How does a good night’s sleep re-energize the brain? At least in fruit flies, by burning damaged lipids. So claim scientists led by Amita Sehgal at the University of Pennsylvania in Philadelphia. In the February 15 Nature Neuroscience, they reported that during waking, neurons offload oxidized lipids to glia, which store them as droplets. Later, as the insects sleep, glia burn off the fat, and any decrepit mitochondria in glia or neurons are cleared to ready the flies for a new day. Interrupting this neuron-glia lipid shuttle interrupts sleep as well.

  • In fruit flies, brain lipid metabolism syncs with the sleep cycle.
  • During the day, neurons ship oxidized lipids to glia, which store them as droplets.
  • At night, the glia metabolize the fat.

“[This] study is a beautiful example of the complex interaction of astrocytes and neurons [and] is exciting because it expands neuroglia interactions into sleep,” Maiken Nedergaard, University of Rochester Medical Center, New York, wrote to Alzforum.

This work may help scientists understand links between poor sleep and pathologies in Alzheimer’s disease brain, for example the known habit of microglia to accumulate lipid droplets (LDs) in response to Aβ (Aug 2019 news; Apr 2023 conference news).

“The connection between sleep and AD was obvious but it could never be linked to a molecular pathway,” Rik van der Kant, Vrije University, Amsterdam, told Alzforum. “This paper provides a novel way of connecting sleep to lipid homeostasis, which is directly linked to AD.”

Hugo Bellen and colleagues at the Baylor College of Medicine in Houston had discovered the same neuron-glia lipid shuttle in flies carrying AD risk genes, finding their glia bloated with lipid droplets (Moulton et al., 2021). “I thought it was a pathological phenomenon, but Seghal’s paper shows that this shuttle is physiological,” Bellen told Alzforum. “We now know that glial LD formation, which is a highly conserved pathway across multiple species, is promoted by neuronal activity and that glial LDs are consumed during sleep in preparation for another day of activity,” Bellen and colleagues wrote in a Nature Neuroscience News & Views article.

To understand how sleep supports brain metabolism, first author Paula Haynes let Drosophila go about their normal days—i.e., with lights on and off for 12 hours at a time—or she sleep-deprived them for a night by shaking the vials they are kept in every 30 seconds. She studied flies either just after turning the lights on, two hours later, just after turning the lights off, or in the middle of their night. To analyze the poppy seed-sized fly brain, Haynes added fluorescent reporters of reactive oxygen species (ROS), lipid metabolism, and mitochondrial health, i.e., mitophagy, and then imaged the brain under a confocal microscope.

Neuronal activity demands copious energy, generated through mitochondrial oxidative phosphorylation, hence the authors measured oxidative damage in neurons and glia using fluorescent redox proteins, expressed in mitochondria under cell-specific promoters. These proteins switch colors when oxidized, reflecting accumulating ROS. Surprisingly, cortex glia, not neurons, built up mitochondrial ROS during the day (image below). Fruit flies have four types of glia: cortex glia that wrap around neuron bodies, ensheathing glia that surround neuronal projections, blood-brain barrier glia, and astrocytes. They do not have microglia analogous to those of the immune cell lineage found in mammals.

ROS Refresh. After a night’s sleep (top), a fly’s brain showed few oxidized proteins (yellow) under a glial promoter, aka mitochondrial ROS in glia. After sleep deprivation (middle) or by the end of a day (bottom), fly brains had a lot. [Courtesy of Haynes et al., Nature Neuroscience, 2024.]

ROS can oxidize lipids. Indeed, lipid droplets that lit up with a fluorescent antibody against peroxidated lipid byproducts accumulated in the cortex glia, matching the distribution of ROS. The droplets peaked in the middle of the night then troughed right around waking, suggesting that the glia accumulate these peroxidated lipids during the day and clear them overnight.

Peroxidated lipids are generated by ROS inside neurons. How did they end up in glia? The authors surmised they were brought there courtesy of apolipoproteins. Flies express ApoE orthologs NLaz in neurons and GLaz in glia. When Haynes knocked down either, glia formed fewer lipid droplets. Neurons did not accumulate lipid droplets. Further, neuronal but not glial mitochondria became oxidized. This suggests that neurons were indeed the source of the peroxidated lipids and that, by shuttling them to glia, neurons protected their own mitochondria from oxidative stress.

To test this idea, Haynes knocked down neuronal Drp1, an ROS-activated mitochondrial damage response protein. Without Drp1 in neurons, glia accumulated fewer lipid droplets by the end of the day, suggesting that mitochondrial damage control drives the neuron-glia lipid shuttle. In contrast, sans Drp1, glia couldn’t metabolize lipid droplets overnight as usual, suggesting they need mitochondria to power lipid catabolism. All told, the data suggest that mitochondrial activity drives the wake/sleep cycle that transfers oxidized lipids from neurons to glia for breakdown. “We propose that a mitochondrial lipid metabolic cycle between neurons and glia reflects a fundamental function of sleep relevant for brain energy homeostasis,” wrote the authors.

Would tinkering with the proteins in this cycle mess with sleep? Indeed, flies without NLaz, GLaz, or Drp1 in either neurons or glia cell slept less, and the shut-eye they got was fragmented. The same happened when the scientists suppressed genes that break down fatty acids, in glia.

Interrupted sleep cycles have been linked to an increased risk of AD and worse amyloid pathology (Feb 2017 news; Apr 2018 news). Even cognitively normal older adults with one or two copies of the AD risk gene APOE4 have disrupted sleep (Blackman et al., 2022). In fruit flies expressing human APOE4 instead of GLaz, glia poorly accumulated lipid droplets, which hastened neurodegeneration (Liu et al., 2017). Julia TCW of Boston University takes these results, and Sehgal’s, to mean that ApoE4 hobbles the neuron-glia lipid shuttle, which might triggers sleep problems and neuronal damage in people.

Russell Swerdlow, Kansas University Medical Center, Kansas City, noted that sleep troubles seem to start early in the AD trajectory; he attributes this primarily to floundering mitochondria. “Sleep disruption may prove to be less of a driver of AD, and more a consequence of mitochondrial strain,” he wrote (full comment below).—Chelsea Weidman Burke


  1. This is a fantastic study. It addresses fundamental issues of brain bioenergetic metabolism that relate and link to several AD-relevant phenomena, including neuron-glia interplay, apolipoprotein E biology, lipid metabolism, oxidative stress, circadian rhythms, sleep, and mitochondrial biology. While “mitochondriacs” will find the overall findings unsurprising, it is nevertheless quite exciting to see the connections between mitochondria and sleep laid out at this level of detail. It is thrilling to think about how this system arose through evolution and how it ties to AD.

    In general, we continuously run our neuron mitochondria when awake, which initiates the need for regular mitochondrial and probably general cell tune-ups. To accomplish this, worn-out, likely oxidized neuron membrane material is shipped to astrocytes through an ApoE-dependent transfer event, and the transferred membrane is first stored and then catabolized through fatty acid beta oxidation in astrocyte mitochondria.

    To better understand the need for this neuron-astrocyte interaction, one needs to know that neuron mitochondria are not wired to perform fatty acid oxidation. The astrocyte mitochondria fatty acid catabolism event happens during sleep, and neuron mitochondria also take advantage of the sleep-related downtime to repair themselves. By the time awakening occurs, the neuron mitochondria are ready to go another round, and the bioenergetic metabolism-oxidized fatty acid byproducts have been eliminated and likely even recycled. For all this to happen, of course, we simply need to sleep.

    The authors speculate that when sleep breaks down, this biological system cannot perform, and a resultant disruption of the cycle may promote neurodysfunction and neurodegeneration. This could turn out to be the case, but the nature of the modeling arguably limits our ability to extrapolate to the human AD state.

    From the perspective of the mitochondrial cascade hypothesis, one might speculate that failing mitochondria would throw this whole system out of whack, with ApoE isoform status helping to determine the extent of the resulting disequilibrium. In other words, sleep disruption may prove to be less of a driver of AD, and more a consequence of AD mitochondrial strain.

  2. This exciting new paper from the Seghal lab makes an unexpected connection between reactive oxygen species (ROS), lipid metabolism, and sleep. Previous work had shown that neurons release ROS-damaged peroxidated lipids when overexcited or stressed; these lipids are then taken up by glia and temporarily stored in lipid droplets before being metabolized in mitochondria (Liu et al., 2015; Ioannou et al., 2019). Most previous work on this process focused on disease conditions. However, Haynes and colleagues discovered that glial lipid droplet accumulation occurs as part of a daily rhythm: in flies, glia accumulate lipid droplets during the day in response to normal neuronal activity, which are then metabolized by glial mitochondria at night during sleep. The daily transfer of lipids from neurons to glia required neuronal lazarillo (NLaz), a fly apolipoprotein similar to human apolipoprotein E (APOE): knocking down NLaz reduced glial lipid droplet accumulation.

    Interestingly, reducing the expression of glial lazarillo (GLaz) resulted in more lipid droplets rather than fewer, which was unexpected. The authors suggest that GLaz may play a function in lipid droplet formation or metabolism within glia, in addition to its role in transporting lipids between cells. This is consistent with recent results from my lab showing that APOE can localize directly to the surface of glial lipid droplets, with APOE knockdown causing larger lipid droplets (Windham et al., 2024). Importantly, reducing either NLaz or GLaz expression resulted in sleep disruption.

    This work identifies neuron-glia lipid transport and metabolism as an important new regulator of sleep and raises fascinating questions about how these findings might translate to humans, especially in the context of Alzheimer’s disease. The E4 variant of APOE is the biggest genetic risk factor for developing late-onset Alzheimer’s disease. APOE4 mice were previously found to be more sensitive to sleep disruption than APOE3 mice, with sleep disruption in E4 mice resulting in increased Aβ deposition and tau seeding (Wang et al., 2023). Sleep disturbances are also associated with Alzheimer’s and other neurodegenerative diseases.

    This work opens new avenues to explore the mechanistic relationship between APOE, sleep, and AD. Questions remain about how exactly glial lipid metabolism regulates sleep need, or sleep behavior. In addition, it remains to be seen whether human APOE is important for intercellular lipid transfer and/or glial lipid droplet metabolism in the context of sleep, and whether any of these functions are affected by APOE genotype.


    . Glial lipid droplets and ROS induced by mitochondrial defects promote neurodegeneration. Cell. 2015 Jan 15;160(1-2):177-90. PubMed.

    . Neuron-Astrocyte Metabolic Coupling Protects against Activity-Induced Fatty Acid Toxicity. Cell. 2019 May 30;177(6):1522-1535.e14. Epub 2019 May 23 PubMed.

    . APOE traffics to astrocyte lipid droplets and modulates triglyceride saturation and droplet size. J Cell Biol. 2024 Apr 1;223(4) Epub 2024 Feb 9 PubMed.

    . APOE-ε4 synergizes with sleep disruption to accelerate Aβ deposition and Aβ-associated tau seeding and spreading. J Clin Invest. 2023 Jul 17;133(14) PubMed.

  3. This is an exciting finding. It builds on work from Hugo Bellen’s lab, which showed that astrocytes can take up lipid droplets from stressed/injured neurons as a neuroprotective mechanism. This new paper provides a novel connection to sleep deprivation, which is bi-directional. This is really interesting.

    Our lab recently published data showing that the circadian clock protein REV-ERBalpha can regulate microglial lipid droplets, so both sleep and circadian rhythms may influence lipid droplets in the brain. One implication of this paper is that a possible novel function of sleep (or circadian rhythms) might be to limit activity-induced lipid peroxidation and mitochondrial dysfunction.

    This work helps us understand how sleep restriction/disruption can influence brain health and perhaps neurodegenerative pathology, though this was not specifically tested here. We know that people with AD have sleep disruptions, and this study gives us a new idea of why that might happen.

    There are a couple of caveats to consider, however. Lipid droplets seem to be very prominent in flies and are seen in a variety of pathological conditions, as well as in cell culture, but they are less common in mouse brain and more difficult to induce in vivo. Moreover, this paper does not examine this phenomenon in any AD models, so the relevance to AD is speculative. However, it is an exciting finding and will certainly spur more research on this topic in mammalian systems and AD models.

  4. This is a beautiful study with compelling results. That the authors discovered a role for glia in protecting neurons from wake-related oxidative stress is quite unexpected and eye-opening.

    The next step will be to test the extent to which this mechanism applies to mammalian sleep. Mice and humans spend 15-20 percent of their total sleep in REM sleep, during which neuronal activity is even higher than in waking. Further, REM sleep accounts for most of the sleep during development. Thus, it will be very interesting to see whether this newly discovered shuttle system is restricted to NREM sleep in mammals.

    Also, neurons in several brain regions, including the hippocampus, are equally active in waking as in sleep, and often more active in sleep (for instance, many cells in the dentate gyrus). Thus the “stress” of waking, and therefore the need to sleep, is likely to depend not only on neuronal activity, but also on other factors.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Newly Identified Microglia Contain Lipid Droplets, Harm Brain
  2. Dysregulated Lipid Metabolism Comes to the Fore at AD/PD
  3. A Change in Sleep Habits from Normal to Long: Harbinger of Dementia?
  4. Does Amyloid Accumulate After a Single Sleepless Night?

Paper Citations

  1. . Neuronal ROS-induced glial lipid droplet formation is altered by loss of Alzheimer's disease-associated genes. Proc Natl Acad Sci U S A. 2021 Dec 28;118(52) PubMed.
  2. . APOE ε4, Alzheimer's disease neuropathology and sleep disturbance, in individuals with and without dementia. Alzheimers Res Ther. 2022 Mar 30;14(1):47. PubMed.
  3. . The Glia-Neuron Lactate Shuttle and Elevated ROS Promote Lipid Synthesis in Neurons and Lipid Droplet Accumulation in Glia via APOE/D. Cell Metab. 2017 Nov 7;26(5):719-737.e6. Epub 2017 Sep 28 PubMed.

Further Reading

Primary Papers

  1. . A neuron-glia lipid metabolic cycle couples daily sleep to mitochondrial homeostasis. Nat Neurosci. 2024 Feb 15; PubMed.
  2. . Glial lipid droplets resolve ROS during sleep. Nat Neurosci. 2024 Feb 15; PubMed.