Nature News · Feb 11, 2026 · Collected from RSS
MainSleep is a behavioural state shared by almost all animals. It is defined as a quiescent state associated with reduced consciousness that is different from coma or anaesthesia because it is rapidly reversible with a stimulus. The circadian system regulates sleep on a 24-h cycle, but sleep is also regulated by homeostatic mechanisms, whereby the pressure to sleep increases with extended periods of wakefulness1. The importance of sleep is recognized widely, but the underlying mechanisms and functions are still debated1,2.Whereas studies of sleep focus on the brain, sleep loss also has an effect on the periphery3. In addition, there is now reason to believe that peripheral tissues can affect sleep4,5. For instance, the immune system is implicated in the control of sleep, particularly during sickness. In Drosophila, the nuclear factor kappa B protein Relish acts in the fly fat body (functional equivalent of the liver) to regulate sleep following infection6. Sleep, in turn, influences recovery from bacterial and viral infections in both mammals and flies7,8. Sleep deprivation increases the expression of sleep-promoting cytokines, such as tumour necrosis factor (TNF) or interleukin (IL)-6 and it may do so in the same way as inflammation, by increasing levels of the glucocorticoid hormone through the hypothalamic–pituitary–adrenal axis or noradrenaline through the sympathetic nervous system7. TNF is also implicated in Drosophila sleep through its expression in astrocytes, but it can act systemically as well9. In general, interactions between sleep and the immune system have focused largely on signalling molecules, and not on regulation by immune cells in the periphery. Little is known about the role of peripheral mechanisms in the function of sleep.Using Drosophila as a model system, we addressed a sleep function for circulating blood cells called haemocytes, 95% of which are macrophage-like plasmatocytes10 that function in immune responses. We show that, at times of high sleep, haemocytes localize to the brain and take up lipids accumulated in cortex glia. As lipid droplets (LDs) in cortex glia reflect the transfer of wake-associated oxidative damage from neurons11, this uptake by haemocytes is expected to ease metabolic stress in the brain. Indeed, loss of the Eater receptor, which mediates lipid uptake by haemocytes, causes increased acetylation in the brain, along with mitochondrial oxidation and reduced NAD+ levels. Thus, haemocytes, and Eater in particular, act in a sleep-dependent fashion to maintain metabolic homeostasis in the brain.Haemocytes interact with the BBB during high sleepTo explore a possible interaction between haemocytes and the brain, we first used a tissue-clearing method12 to visualize circulating haemocytes in the fly head (Fig. 1a and Extended Data Fig. 1a). We used the HmlΔ-LexA fly line to label haemocytes, and detected their localization throughout the head area including the proboscis (pb), maxillary palp (mp) and ocellar (oc) regions, but not in the eyes or antenna (at) (Fig. 1a). To determine whether these populations of haemocytes mostly circulate or actually contact the brain, we dissected fly brains and visualized haemocytes using various markers (Fig. 1b,c and Extended Data Fig. 1b–d). Consistent with previous studies, Hml+ haemocytes were located mostly near the dorsal part of the brain13, especially in the dorsally located pars intercerebralis region (Extended Data Fig. 1c). These Hml+ cells were also positive for other haemocyte markers14 such as Srp-Hemo (Fig. 1b) and NimC1 (Fig. 1c). Moreover, they were observed with other haemocyte Gal4 drivers14, such as Srp-Gal4, Ppn-Gal4 and Srp-Hemo-split-Gal4, and they expressed eater-dsRed, which is specific to haemocytes (Extended Data Fig. 1b). Haemocytes were not detected near the ventral nerve cord or inside the brain (Extended Data Fig. 1d). Based on these observations, we conclude that haemocytes circulate in the fly head cavity and possibly interact with the brain at specific sites.Fig. 1: Blood cells (haemocytes) circulate in the fly head cavity and contact the BBB.a, Haemocytes within the fly head cavity, visualized with mCherry (red) driven by an Hml+ driver. b,c, Haemocytes labelled with GFP driven by Hml+ (green) and Srp-hemo+ (red) localize to the dorsal middle area of the brain (left). Dotted box shows magnified area. Brain is visualized with brp (magenta) (b). Hml+ (green) and NimC1+ (red) haemocytes are in the same area as Srp+ cells (left) (c). d,e, Localization of haemocytes (red) near the BBB (green) (d). NimC1+ (cyan) haemocytes are next to PG cells (R85G01-LexA, green) (top) or SPG glial cell (R54C07-LexA, green) (bottom). Cortex glia (CG) are visualized with RFP (NP2222-Gal4, red) (e). f–h, Sleep dependence of haemocyte recruitment to the head. Haemocytes (red) are visualized at different ZT times, with sleep deprivation or gaboxadol feeding (f). Haemocytes were quantified at different times of day (n = 22, 26, 19, 20, 9 and 22 from left to right on the graph, g) or after sleep deprivation (SD) or gaboxadol (n = 14, 16, 18, 20, 17, 18, 20, 16, 20 and 21 from left to right, h). Yellow line, average; error bars, s.d. i–j, Effects on haemocyte recruitment of manipulating sleep-promoting and wake-promoting neurons. Haemocytes (red) were visualized following wake- or sleep-promoting manipulations (i). Data in i were quantified (n = 12, 14, 13, 14, 28, 28, 23 and 24 from left to right on the graph; horizontal bars, median) (j). NS, not significant; P > 0.01); *P < 0.1; **P < 0.01; ***P < 0.001; ****P < 0.0001. n represents biologically independent samples. Two-sided Tukey’s multiple comparisons test was performed for all data analysis. Detailed statistics in Supplementary Table 1. Scale bars, 100 μm (a, b–d (left panels) f, i); 50 μm (b–d, right panels); 10 μm (e). oc, ocelli; at, antenna; mp, maxillary palp; pb, proboscis.Full size imageThe fly brain is separated from the periphery by the blood–brain barrier (BBB)15. To assess whether haemocytes near the pars intercerebralis region interact with the BBB, we visualized haemocytes together with the BBB-specific Gal4 line that marks both perineurial glia (PG) and sub-perineurial glia (SPG) cells of the BBB15 (Fig. 1d and Extended Data Fig. 1b), or using Gal4 lines that individually mark PG (NP6293-Gal4) or SPG (moody-Gal4) cells (Fig. 1e and Extended Data Fig. 1e). We found that Srp-Hemo+, eater+ or NimC1+ haemocytes were located adjacent to the BBB (Fig. 1d,e and Extended Data Fig. 1b,e). Moreover, when we visualized haemocytes together with PG markers, we observed that extensions of the PG membrane physically contact haemocytes (Fig. 1e and Extended Data Fig. 1f). Although SPG and haemocytes are separated by PG cells, haemocyte membranes are also in contact with SPG cells (Fig. 1e). Indeed, we used the green fluorescent protein (GFP) reconstitution across synaptic partners (GRASP) technique16 to confirm direct physical interaction between haemocytes and SPG cells (Extended Data Fig. 1g,h). Similar direct interaction of SPG cells with haemocytes was observed in pupal stages by electron microscopy17. A previous study found that cortex glia cells contact or share membranes with SPG18 cells, so it is possible that haemocytes also directly contact cortex glia cells but we did not detect a GRASP signal (data not shown). Based on these findings, we conclude that haemocytes exist within the fly head cavity and interact physically with glial cells, particularly glia of the BBB.We also asked whether haemocyte recruitment to the brain is influenced by the sleep:wake cycle (Fig. 1f,g). At Zeitgeber time (ZT) 8 and ZT20 (ZT0 = lights on, in circadian terms), which are times of the afternoon siesta and night-time sleep, respectively, the number of haemocytes in the head was higher than at other times of day (Fig. 1f,g). To confirm sleep-dependent haemocyte recruitment to the fly head, we compared haemocyte numbers following sleep deprivation or gaboxadol feeding to induce sleep (Fig. 1f,h). Sleep deprivation reduced haemocyte numbers in the head, but the numbers recovered during rebound sleep (Fig. 1f,h). By contrast, feeding gaboxadol increased haemocyte numbers in the fly head, with no significant differences across ZT time points (Fig. 1f,h). These results were supported by manipulations of neuronal activity to increase/decrease sleep (Fig. 1i,j). Thermogenetic stimulation of wake-promoting neurons (C584-Gal4 UAS-TrpA1or TH-Gal4 UAS-TrpA1)19 decreased haemocyte number in the fly head, whereas similar stimulation of sleep-promoting neurons (R23E10-Gal4 UAS-TrpA1)20 increased haemocyte recruitment to the head (Fig. 1i,j). From these results, we surmise that the interaction between haemocytes and the brain is increased during sleep. eater expressed in haemocytes regulates sleepGiven that haemocytes circulate in the fly head and are more abundant during sleep than wake, we hypothesized that the function of haemocytes may be relevant for sleep. To use an unbiased approach towards such function, we examined recent single-cell RNA sequencing data21 for transcripts expressed highly in haemocytes. Analysis of the biological functions of the top 100 genes in haemocytes through g:Profiler22 revealed significant annotations for defence response against other organisms, phagocytosis and immune system processes. As the phagocytosis of Gram-positive or Gram-negative bacteria in Drosophila is mediated typically by Nimrod receptor family genes23, we focused on a possible role for this family (NimA, NimB1-5, NimC1-4, drpr, eater, Col4a1, PGRP-LC) in the regulation of sleep. In addition, because we observed that more haemocytes are localized in the head cavity during the sleep state in flies (Fig. 1f–j), we also added genes previously identified as being involved in haemocyte migration. These include activin-β signalling factors (babo, put, Smox) that are important for sessile localization of ha