished in ablated 40-day-old flies, while there were no age-dependent differences were noted in intact flies (Fig 1D and 1E). These findings confirm that the extension of ensheathing glial processes into the antennal lobe is impaired in aged flies.

Neuronal injury results in upregulation of the phagocytic protein Draper within olfactory ensheathing glia, leading to clearance of the damaged neurites [9]. To confirm that age-related deficits in glial plasticity are associated with the loss of Draper and reduced clearance of damaged neurites, we genetically labeled the olfactory receptor neurons (ORNs) using the Or22a::GFP transgene, which selectively labels ORNs from the antennae while excluding those from the maxillary palp [38]. The antennae of 5- or 40-day-old flies were ablated and the brains were subsequently immunostained for Draper 24 hours following injury. There were no differences in Draper levels in 5- and 40-day-old intact flies (Fig 1F-1G). Antennal ablation robustly induced Draper within the antennal lobes of 5-day-old flies, but not 40-day-old flies, confirming Draper induction is reduced in aged flies (Fig 1F-1H). Conversely, in 5-day-old flies, there was a significant reduction in Or22a::GFP signal, indicating robust glial elimination of damaged neurites (Fig 1F). Or22a::GFP signal was significantly higher in ablated 40-day-old flies compared to that in 5-day-old flies, revealing reduced clearance in aged flies (Fig 1F-1G and 1I). Together, these findings confirm that glial recruitment, Draper induction, and neurite engulfment in response to injury are reduced in 40-day-old flies.

Injury and immune activation are associated with increased sleep in Drosophila [18,33]. In flies, sleep is increased immediately following antennal ablation, providing a system to examine the effects of aging on neural injury response [18]. To examine the effects of aging on sleep following injury, we quantified sleep immediately following injury in 5- and 40-day-old flies (Fig 2A). In agreement with previous findings, 5-day-old antennal ablated flies slept significantly more than intact controls (Fig 2B) [18]. Conversely, there was no increase in daytime sleep following ablation in 40-day-old ablated flies (Figs 2C, 2D, and S2A). In 40-day-old flies there was an increase in nighttime sleep 12 hours from injury, suggesting a reduced and delayed response (S2A Fig). At day two following ablation sleep did not differ from controls for both young and old flies, revealing the immediate effects of injury do not result in long-lasting changes (Fig 2B and 2C). The increased sleep following injury in 5-day-old flies is due to longer bout lengths, consistent with greater sleep consolidation (Figs 2D, and S2D-S2K).

We applied a Markov model that determines propensity to remain asleep, which serves as an indicator of sleep depth [37]. Following injury, sleep propensity, P(Doze), was elevated and wake propensity, P(Wake), was reduced in 5-day-old flies, but not in 40-day-old flies, supporting the notion that sleep drive is diminished in aged flies (Figs 2E, 2F, and S2B-S2C). Furthermore, ablation induced an increase in average bout length during the day but not at night in 5-day-old flies, whereas no such effect was observed in 40-day-old flies (S2F-S2G and S2J-S2K Fig). Therefore, aging disrupts sleep-induction following injury. These findings support the idea that both reduced levels of basal sleep, and impaired sleep induction following injury contribute to age-related loss of glial activation and Wallerian degeneration.

Draper is required for the clearance of damaged neurites, but its role in sleep regulation has not been investigated. To determine whether Draper also regulates sleep, we selectively knocked down draper in all glia using Repo-GAL4, or specifically in ensheathing glia using mz0709-GAL4 or R56C01-GAL4. In all three cases, knockdown flies slept less than those carrying either GAL4 or draper-RNAi alone (S3A Fig). To test whether Draper is required for post-injury sleep, we injured mz0709-GAL4 > draper-RNAi flies and measured sleep immediately following antennal ablation. Knockdown of draper in ensheathing glia abolished post-injury sleep, and injured flies were indistinguishable from intact draper knockdown controls (S3B and S3C Fig). We next asked whether Draper is required for injury-induced changes in waking activity. In mz0709-GAL4 > draper-RNAi flies, there was no difference in waking activity between intact and ablated animals (S3D Fig). Together, these findings demonstrate that Draper functions in ensheathing glia to promote basal sleep and is essential for the post-injury sleep response.

The engulfment phenotypes observed in aged flies phenocopy those previously reported in response to acute sleep deprivation, raising the possibility that sleep deficits contribute to the age-related decline in glial clearance of damaged neurites [19]. To examine whether age-associated sleep loss contributes to loss of ensheathing glial response to injury we sought to increase sleep and measure the effect on clearance of damaged neurons. We pharmacologically enhanced sleep in young and old flies following antennal ablation and measured the effects on glial plasticity and axon engulfment [35,38]. Five- or 40-day old flies were fed the sleep-inducing GABA agonist gaboxadol for 48 hours following injury, after which Wallerian degeneration was quantified (Figs 3A and S4A) [38]. Gaboxadol treatment significantly increased sleep in both young and old flies by increasing the average bout length (Figs 3B, 3C, S4B and S4C). Signal from Or22a::GFP-labeled neurites was significantly reduced in ablated 5- and 40-day-old flies treated with Gaboxadol compared to solvent treated controls revealing that pharmacological induction of sleep restores glial elimination to old flies (Figs 3D, 3E, S4D and S4E). In both 5- and 40-day-old flies there was no effect of gaboxadol treatment on unablated controls, confirming that gaboxadol does not promote axonal engulfment in uninjured controls (Figs 3F and S4F). Similarly, Draper levels were elevated in both 5 and 40-day-old ablated flies treated with gaboxadol compared to age-matched untreated controls, while there was no effect of gaboxadol on Draper levels in flies with intact antennae (Figs 3F’, 3G’, 3I, S4F’, S4G’, and S4I). Therefore, feeding flies gaboxadol, which reportedly increases sleep, restores age-related deficits in neurite clearance and glial activation.

It is possible that the effects of gaboxadol on glial activation and neurite engulfment are due to activation of GABA signaling rather than induction of sleep per se. To differentiate between these possibilities, we genetically activated R23E10 neurons that label sleep promoting neurons in the brain and/or the ventral nerve cord [39–41] and measured the effect on clearance of damaged neurites. Briefly, we ablated the antennae of flies expressing the thermosensitive cation channel TrpA1 in R23E10 neurons (R23E10-Gal4 > TrpA1), and increased the temperature from 22°C to 29°C for 48 hours following injury. Following sleep induction, brains were immunostained for Draper levels (Fig 4A). Consistent with previous findings in young flies, thermogenetic activation of 23E10 neurons increased sleep in 5- and 40-day-old flies (Figs 4B–4D and S5A-S5E) [39]. Induction of sleep enhanced Draper levels in 5-day-old flies compared to genetic controls (Fig 4E and 4F). There was no effect of sleep induction on Draper levels in intact 5-day-old flies, suggesting the effects of sleep on Draper is specific to post-injury response (Fig 4F). Similarly, in 40-day-old flies, genetic induction of sleep increased Draper levels in ablated flies but not in unablated controls (Fig 4G and 4H). Together, these findings confirm that genetically inducing sleep is sufficient to restore post-injury Draper levels in aged flies.

Numerous genetic pathways have been identified as contributing to glial activation and increased Draper levels following axotomy, yet much less is known about how aging impacts glial function [8]. To identify age-dependent changes in ensheathing glia we performed snRNA-seq on the central brains of both young and old flies under intact and axotomized conditions (Fig 5A). We dissected 120–140 central brains of intact or ablated flies at 5 and 40 days of age. This yielded at least 295k nuclei for each of the four experimental groups, resulting in 68,361 nuclei that met the criteria for inclusion in analysis (see Materials and Methods). There were no significant differences in nuclei number between any of the groups tested, suggesting consistency between sample collections. We were able to classify nuclei into diverse types of brain cell types based on known markers (Figs 5B–5C and S6A-S6D). Clustering analysis revealed six distinct populations and chiasm giant glia, and one unannotated group (Figs 5B-5D and S6E). The total number, and types of glia did not including cortex glia (CG), ensheathing glia (EG), astrocyte-like glia (ALG), perineural glia (PG), subperineural glia (SPG) and chiasm giant glia, and one unannotated group (Figs 5B–5D and S6E). The total number, and types of glia did not vary significantly across experimental conditions, supporting the notion that loss of neurite engulfment in aged flies is not due to an overall loss of ensheathing glia (Figs 5E and S6F).

To identify genes that are transcriptionally regulated in response to neural injury in young and old flies, we compared gene expression across experimental groups within the ensheathing glia cluster. We first sought to specifically identify the olfactory ensheathing glia based on the expression of repo and the Toll-1 receptor Sarm, and draper, all of which are highly expressed in olfactory ensheathing glia post neural injury [42–44]. We specifically analyzed the ensheathing glia group. Of these, 16.31% co-expressed draper and Sarm, and were therefore defined as olfactory ensheathing glia (Figs 6A and S7A). This subset was further examined for differential gene expression.

To determine the differences in gene expression between young and old individuals in response to neuronal injury, we compared the transcriptional changes in genes between the young and old groups of olfactory ensheathing glial nuclei, using Seurat’s FindMarkers function. In the comprehensive gene expression analysis, we observed distinct patterns of differential gene regulation between intact and ablated groups in young and old cohorts separately. In the young cohort, the ablated group showed a significant divergence, with 927 genes upregulated and 670 genes downregulated compared to the intact group (Figs 6B and S7B). This robust response suggests a pronounced compensatory mechanism for engulfment of damaged neurites or altered developmental trajectory in response to ablation. Conversely, in the aged cohort, the ablated flies displayed a distinct profile, with 877 genes significantly upregulated, of which 325 overlapped with those upregulated in the young group (Figs 6C and S7B). This overlap indicates a subset of genes that maintain their function throughout the lifespan in response to neuronal injury. Among the 602 uniquely upregulated genes in the young, ablated group, we identified key candidates, including autophagy-related genes Atg12, Atg8a, Rab10, Rab14, Rab7, and Rab3GAP1, along with several ribosomal subunit genes, as well as draper and Sarm. These genes may play pivotal roles in efficient neurite engulfment within 24 hours post-injury and warrant further investigation regarding their functional implications in development and aging (Fig 6B-6D).

Next, we performed pseudobulk differential expression analysis on olfactory EG cell cluster from the raw count data and subsequently used z-score clustering expression to identify different patterns of gene expression across all four experimental groups (S7C Fig). We identified 1,356 genes in cluster 2 that were selectively upregulated in young flies following ablation, but not in old flies. By conducting these two analyses and overlapping the 602 genes with the genes in cluster 2, we specifically focused our analysis on the genes that were upregulated in the young cohort but not in aged flies following neural injury (S7B and S7C Fig). The candidate genes mentioned above, including a set of ribosomal subunit genes and several autophagy-related genes involved in phagocytosis and engulfment pathways, all exhibit this expression pattern (Figs 6D, 6E and S7C). Strikingly, seven large ribosomal protein subunit (RpL) transcripts and five small ribosomal protein subunit (RpS) transcripts were significantly upregulated following ablation in young flies, but not in old ablated flies (Figs 6D and S7D). These genes present numerous potential candidate regulators of glial recruitment and the injury response, highlighting the role of increased ribosomal synthesis following injury (Figs 6E and S7D). Gene ontology analysis revealed that genes related to both ribosome function and autophagy are equally important for cellular processes following injury, specifically associated with Wallerian degeneration (Fig 6E). Together, these analyses identify numerous genes that are transcriptionally regulated by both neural injury and aging, providing candidate regulators for the age-dependent loss of Wallerian degeneration.

Here, we show a link between reduced sleep and age-dependent loss of Wallerian degeneration in Drosophila. Both aging and acute sleep loss are associated with reduced brain function including changes in structural plasticity, long-term potentiation, memory, and elevated oxidative stress levels [25,27,30,45]. Previous work in mice and flies has shown that both aging and sleep loss impair glial engulfment, but the relationship between sleep loss and aging has been unclear [29,19]. Here, we show that pharmacological and genetic induction of sleep restores induction of Draper and glia-mediated engulfment to aged flies.

Age-related loss of glial function is associated with many aspects of declining brain function. In flies and mammals, the clearance of debris and toxic substances from the brain is critical for healthy aging. In mammals, both the overactivation and reduced function of microglia is associated with neurodegeneration [24,46,47], while the glymphatic systems play a critical role in regulating the clearance of amyloid beta from cerebral spinal fluid [22]. In Drosophila, loss of the ensheathing glia is associated with lipid accumulation and reduced lifespan, while genetically increasing ensheathing glia number is protective in a model of Alzheimer’s disease [48]. Further, loss of glial function accelerates age-related changes in memory [49]. These findings suggest ensheathing glia are critical for healthy aging. Here, we find that age-related sleep loss contributes to reduced glial function. These findings raise the possibility that loss of glia function may contribute to the negative impacts of aging.

There are many functional parallels between Drosophila and mammalian glia and their roles in sleep regulation. Both astrocytes and microglia are essential for normal brain function in animals ranging from flies to mice. These effects appear to be highly conserved because astrocytic-like and microglia-like ensheathing glia also regulate sleep in Drosophila [16,50,51]. In some cases, glia appear to impact global brain function, such as regulation of synaptic strength and neural excitability, while in other cases glia appear to modulate levels of specific sleep transmitters or neural circuits [52–55]. It is possible that the effects of sleep and aging in modulating glia function extend beyond olfactory ensheathing glia, to other types of glia in the brain. Supporting this notion, our snRNA-seq data identified Sarm- and draper-positive glial cells in the CG, PG, SPG and ALG subtypes (S7E-S7H Fig). Future studies examining the effects of aging on astrocyte-like glia, cortex glia, and non-olfactory ensheathing glia is critical to understand whether aging and sleep loss impact all types of glia or are specific to the engulfing properties of olfactory ensheathing glia.

In these experiments, we used two complementary methods to increase sleep: activation of neurons labeled by the GAL4 driver 23E10 and administration of the GABA receptor agonist gaboxadol. While both approaches restored Draper induction following injury in aged flies, each approach may not fully recapitulate natural sleep. Specifically, 23E10 activation and gabaxadol differ dramatically in their effects on global neural activity and transcriptional regulation [35,56]. Moreover, although many studies have characterized activation of 23E10 neurons as sleep-promoting, others report no detectable sleep phenotype with this manipulation [57,40]. Numerous additional sleep-promoting neurons have been identified including RDL neurons, DPM neurons, and subsets of mushroom body neurons [58–60]. Furthermore, alternative experimental manipulation, including mechanical stimulation such as vibration to induce sleep offer further avenues to test the hypothesis that sleep restores age-dependent loss of glial function [61]. Incorporating these additional approaches will be critical for rigorously establishing whether sleep is sufficient to rescue age-related declines in glial responsiveness.

Drosophila increase sleep immediately following ablation injury to the antennae or wing suggesting this is a naturally occurring mechanism to promote recovery and glial-mediated engulfment [18]. The degree of post-injury sleep is thought to directly result from synaptic pruning, because expression of neuroprotective factors in the olfactory neurons blunt sleep following injury [18]. Here, we show that post-injury sleep is reduced in aged flies, consistent with the notion that pruning promotes sleep following injury. These findings raise the possibility that bi-directional interactions between neurite clearance and sleep regulation are both impacted by aging. The effects were greatest in the first 12 hours following injury. It is important to note that the analysis in this manuscript was limited to activity-based monitoring. Additional investigations, such as assessments of arousal threshold, whole-brain imaging, or whole-body metabolic rate, are critical to determine the specific type of sleep induced by injury [35,62,63]. Nevertheless, the findings raise the possibility that greater upregulation of neuroprotective factors contribute to the blunted sleep response. Because ORN cell bodies are removed in the process, we are unable to address this using our snRNA-seq protocol. However, future work specifically examining interactions between post-injury sleep induction and neuroprotective factors in aged flies may be informative.

Transcriptional regulation in glia is modulated by sleep, circadian timing, and age [64,65]. These molecular factors directly impact glial function including clearance of debris, regulation of neurotransmitter release, and secretion of sleep-regulating factors [51,66]. For example, insulin signaling promotes sleep and activates ensheathing glia phagocytotic mechanisms in Drosophila, while diet induced insulin resistance inhibits neurite clearance, revealing a critical role for insulin signaling in debris clearance [13,28]. To identify new factors that contribute to age related changes in sleep we performed snRNA-seq in the central brains of old and young flies with ablated antennae. We find numerous genes previously known to be under transcriptional regulation including draper and the autophagy regulators Atg8a and Rab5 are not upregulated following axotomy in aged flies. Altered autophagy is linked to sleep, as autophagosome levels accumulate during periods of wakefulness. Sleep loss is associated with reduced proteostasis, ultimately leading to neurodegeneration [60,61].

Our findings indicate that aging inhibits induction of ribosomal subunit transcription in glia, which is consistent with studies ranging from C. elegans to humans that reveal an age-associated loss of ribosomal protein subunits [67,68]. Consequently, the loss of ribosomal biogenesis may inhibit the structural and functional plasticity required to engulf aging neurites. It is possible that the injury-induced ribosomal subunits are generally required for enhanced translation following injury, or that these subunits confer unique functions. In Drosophila, selective loss of individual ribosomal subunits is associated with defects in development and cellular function [69–71] Furthermore, the TOR pathway has been implicated more generally in age-associated loss of ribosomal biogenesis including reduced TOR signaling [72]. Therefore, it is possible that pharmacological enhancement of ribosomal biogenesis through these pathways has potential to ameliorate the cellular, and possibly functional, effects of age-related sleep loss [73].

Ensheathing glia are critical for maintaining brain health, and preventing their age-dependent loss extends lifespan in Alzheimer’s model Drosophila [48]. A central question is how the transcriptional changes identified in aged flies alter the physiology of ensheathing glia, and whether these changes in turn influence longevity. Multiple studies have applied Ca² ⁺ imaging to characterize physiological changes across diverse classes of glia [74]. In ensheathing glia specifically, Ca² ⁺ activity is highly dynamic and responsive to shifts in neurochemistry and behavior [75]. Given that optogenetic activation of ensheathing glia promotes sleep, age-related changes in brain-wide neurochemistry or transcriptional regulation within glia could have important consequences for both glial function and behavior [75]. Future studies examining the impact of ribosomal dysregulation on ensheathing glia physiology will be particularly valuable for pinpointing gene-specific contributions to the age-dependent decline in glial function.

Taken together, our findings demonstrate a role for sleep in the age-related loss of glial-mediated phagocytosis. Sleep is sufficient to restore glial membrane plasticity, facilitate the induction of Draper, and enhance neurite engulfment in aged flies. These findings provide a framework for examining the effects of sleep on glial-mediated phagocytosis and how sleep and aging impact glial function more broadly. Given that many aspects of glial function, including their roles in regulating brain health, are conserved across species from flies to mammals, this system has the potential to advance our understanding of changes in glial function throughout the lifespan.