Glucose uptake in pigment glia suppresses Tau-induced inflammation and photoreceptor degeneration Open Access

Alzheimer's disease (AD) is a progressive neurodegenerative disorder and the most common cause of dementia among older people (Knopman et al., 2021). AD is characterized by the extracellular deposition of β-amyloid and intracellular accumulation of abnormally phosphorylated forms of the microtubule-associated protein Tau (MATP in humans, hereafter referred to as Tau) (Samudra et al., 2023). Neuroinflammation is another core pathology in AD and other neurodegenerative disorders (Sobue et al., 2023). Inflammatory responses in the central nervous system are mediated by mainly glial cells, including microglia and astrocytes, to protect against infections and injuries; however, under neurodegenerative disease conditions, glial cells lose their neuroprotective functions and acquire reactive phenotypes (Ransohoff, 2016; Streit, 2002). Although they can remove protein aggregates by phagocytosis, activated glial cells can also harm neurons via excess phagocytosis and the release of neurotoxic cytokines (Chen and Holtzman, 2022). However, what triggers glial activation under disease conditions is not fully understood.

A decline in brain glucose metabolism has been pathologically linked to AD (Butterfield and Halliwell, 2019). Most of the ATP required to support brain function is supplied by glucose metabolism (Bélanger et al., 2011), and glucose utilization is reduced in the brains of AD patients (Dukart et al., 2013; Huang et al., 2020; Oka et al., 2021). Dysregulated glucose metabolism in AD may be due to cerebral hypoperfusion (Park et al., 2019) or downregulation of glucose transporter proteins (GLUTs) that mediate glucose uptake across the blood-brain barrier and delivery to glial cells (Bélanger et al., 2011; Koepsell, 2020; Kyrtata et al., 2021; Liu et al., 2008). Glial cells also rely on glucose, and cellular metabolism can regulate their activation (Wang et al., 2019; Xiang et al., 2021). The mitochondrial oxidative phosphorylation is reduced in the activated microglial cells (Nair et al., 2019), and mitochondrial dysfunction in astrocytes has been reported in amyotrophic lateral sclerosis and neuroinflammation models (Fiebig et al., 2019; Martínez-Palma et al., 2019). While these reports suggest that targeting glial metabolism is a strategy to decrease neuroinflammation and suppress neurodegeneration, mechanistic investigation in an in vivo model is required to explore this possibility.

Here, we investigated the roles of glucose metabolism in glial cells on Tau-induced neurodegeneration in a Drosophila model. We identified glia-associated phenotypes in the fly retina expressing human Tau, such as inclusion-like structures by phagocytosis and induction of antimicrobial peptide (AMP) expression. Drosophila retina contains neurons and several glial cells, and we found that pigment glia plays a critical role in Tau-induced photoreceptor degeneration. Enhancement of glucose uptake in pigment glial cells ameliorates Tau-induced laminal cortex swelling as well as photoreceptor degeneration. Our results suggest that glucose hypometabolism in glial cells contributes to neurodegeneration in tauopathy.

Expression of human Tau in Drosophila by using the pan-retinal eye-specific GMR-GAL4 driver (Freeman, 1996) causes a rough-eye phenotype in the compound eye surface accompanied by the formation of vacuoles caused by apoptosis in the retina and axonal degeneration in the lamina (Fig. 1A) (Iijima-Ando et al., 2012; Jackson et al., 2002; Wittmann et al., 2001). UAS-Tau transgene alone does not cause these phenotypes, as described previously (Wittmann et al., 2001 and Fig. S1). Transmission electron microscopy (TEM) revealed degenerating photoreceptor neurons with disoriented rhabdomeres in Tau-expressing flies (Fig. 1B,C, ommatidium colored in yellow). We also noticed swelling of the laminal cortex (shaded in blue in Fig. 1A) that is located between the lamina and retina, and consists of layers of L1-L5 neurons, surface glia – i.e. perineurial glia (PNG) and subperineurial glia (SPG) − and cortex glia (Chotard and Salecker, 2007; Kremer et al., 2017). We analyzed neuronal and glial cell distribution in the laminal cortex by using glial expression of mCherry carrying the nuclear localization signal (mCh::NLS) and immunostaining for the neuronal marker protein elav (Fig. 1D). In the laminal cortex, surface glia and cortex glia form two layers, and cell bodies of L1-L5 neurons another layer. With Tau expression in the retina, the neuronal layers in the laminal cortex were enlarged by an increased number of neurons. In contrast, glial cells in the laminal cortex were degenerated (Fig. 1D).

Tau expression caused the formation of dense inclusion-like structures in the retina and lamina (Fig. 1A). TEM analysis of these structures revealed two distinct ultrastructures: an invaginated corneal lens (Fig. 1A,E,G, arrowheads) as well as dark inclusions containing cellular debris (Fig. 1A,F,G, arrows) wrapped by glia (Fig. 1F, white arrows). These inclusions contain ubiquitylated proteins and cytoskeleton proteins, such as filamentous actin (Fig. 1G), suggesting that they are dead cells that harbor abnormal protein accumulation and are engulfed by glia. These results suggest that glial cells are also affected by Tau protein expression.

The appearance of inclusion-like structures (Fig. 1F) was similar to that of glial bodies, membranous structures that represent abnormal multilayered wrappings by glial cells (Dutta et al., 2016). We were motivated to test if these inclusions are formed through glial phagocytosis. Glial phagocytic activities in the adult fly brain are mediated by the two glial transmembrane phagocytic receptors Draper (Drpr) and Nimrod C4 (NimC4, also known as SIMU) (Kurant et al., 2008; Scheib et al., 2012). We found that blocking phagocytosis in the retina by RNAi-mediated knockdown of drpr and NimC4 did not significantly affect the eye size or photoreceptor degeneration but reduced the number of inclusions in the tau-expressing retina (Fig. S2, Fig. 2A). These results indicate that these inclusions are formed by glial engulfment activity.

Proinflammatory responses in Drosophila involve the expression of AMPs downstream of Toll and immune-deficiency (Imd) pathways (Harnish et al., 2021; Hedengren et al., 1999; Sakakibara et al., 2023; Shukla et al., 2019). We found significantly elevated expression levels of AMPs, including Drosomycin (Drs), Cecropin A1 (CecA1), Metchnikowin (Mtk), Diptericin B (DptB), and Attacin C (AttC), in flies that express Tau driven via GMR-GAL4 (Fig. 2B). These results suggest that Tau expression in the fly retina induces glial phagocytosis and inflammatory responses, which are both hallmarks of glial activation.

We analyzed the effects of enhancement of glucose uptake in Tau-induced phenotypes in the retina by expressing human SLC2A3 (officially known and hereafter referred to as GLUT3), which has been shown to enhance glucose uptake in cells effectively (Besson et al., 2015; Manzo et al., 2019; Oka et al., 2021). Coexpression of GLUT3 increased the eye size in Tau flies (Fig. 3A, compare Tau and Tau+GLUT3, P<0.001). The eyes of flies coexpressing Tau and mCD8::ChRFP were similar to those in flies expressing Tau alone, indicating that the increase in eye size caused by coexpression of GLUT3 is neither due to non-specific effects of an exogenous protein nor to reduction in GAL4 activity (Fig. 3A, compare Tau and Tau+mCD8::ChRFP, P>0.05). Expression of GLUT3 alone did not affect eye size (Fig. 3A, compare Control and GLUT3, P>0.05). Rhodopsin 1 (Rh1) protein levels are significantly lower in Tau flies, probably because the loss of photoreceptor neurons (Fig. 3B, compare Control and Tau). Coexpression of GLUT3 restored Rh1 protein levels (Fig. 3B, compare Tau and Tau+GLUT3, P<0.001). There were also significantly fewer vacuoles in the retina of flies coexpressing Tau and GLUT3 compared with that in flies expressing Tau alone (Fig. 3C-D, P<0.01). However, at ultrastructural levels, while the retina of Tau-expressing flies contained degenerating photoreceptor neurons with distorted rhabdomeres (Fig. 1C, ommatidium colored in yellow), ommatida of flies that coexpress Tau and GLUT3 retained rhabdomeres with proper orientation (Fig. 3F, Tau+GLUT3). Coexpression of GLUT3 did not affect the number of inclusion-like structures (Fig. 3E), but suppressed swelling of the laminal cortex (Fig. 3E). Overexpression of Drosophila Glut1 (Fig. S3), a fly homolog of human GLUT3, also suppressed tau-induced retinal degeneration and swelling of the laminal cortex (Fig. 3G-I). These results indicate that enhancement of glucose uptake suppresses Tau-induced neurodegeneration and glial phenotypes.

To understand the mechanisms underlying the neuroprotective effects of enhanced glucose uptake, we first analyzed the effect of GLUT3 expression on phosphorylation and accumulation of Tau by using western blotting with monoclonal antibodies. Coexpression of GLUT3 did not affect the levels of total Tau (Total Tau) or its phosphorylation at disease-associated sites, including phosphorylation at Ser262 (pSer262), Ser396/404 (PHF1) or Ser202 (CP13) (Fig. 4). These results suggest that enhanced glucose uptake mitigates retinal degeneration downstream of Tau phosphorylation and accumulation.

The regulatory sequences of the Drosophila ninaE gene used in GMR-GAL4 drives expression in all cells behind the morphogenetic furrow, including pigment glial cells (Freeman, 1996). Pigment glial cells surround photoreceptor neurons and support their functions by providing nutrition and by taking up toxic materials (Liu et al., 2017). Expression of GMR-GAL4 and 54C-GAL4 in pigment glial cells was confirmed by using UAS-mCh::NLS (Fig. S4). Although Tau expression in pigment cells does not cause a rough-eye phenotype (Fig. S5), GLUT3 expression in pigment cells might mediate the protective effects against photoreceptor degeneration. To dissect the effects of GLUT3 expression in glial cells during Tau-induced retinal degeneration, we used GMR-Tau (P301L) − which expresses a human Tau variant carrying the pathological Pro 301 to Leu mutation (P301L), directly under the control of the GMR regulatory sequence (Jackson et al., 2002) − in combination with the GAL4/UAS system to express GLUT3 under the 54C-GAL4, a pigment glia-specific driver (Nagaraj and Banerjee, 2007). We found that laminal cortex swelling (Fig. 5A) and expression of AMPs (Fig. 5B) was suppressed by expression of GLUT3 in pigment glia. Intriguingly, pigment cell expression of GLUT3 in the Tau-expressing retina also suppressed the degeneration of photoreceptors (Fig. 5A,C). Vacuole formation was suppressed, eye size was significantly increased and protein levels of Rh1 were restored upon GLUT3 expression in pigment glia cells (Fig. 5A,C,D). In contrast, expression of GLUT3 in photoreceptor neurons by Rh1-GAL4 did not mitigate photoreceptor degeneration (Fig. 5E-G), indicating that the protective effects of GLUT3 coexpression against Tau toxicity observed with GMR-GAL4 (Fig. 3) was mediated by pigment glia. These results suggest that enhancement of glucose uptake in pigment glia mitigates inflammatory responses and Tau-induced degeneration of photoreceptors.

Enhancing glucose uptake by neuronal overexpression of GLUTs has been reported to ameliorate neurodegeneration in several neurodegenerative conditions, such as cell death caused by expression of huntingtin, β-amyloid or TDP-43 in Drosophila (Besson et al., 2015; Manzo et al., 2019; Niccoli et al., 2016; Cabirol-Pol et al., 2018), as well as in age-related neuronal dysfunction (Oka et al., 2021). However, the beneficial effect of enhanced glucose uptake in glial cells has not been examined. In this current study, we characterized glial phenotypes in the retina of flies that express human Tau and examined the effects enhanced glucose uptake has within their retinal pigment glia cells. We found that Tau expression causes phenotypes mediated by glia, such as inclusion formation, swelling of the laminal cortex and expression of AMPs. We also found that the expression of glucose transporters in pigment glia suppresses many of these phenotypes and mitigates photoreceptor degeneration downstream of Tau. Our results suggest that glucose metabolism in glial cells contributes to neuroinflammation. The role of glucose metabolism in glia cells to metabolically support neurons has been well established (Morrison et al., 2013; Pellerin and Magistretti, 1994; Chatterjee and Perrimon, 2021). In this study, we revealed an underappreciated role of glucose within pigment glia to regulate their inflammatory responses.

Our study also found that pigment glial cells can function as immune cells in the fly retina, as they are activated in response to Tau-induced lesions. Pigment glia ensheathe each individual ommatidium, the minimum unit of the compound eye, to provide an optical insulation that prevents extraneous light rays from inappropriately activating photoreceptors (Tomlinson, 2012). Pigment glia also accumulates lipid droplets in response to elevated oxidative stress in photoreceptor neurons (Liu et al., 2015), and lipid accumulation has been linked to microglial activation in mammalian microglia (Jung and Mook-Jung, 2020; Marschallinger et al., 2020). Our finding is consistent with the results from a comprehensive single-cell RNA sequence that expression of AMPs overlaps with the pigment glia marker ‘santa-maria’ (Li et al., 2022; Yeung et al., 2022). Further characterization of pigment glia to identify the equivalent mammalian glial cell types will make the Drosophila retina an in vivo platform to study neuroinflammation and neuron-glia interaction.

We found that Tau expression driven by GMR-GAL4 causes swelling of the laminal cortex, and degeneration of surface and cortex glia, in which GMR-GAL4 had not been expressed. These non-cell autonomous effects might be mediated by AMPs, since GLUT expression in the pigment glia suppressed expression of AMPs and mitigated the laminal cortex swelling. Although the primary function of AMPs is to kill bacteria (Ganesan et al., 2011), recent reports suggest that AMPs affects diverse physiological processes, such as sleep and memory formation (Toda et al., 2019; Barajas-Azpeleta et al., 2018), implying that AMPs can work as intercellular signaling molecules. In a fly model of traumatic brain injury, downregulation of the NF-κB-signaling pathway in response to mutation of Relish or loss of Mtk protects against detrimental effects (Swanson et al., 2020). Secreted AMPs from pigment cells in Tau-expressing retina may damage glial cells in the laminal cortex. Another possible mechanism underlying this non-cell autonomous toxicity may be intercellular lipid transfer. During oxidative stress, neurons release lipids that are internalized by glia, thereby causing them to swell (Liu et al., 2015, 2017). Pigment cells primarily accept lipid droplets from photoreceptor neurons (Liu et al., 2015) but − under neurodegenerative conditions − oxidative stress and lipid shuttling from neurons might exceed the capacity of the pigment cells, and overflowed lipids may harm distal glial cells. Enhanced glucose uptake of pigment cells might enhance their ability to degrade lipid droplets from neurons and reduce the overflow of lipid droplets that affect other glia.

Enhanced glucose uptake in pigment cells suppressed expression of AMPs dramatically; however, photoreceptor degeneration was only mildly rescued. Tau affects many cellular components and processes, such as the cytoskeleton, mitochondria, chromatin structures, lipids and stress signaling that, independently, can contribute to neurodegeneration (DuBoff et al., 2012; Frost et al., 2014; Fulga et al., 2007; Goodman et al., 2024; Papanikolopoulou et al., 2019; Voelzmann et al., 2016). Our results suggest that secretion of AMPs from pigment glia in the Tau-expressing retina is one of the events that, collectively, cause degeneration when induced in parallel within multiple cell types.

Changes in glial activity under disease conditions are often associated with alterations in the metabolic signatures, including glucose metabolism pathways (Edison, 2020; Jha and Morrison, 2018; Wang et al., 2019). Multiple signaling pathways that sense metabolic status are conserved across species and are also known to regulate inflammatory responses (Saito et al., 2019). For example, liver kinase B1 (LKB1, officially known as STK11) is a key regulator of metabolism and activates members of the AMP-activated protein kinase (AMPK) and AMPK-related kinase (ARK) families (Compton et al., 2023), which have been reported to promote anti-inflammatory responses (Compton et al., 2023) by negatively regulating nuclear factor κB (NF-κB) signaling in microglia (Ramamurthy and Ronnett, 2006). Another critical energy regulator, Sirtuin 1 (SIRT1), can suppress NF-κB signaling (Zhang et al., 2020). Many regulatory mechanisms of NF-κB signaling are conserved among species (Buchon et al., 2014), and these pathways can also regulate inflammatory responses in Drosophila (Hetru and Hoffmann, 2009). Multiple AMPs downstream of NF-κB signaling are activated upon Tau expression and suppressed upon GLUT3 coexpression. Further studies using metabolic and transcriptomic approaches will facilitate the mechanistic understanding of neuroinflammation.

The following fly stocks were obtained from the Bloomington Drosophila Stock Center (BDRC): UAS-mCherry.NLS (39434), Repo-GAL4 (7415), UAS-mCD8::ChRFP (27391), UAS-Luciferase (35788), UAS-Luciferase RNAi (31603), GMR-Tau P301L (51377), UAS-CD8::GFP (5137), GMR-GAL4 (9146), Rh1-GAL4 (8691), and 54C-GAL4 (27328). UAS-Glut1^d05758, which is a P{XP} insertion containing UAS elements upstream of Glut1, was obtained from Harvard Medical School (http://flybase.org/reports/FBti0055936.htm, Exelixis Drosophila Collection) (Thibault et al., 2004). UAS-drpr RNAi (HMJ30231) and UAS-NimC4 RNAi (HMJ23355) were obtained from the Japanese National Institute of Genetics (NIG-FLY). w¹¹¹⁸ and UAS-GLUT3 were a gift from Marie Thérèse Besson (Aix-Marseille Université, Marseille, France) (Besson et al., 2015). UAS-Tau (wildtype 0N4R) was a gift from Mel. B. Feany (Brigham and Women's Hospital and Harvard Medical School, Boston, USA) (Wittmann et al., 2001). Flies were reared in a standard medium containing 10% glucose, 0.7% agar, 9% cornmeal, 4% brewer's yeast, 0.3% propionic acid, and 0.1% N-butyl p-hydroxybenzoate [w/v]. Flies were maintained at 25°C under light-dark cycles of 12:12 h. Experiments involving transgenic Drosophila were approved by the Tokyo Metropolitan University research ethical committee (#G5-15). Fly genotypes for each experiment are listed in Tables S1 and S2.

Images of fly eyes were captured with an Olympus digital microscope DSX110 or the Leica MZ16 stereomicroscope, and the area of the eye surface was measured using Fiji (NIH). Eyes from more than 11 flies were analyzed for each genotype.

Flies were collected at 3-5 days old, and brains were dissected in PBS, fixed in 4% PFA/PBS (Thermo Scientific) for 30 min at room temperature (RT). Brains were washed with PBST for 10 min three times. Normal donkey serum at 4% (abcam) was used as a blocking buffer for 30 min at RT. Brains were stained with antibody against Drosophila elav (1:1000, DSHB cat# Elav-9F8A9, RRID:AB_528217), Repo (1:50, DSHB cat# 8D12, RRID:AB_528448) or ubiquitin (FK2, 1:1000, StressMarq Biosciences cat# SMC-214D-DY405, RRID:AB_2820835) and visualized with goat anti-rat antibody conjugated to Alexa Fluor 488 (1:1000, Thermo Fisher Scientific cat# A-11006, RRID:AB_2534074) or goat anti-mouse antibody conjugated to Alexa Fluor 488 (1:1000, (Thermo Fisher Scientific cat# A-11001, RRID:AB_2534069). Nuclei were stained with DAPI (1:1000, abcam). Filamentous actin was stained with Alexa Flour 488 Phalloidin (1:1000, Invitrogen). Brains were mounted in VectaShield (Vector Laboratories) and imaged using Nikon spinning-disc confocal microscope.

Flies were frozen in liquid nitrogen, and fifteen fly heads per genotype were homogenized in SDS-Tris-Glycine sample buffer. The same amount of the lysate was loaded to each lane of multiple 10% Tris-Glycine gels and transferred to PVDF membrane. Membranes were blocked with 5% skimmed milk in tris-buffered saline with 0.1% Tween 20 (TBST), blotted using the antibodies described below, incubated with appropriate secondary antibodies and detected using Immobilon Western Chemiluminescent HRP Substrate (Merck Millipore). Membranes were probed with anti-actin antibody (Sigma) as the loading control. Antibodies against Tau pS202 (CP13) and phosphorylated-Ser396/404-Tau (PHF1) were a gift from Peter Davis (Albert Einstein College of Medicine, New York, USA). Antibodies against Tau (T46, Thermo Fisher Scientific cat# 13-6400, RRID:AB_2533025), posphorylated-Ser262-Tau (pSer262, abcam, cat# ab92627, RRID:AB_10563129) rhodopsin 1 (4c5, DSHB cat# 4c5, RRID:AB_528451) and actin (Sigma-Aldrich cat# A2066, RRID:AB_476693) were purchased. The chemiluminescent signals were detected by using Fusion FX software (Vilber) and intensity was quantified with Fiji (NIH). Western blots were repeated at least three times with different animals, and representative blots are shown. Flies used for western blotting were 5-7 days old after eclosion.

Heads of flies aged 3-5 days after eclosion were fixed in Bouin's fixative for 48 h at RT and incubated for 24 h in a leaching buffer (50 mM Tris/150 mM NaCl). Samples were dehydrated in a series of ethanol baths (70%, 80%, 95% and 100%) and xylene at RT, and embedded in paraffin. Serial sections (6 μm thick) through the entire heads were stained with hematoxylin and eosin (H&E). Sections were mounted to glass slides, de-paraffined in xylene and a series of ethanol baths (100%, 95%, 80%) for 3 min each. Slides were stained with 100% hematoxylin for 2 min and washed in tap water for 5 min. They were then stained with 100% eosin for 30-45 s, washed with 95% ethanol and 100% ethanol for 6 min each, incubated with xylene for 10 min, dried and mounted with a coverslip. Each section was examined using bright-field microscopy. Images of the sections that include the lamina were taken using a BZ-X700 (Keyence) or Mica (Leica) microscope. Retinal degeneration or the vacuole area was analyzed using Fiji (NIH). The area of the laminal cortex was measured and normalized to the whole laminal area to calculate the laminal cortex area.

Decapitated heads were cut in half vertically, then incubated in primary fixative solution (2.5% glutaraldehyde and 2% paraformaldehyde in 0.1 M sodium cacodylate buffer) at RT for 2 h. After washing heads with 3% sucrose in 0.1 M sodium cacodylate buffer, fly heads were post-fixed for 1 h in secondary fixation (1% osmium tetroxide in 0.1 M sodium cacodylate buffer) on ice. After washing with H₂O, heads were dehydrated with ethanol and infiltrated with propylene oxide and Epon mixture (TAAB and Nissin EM) for 3 h. After infiltration, specimens were embedded with an Epon mixture at 70°C for 2∼3 days. Some longitudinal sections cut out by glass knives were stained with toluidine blue to check the position. Thin sections (70 nm) of retinal sections were collected on copper grids. The sections were stained with 2% uranyl acetate in 70% ethanol and Reynolds' lead citrate solution. Electron micrographs were obtained with a VELETA CCD Camera (Olympus Soft Imaging Solutions GMBH) mounted on a JEM-1010 electron microscope (Jeol Ltd.).

Heads from more than 25 flies were mechanically isolated, and total RNA was extracted using ISOGEN (NipponGene), followed by reverse transcription using the PrimeScript RT reagent kit (Takara). The resulting cDNA was used as a template for PCR with THUNDERBIRD SYBR qPCR mix (TOYOBO) on a Thermal Cycler Dice real-time system TP800 (Takara). Expression of genes of interest was standardized relative to rp49 or actin. Relative expression values were determined by the ΔΔCT method (Livak and Schmittgen, 2001). Experiments were repeated three times, and a representative result was shown. Primers were designed using DRSC FlyPrimerBank (Harvard Medical School). Primer sequences are: Attacin C for 5′-CTGCACTGGACTACTCCCACATCA-3′ and Attacin C rev 5′-CGATCCTGCGACTGCCAAAGATTG-3′, Cecropin A1 for 5′-CATTGGACAATCGGAAGCTGGGTG-3′ and Cecropin A1 rev 5′-TAATCATCGTGGTCAACCTCGGGC-3′, Diptericin B for 5′-AGGATTCGATCTGAGCCTCAACGG-3′ and Diptericin B rev 5′-TGAAGGTATACACTCCACCGGCTC-3′, Drosomycin for 5′-AGTACTTGTTCGCCCTCTTCGCTG-3′ and Drosomycin rev 5′-CCTTGTATCTTCCGGACAGGCAGT-3′, Metchnikowin for 5′-CATCAATCAATTCCCGCCACCGAG-3′ and Metchnikowin rev 5′-AAATGGGTCCCTGGTGACGATGAG-3′, drpr for 5′-GCAGATGCCTGAATAACTCCTC-3′ and drpr rev 5′-TCCTTGCATTCCATGCCGTAG-3′, NimC4 for 5′-GAACGAGACGATACGAGCCAC-3′ and NimC4 rev 5′-GGTGACTTGTTCCTCCTCTGA-3′, Glut1 for 5′-TTACCGCGGAGCTCTTCTCC-3′ and Glut1 rev 5′-GCCATCCAGTTGACCAGCAC-3′, rp49 for 5′-GCTAAGCTGTCGCACAAATG-3′ and rp49 rev 5′-GTTCGATCCGTAACCGATGT-3′, actin 5C for 5′-TGCACCGCAAGTGCTTCTAA-3′ and actin 5C rev 5′-TGCTGCACTCCAAACTTCCA-3′.

We thank the Japanese National Institute of Genetics (NIG-FLY), Bloomington Drosophila Stock Center (BDSC), and Vienna Drosophila Resource Center (VDRC) for fly strains. We thank Dr Shin-Ichi Hisanaga (Tokyo Metropolitan University) for critical comments and Drs Taro Saito and Akiko Asada (Tokyo Metropolitan University) for technical support. We thank Drs Mel B. Feany (Harvard University) for UAS-Tau and Marie Thérèse Besson (Aix-Marseille Université) for UAS-GLUT3. We also thank Dr Ismael Al-Ramahi (Baylor College of Medicine) for help with histological studies.