Dysfunction of astrocytic glycophagy exacerbates reperfusion injury in ischemic stroke

Haiyun Guo; Yumeng Li; Shiquan Wang; Yongheng Yang; Tiantian Xu; Jianshuai Zhao; Jin Wang; Wenqiang Zuo; Pengju Wang; Guangchao Zhao; Huaning Wang; Wugang Hou; Hailong Dong; Yanhui Cai

doi:10.1016/j.redox.2024.103234

. 2024 Jun 7;74:103234. doi: 10.1016/j.redox.2024.103234

Dysfunction of astrocytic glycophagy exacerbates reperfusion injury in ischemic stroke

Haiyun Guo ^a,¹, Yumeng Li ^a,¹, Shiquan Wang ^a,¹, Yongheng Yang ^a, Tiantian Xu ^a, Jianshuai Zhao ^a, Jin Wang ^a, Wenqiang Zuo ^a, Pengju Wang ^b, Guangchao Zhao ^a, Huaning Wang ^c, Wugang Hou ^a,^⁎⁎⁎, Hailong Dong ^a,^⁎⁎, Yanhui Cai ^a,^c,^⁎

PMCID: PMC11215420 PMID: 38861834

Abstract

Glycophagy has evolved from an alternative glycogen degradation pathway into a multifaceted pivot to regulate cellular metabolic hemostasis in peripheral tissues. However, the pattern of glycophagy in the brain and its potential therapeutic impact on ischemic stroke remain unknown. Here, we observed that the dysfunction of astrocytic glycophagy was caused by the downregulation of the GABA type A receptor-associated protein like 1 (GABARAPL1) during reperfusion in ischemic stroke patients and mice. PI3K-Akt pathway activation is involved in driving GABARAPL1 downregulation during cerebral reperfusion. Moreover, glycophagy dysfunction-induced glucosamine deficiency suppresses the nuclear translocation of specificity protein 1 and TATA binding protein, the transcription factors for GABARAPL1, by decreasing their O-GlcNAcylation levels, and accordingly feedback inhibits GABARAPL1 in astrocytes during reperfusion. Restoring astrocytic glycophagy by overexpressing GABARAPL1 decreases DNA damage and oxidative injury in astrocytes and improves the survival of surrounding neurons during reperfusion. In addition, a hypocaloric diet in the acute phase after cerebral reperfusion can enhance astrocytic glycophagic flux and accelerate neurological recovery. In summary, glycophagy in the brain links autophagy, metabolism, and epigenetics together, and glycophagy dysfunction exacerbates reperfusion injury after ischemic stroke.

Keywords: Glycophagy, Ischemic stroke, Astrocyte, GABARAPL1, O-GlcNAcylation, Oxidative stress

Graphical abstract

Highlights

•
Brain glycophagy is dysfunctional due to GABARAPL1 downregulation after reperfusion.
•
Astrocytic PI3K-Akt pathway takes part in driving GABARAPL1 downregulation.
•
Glycophagy-induced glucosamine decline aggravates its dysfunction by glycosylation.
•
Glycophagy restoration is neuroprotective for brain reperfusion injury.
•
Hypocaloric diet helps brain recovery by promoting glycophagy after reperfusion.

1. Introduction

Ischemic stroke is the leading cause of disability and mortality among adults and has limited therapeutic strategies [1]. Thrombolysis is the main effective therapeutic strategy used to salvage cells in the ischemic penumbra but inevitably increases the incidence of hemorrhagic transformation and exacerbates cerebral damage [2,3], a phenomenon called ischemia/reperfusion (I/R) injury. Blood deficiency and resupply during I/R disrupt cerebral energy homeostasis, which exacerbates free radical formation, inflammation, calcium overload, and glutamate excitotoxicity, known as the classical mechanisms underlying I/R injury [4,5]. There is an urgent need for novel strategies aimed at alleviating reperfusion injury for patients suffering from ischemic stroke.

Glycophagy is a noncanonical pathway that decomposes glycogen into glucose via lysosomes [6,7]. Recent research has indicated that glycophagy is crucial for maintaining myocardial energy balance and is not simply a redundant “alternative” degradation route for glycogen [8]. For instance, glycophagy modulates cellular redox homeostasis, glutamine utilization, and fatty acid oxidation in the skeletal muscle [9]. Furthermore, glycophagy drives lipid droplet formation in adipocytes [10]. Glycophagic dysfunction is involved in many diseases, including Lafora disease and Pompe disease [11,12], and exacerbates diabetic cardiomyopathy damage [13,14]. However, alterations in brain glycophagy during reperfusion and its association with the prognosis of ischemic stroke remain unknown.

In contrast to muscle, heart, and liver, brain glycogen contains approximately 25 % glucosamine [15]. It can degrade glycogen into glucosamine-6-phosphate by glycogenolysis [15], which suggests that brain glycophagy may participate in regulating protein posttranslational modification since glucosamine is an important donor for cellular glycosylation. In addition, brain glycogen predominantly exists in astrocytes but not in neurons [16], and many metabolic interactions, such as lactate shuttle, glutamate cycle, and glutathione (GSH) shuttle, occur between astrocytes and neurons [17]. Astrocytic glycophagy may not only enhance the resistance of astrocytes but also help neighboring neurons when the brain as a whole suffers from energy stress. Accordingly, brain glycophagy might have more nonmetabolic characteristics than peripheral glycophagy and deserves greater attention.

Here, we found that astrocytic glycophagy was dysfunctional due to the downregulation of GABA type A receptor-associated protein like 1 (GABARAPL1) after cerebral reperfusion in humans and mice. In addition, brain I/R-induced phosphoinositide 3-kinase (PI3K)-protein kinase B (Akt) pathway activation was involved in driving GABARAPL1 downregulation in astrocytes. Astrocytic glycophagic dysfunction was exacerbated due to an unexpected brain-specific positive feedback loop. Restoring brain glycophagy improved the outcome of reperfusion injury by reducing DNA damage and oxidative injury in astrocytes and accelerating the survival of surrounding neurons. In addition, a hypocaloric diet could enhance astrocytic glycophagic flux and improve neurological outcomes during the acute phase after cerebral reperfusion.

2. Materials and methods

2.1. Human samples

Human tissues were obtained from the China Brain Bank, Zhejiang University (Hangzhou, China), as previously reported [18] and used following the ethical standards of the Zhejiang University Ethics Committee (ID: 2019-001) and in accordance with the World Medical Association's Declaration of Helsinki. The patient information was shown in Fig. S2. Informed consent was obtained from the participants before inclusion in the study. The brain tissues of stroke patients were cut to identify the different regions, including the core infarction area and penumbra area, and then frozen in liquid nitrogen for metabolic analysis or fixed with paraformaldehyde for immunofluorescence and glutaraldehyde for electron microscopy analysis. The postmortem interval (PMI) was approximately 30–40 min, and the time for dissection to obtain the penumbra tissue was approximately 20 min.

2.2. Mouse

Four-week-old male C57BL/6J mice, one-day-old neonatal C57BL/6J pups, and embryonic Day 15–16 female C57BL/6J mice were purchased from the Experimental Animal Center of the Fourth Military Medical University. ALDH1L1-CreERT2 and GABARAPL1-loxP mice were generated by Cyagen (Santa Clara, USA). The Animal Care Committee of the Fourth Military Medical University approved the mouse experiments. The mice were housed in standard breeding cages with ad libitum access to water and food in groups of four at 23 ± 1 °C.

2.3. Hypocaloric diet

The normocaloric diet (1314, 3339 kcal/kg) and hypocaloric diet (C1012, 1303 kcal/kg) for mice were obtained from Altromin International (Germany). The hypocaloric diet was maintained for one week immediately post-MCAO operation and subsequently changed to a normocaloric diet in the second week post-MCAO operation.

2.4. Primary astrocyte and neuron culture

One-day-old neonatal C57BL/6J pups were sterilized with 75 % ethanol and decapitated to obtain primary astrocytes. The cortices were removed from the skulls and placed on ice. The meninges were peeled from the brain under a microscope on ice, and the tissues were digested with trypsin (Thermo Fisher Scientific, 25200056) at 37 °C for 10 min. The trypsin activity was terminated with Dulbecco's modified Eagle's medium (DMEM, 5.56 mM glucose, Thermo Fisher Scientific, 11885092) supplemented with 15 % fetal bovine serum (Thermo Fisher Scientific, 16140071). The mixture was filtered through a sterile 200-mesh screen. The filtrate was centrifuged at 300×g for 10 min, and the cells were seeded onto poly-d-lysine-coated flasks (0.1 mg/mL, Sigma‒Aldrich, P0296) at a density of 10,000 cells/cm². The cells were maintained for seven days, and the culture medium was changed every three days. The cells were then shaken for 19 h at 190 rpm at 37 °C to remove oligodendrocytes and microglia.

Primary neurons were obtained from the cortices of embryonic Day 15–16 C57BL/6J embryos. The procedure was the same as that used to prepare primary astrocytes, except digestion time for brain tissues after removing the meninges was 5 min at 37 °C. The culture medium used for the neurons consisted of glucose-free neurobasal medium (Thermo Fisher Scientific, A2477501) supplemented with B-27 (2 %, Thermo Fisher Scientific, 17504044), glucose (5.56 mM, Thermo Fisher Scientific, A2494001), and glutamine (1 %, Thermo Fisher Scientific, 35050061).

2.5. Astrocyte–neuron coculture system

An astrocyte-neuron coculture system was established using hanging inserts with porous membranes (3 μm, Millipore, PTSP24H48). The coculture medium comprised a glucose-free neurobasal medium containing 15 % fetal bovine serum, 2 % B27, 5.56 mM glucose, and 1 % glutamine.

2.6. Low-glucose culture

Glucose-free DMEM (Thermo Fisher Scientific, A1443001) supplemented with 2.5 mM glucose and 15 % fetal bovine serum was used to mimic the low-glucose environment of cultured astrocytes. Glucose-free neurobasal medium supplemented with 15 % fetal bovine serum, 2 % B27, 2.5 mM glucose and 1 % glutamine was used to mimic the low-glucose environment for cocultured neurons.

2.7. I/R model in mice

A middle cerebral artery occlusion/reperfusion (MCAO/R) model was established to mimic I/R injury in mice. Eight-week-old male C57BL/6J mice were anesthetized using 1.4 % isoflurane. A small incision was made on the skin of each mouse at the midline of the anterior neck, and the muscles were dissected under a surgical microscope. Subsequently, an MCAO monofilament (RWD Life Sciences, MSMC21B100PK50) was inserted through a small incision into the right internal carotid artery and extended into the right middle cerebral artery. The filament was withdrawn after 60 min. The penumbra in the mouse was determined according to our previous study [19]. Briefly, the brain of each mouse was divided into three sections from the anterior tip of the frontal lobe. The first section and the third section were 2 mm thick. The second section was 4 mm thick. Then, we identified the midline between the two hemispheres in the second section and made a longitudinal cut about 1 mm from the midline in the right hemisphere. We generated a transverse diagonal cut at approximately “2 o'clock” to separate the core from the penumbra. The region from “0 o'clock” to “2 o'clock” was identified as the penumbra.

2.8. In vitro I/R model

In this study, an oxygen-glucose deprivation/reoxygenation (OGD/R) model was established to mimic I/R in vitro using a humidified hermetic chamber (Embrient Inc). The culture medium was replaced with a glucose-free neurobasal medium for cocultures and glucose-free DMEM for astrocytes. The cells were transferred to a chamber containing an anaerobic gas mixture (5 % CO₂ and 95 % N₂). The duration of OGD was 2 h. The glucose-free medium was replaced with a complete culture medium after OGD.

2.9. Neurobehavioral analysis

Neurobehavioral analysis, including the corner test, grid-walking test and adhesive removal test, was performed according to our previous studies [18,19].

2.10. Immunofluorescence and immunocytochemistry staining

Mice were deeply anesthetized via the intraperitoneal administration of pentobarbital sodium (50 mg/kg) and transcardially perfused with saline, followed by 4 % paraformaldehyde (Sigma-Aldrich, 158127) in phosphate-buffered saline (PBS). The brain tissues of mice and humans were postfixed overnight with 4 % paraformaldehyde and cryoprotected with 20 % and 30 % sucrose solutions (Sigma-Aldrich, S9378) in PBS for one day each. Subsequently, the mouse brains were frozen, and coronal brain sections (12 μm) were collected on slides using a cryostat and blocked with serum. The incubation time was 12 h at 4 °C for primary antibodies and 2 h at room temperature for secondary antibodies. DAPI (300 nM, Thermo Fisher Scientific, D3571) was used to visualize the cell nuclei, and images were captured using a confocal microscope (Olympus). The fluorescence intensity of the colocalization area was calculated using ImageJ software.

For in vitro immunocytochemistry analysis, cultured astrocyte slices were fixed overnight with 4 % paraformaldehyde and permeabilized with 0.1 % Triton X-100 (Sigma-Aldrich, T8787) for 10 min. The slices were then incubated with 1 % bovine serum albumin (BSA, Sigma-Aldrich, A1933) for 1 h. The other procedure was the same as that for immunofluorescence.

2.11. Immunoblotting

Immunoblotting was performed according to our previous study [18].

2.12. RT-qPCR

RT-qPCR was performed according to our previous study [18]. The primers and corresponding accession numbers used for RT-qPCR are shown in Table S1.

2.13. Triphenyl tetrazolium chloride (TTC) staining

The mice were decapitated, and the cerebrum was removed from the skulls. Brain tissues were continuously cut into coronal slices at 1 mm intervals and immersed in 2 % TTC solution (02103126-CF, MP Biomedicals) for 20 min at 37 °C. TTC-stained red areas indicate nonischemic regions, whereas white areas indicate ischemic regions. The infarct volume was calculated as the following ratio: (area of the contralateral hemisphere - nonlesion area in the ipsilateral hemisphere)/area of the contralateral hemisphere.

2.14. Golgi staining

Golgi staining of the neurons in the penumbra was conducted using an FD Rapid Golgi Stain Kit (PK401, FD NeuroTechnologies). The number of branches was calculated at each order away from the cell body. The dendritic spine number was traced, and dendrites 10 μm in length were counted.

2.15. Electron microscopy analysis

The brain tissues from human, mouse or astrocyte pellets were fixed with 2 % glutaraldehyde in 0.1 M sodium cacodylate buffer for 1 h and postfixed with 1 % OsO₄ (Sigma-Aldrich, 419494) in 4 % paraformaldehyde buffer at 4 °C for 1 h. After dehydration with graded ethanol, the samples were embedded, heated at 60 °C for 48 h and cut into 50 nm ultrathin sections with an ultramicrotome (LKB Inc). Ultrathin slices were stained with saturated lead citrate and uranyl acetate and observed using a transmission electron microscope (JEOL Ltd).

The glycophagy-related vacuole in the electron microscopy is recognized as an autophagic vacuole that contains many glycogen granules [10,20,21]. The glycophagy-related vacuole includes the glycophagosome and glycophagosome-lysosome fusion vacuole. The glycophagosome is recognized as an autophagic vacuole with glycogen granules at low electron density. The glycophagosome-lysosome fusion vacuole is recognized as an autophagic vacuole with glycogen granules at high electron density because lysosomes are reported to be round- and oval-shaped structures with moderate or high electron density [22]. The mitophagy-related vacuole is recognized as an autophagic vacuole that contains mitochondria [23].

2.16. Metabolite analysis

Glucosamine levels in cultured astrocytes, cocultured neurons, and coculture medium were measured using the Glucosamine Assay Kit (Megazyme, K-GAMINE). Glucose levels in cultured astrocytes were measured using a Glucose Assay Kit (Abcam, ab65333). Glucose-6-phosphate (G6P) levels in cultured astrocytes were measured using a G6P Assay Kit (Abcam, ab83426). NADPH levels in cultured astrocytes and cocultured neurons were measured using an NADPH Assay Kit (Abcam, ab176724). GSH levels in cultured astrocytes, cocultured neurons, and coculture medium were measured using a GSH Assay Kit (Abcam, ab65322). Reactive oxygen species (ROS) levels in cultured astrocytes and cocultured neurons were measured using a ROS detection assay kit (Abcam, ab287839). Phosphoribosyl diphosphate (PRPP) levels in cultured astrocytes and cocultured neurons were measured using a PRPP Assay Kit (Novocib, K0709-04-2). The UDP-GlcNAc levels in cultured astrocytes were measured using a UDP-GlcNAc ELISA Kit (Meikebio, MK0034MA).

2.17. Glycogen levels in total cells

Human brain tissues were rapidly immersed in liquid nitrogen after being cut to identify the penumbra region and the core infarction area. The mice were decapitated without anesthesia, and brain tissues in the penumbra were frozen in liquid nitrogen. The brain tissues from humans and mice were homogenized with 200 μL of 30 % KOH on ice, and the homogenates were then boiled for 10 min to inactivate the enzymes. The boiled samples were centrifuged at 12,000 rpm for 10 min at 4 °C to remove insoluble materials, and the supernatant was ready for the assay using the Glycogen Assay Kit (Abcam, ab169558). For the cultured astrocytes, the culture medium was discarded, and the cells were washed with PBS buffer, lysed with 30 % KOH, and scraped into Eppendorf tubes. The samples were boiled and centrifuged at 12,000 rpm for 10 min at °C to remove insoluble materials, and the supernatant was ready for the assay using the Glycogen Assay Kit.

2.18. Glycogen levels in the lysosome

Lysosomes in cultured astrocytes and the penumbra region of the mouse and human brains were obtained using a Lysosome Enrichment Kit (Thermo Fisher Scientific, 89839). Briefly, the penumbra tissue was washed with PBS and minced into small pieces (<3 mm³). For the cultured astrocytes, the cells were harvested by a cell scraper and centrifuged at 850×g for 2 min and the supernatant was discarded. Then, 800 μL of Lysosome Enrichment Reagent A was mixed with the tissues or cell pellets, and the mixture was homogenized on ice. Then, 800 μL of Lysosome Enrichment Reagent B was added to the mixture, and the tube was centrifuged at 500×g for 10 min at 4 °C. The supernatant was collected, and the protein concentration was determined with a BCA Protein Assay Kit. A discontinuous density gradient was prepared in an ultracentrifuge tube by carefully overlaying the prepared OptiPrep Gradients in descending concentrations. The sample supernatant was mixed with OptiPrep Cell Separation Media to a final concentration of 15 % OptiPrep Media, and the mixture was overlaid on top of the density gradients in the ultracentrifuge tube. Then, the mixture was ultracentrifuged at 145,000×g for 2 h at 4 °C (Beckman). The lysosomal band, which was located in the top 2 mL of the gradient, was removed, and the sample was placed on ice. The isolated lysosome fraction was mixed with PBS and centrifuged at 18,000×g for 30 min at 4 °C. The supernatant was removed, and the lysosomal pellet was surface-washed with gradient dilution buffer and centrifuged at 18,000×g for 30 min at 4 °C. The supernatant was removed, and the lysosomal pellet was lysed with 30 % KOH on ice. The samples were boiled and then centrifuged at 12,000 rpm for 10 min at 4 °C to remove insoluble materials, and the supernatant was subjected to the assay using a Glycogen Assay Kit (Abcam, ab169558).

2.19. Enzyme activity analysis

Adherent astrocytes were grown to 90 % confluence and dissociated using lysis buffer. Lysis was performed on ice for 30 min to achieve complete dissolution in the lysis buffer. The lysate was centrifuged at 13,000×g for 15 min at 4 °C, and the supernatant was used for further analysis. Acid α-glucosidase (GAA) activity in cultured astrocytes, rGAA, and MACS-selected astrocytes was measured using a GAA Activity Assay Kit (Abcam, ab252887). γ-glutamyl transpeptidase (γ-GT) activity in cultured astrocytes was measured using a γ-GT Activity Assay Kit (Elabscience, E-BC-K126-M).

2.20. Nuclear and cytoplasmic protein extraction

Nuclear and cytoplasmic proteins in the cultured astrocytes were separated and extracted using a Nuclear and Cytoplasmic Extraction Kit (78833, Thermo Fisher Scientific). The astrocytes were harvested with trypsin, and Cytoplasmic Extraction Reagent-Ⅰ was added to 10⁷ cells, followed by vortexing for 15 s and incubation on ice for 10 min. Then, Cytoplasmic Extraction Reagent-Ⅱ was added to the mixture, followed by vortexing for 5 s and incubation on ice for 1 min. The mixture was centrifuged at 16,000 rpm for 5 min at 4 °C, and the supernatant was collected for cytoplasmic extraction. Next, the insoluble pellet was washed with PBS three times, suspended in Nuclear Extraction Reagent, vortexed for 15 s and incubated on ice for 40 min. The mixture was centrifuged at 16,000 rpm for 5 min at 4 °C, and the supernatant was collected for nuclear extraction.

2.21. Immunoprecipitation

Immunoprecipitation was performed using the Pierce Co Immunoprecipitation Kit (26149, Thermo Fisher Scientific).

2.22. Magnetic-activated cell sorting (MACS)

Astrocytes in the penumbra of the mouse and human brain were separated using an Adult Brain Dissociation Kit (Miltenyi Biotec, 130-107-677) and an Anti-ACSA-2 MicroBead Kit (Miltenyi Biotec, 130-097-678). ACSA-2 is specifically expressed in astrocytes [24].

2.23. O-GlcNAcylation analysis

The cultured astrocytes were harvested with trypsin and centrifuged. The cell pellet was immersed in liquid nitrogen and sent to PTM Bio (Hangzhou, China) for O-GlcNAcylation analysis.

2.24. Comet assay

A Comet Assay Kit (Abcam, ab238544) was used to detect DNA damage in cultured astrocytes.

2.25. Reverse chromatin immunoprecipitation (ChiP)-mass spectrometry (MS)

The TFs of the GABARAPL1 promoter were determined using a Reverse ChiP Kit (Bersin Bio, Bes5005, http://www.bersinbio.com/). The samples from reverse ChiP underwent mass spectrometric analysis by Gene Denovo Biotechnology Co (Guangzhou, China). The eluted DNA was extracted using a DNA Extraction Kit (Thermo Fisher Scientific, K182002), and PCR was performed to verify the eluted GABARAPL1 promoter nucleotide sequence using DNA Polymerase (Takara Bio Inc., RR370A). Table S1 presents a list of the primers used for PCR. Agarose gel electrophoresis was performed to analyze the PCR products using UltraPure™ Agarose (Thermo Fisher Scientific, 16550100).

2.26. Metabolomics analysis

Cultured astrocytes were digested with trypsin, and cell numbers were determined. In total, 10⁷ astrocytes were transferred into 500 μL of quenching buffer containing 8.5 mg/mL ammonium bicarbonate (Sigma-Aldrich, A6141), 300 μL of methanol (Sigma-Aldrich, 439193), and 200 μL of dH₂O. The mixture was centrifuged at 1000×g at 4 °C for 5 min and the supernatant was discarded. The cell pellet was frozen in liquid nitrogen and subjected to targeted metabolomic analysis for central carbon metabolism, including glycolysis, the tricarboxylic acid (TCA) cycle, and the pentose phosphate pathway (PPP). The analysis was conducted by Biotree Biotech (Shanghai, China), as previously reported [25].

2.27. ¹³C-labeled glucose metabolic flux analysis

¹³C₆-labeled glucose was obtained from Sigma-Aldrich (389374). The culture medium was replaced with a culture medium containing ¹³C₆-labeled glucose (2 g/L) at 6 h immediately after OGD, and the detection time was 12 h after OGD. The ¹³C-labeled astrocytes were washed with PBS, and quenching buffer (40 % methanol, 40 % acetonitrile (Sigma-Aldrich, 34851) and 20 % dH₂O) was added into the culture plate. The astrocytes were collected with a cell scraper, frozen in liquid nitrogen and sent to Biotree Biotech (Shanghai, China) for metabolic flux analysis, including analysis of glycolysis, the TCA cycle, and the PPP.

2.28. RNA sequencing

Total RNA was extracted from cultured astrocytes using TRIzol reagent (Thermo Fisher Scientific, 15596026). The RNA quality was determined using an Agilent 2100 Bioanalyzer (Agilent Technologies). RNA sequencing was subsequently performed by Gene Denovo Biotechnology Co (Guangzhou, China).

2.29. GABARAPL1 overexpression model in vitro

A lentivirus expressing GABARAPL1 (PubMed Gene ID: 57436) was constructed by GeneChem Ltd (Shanghai, China). Before infection, the astrocytes were cultured until they reached 70 % confluence. The lentiviruses were added to the culture medium at a multiplicity of infection (MOI) of 100 for 48 h. Successful lentivirus-mediated upregulation of GABARAPL1 expression was confirmed by immunoblotting.

2.30. GABARAPL1 knockdown model in vitro

A lentivirus with siRNA (TGGCCAGTTCTACTTCTTAAT) against the GABARAPL1 gene (PubMed Gene ID: 57436) was constructed by GeneChem Ltd. Successful lentivirus-mediated silencing of GABARAPL1 expression was confirmed by immunoblotting.

2.31. GAA knockdown model in vitro

A lentivirus with siRNA (CAAGAACAATACCATTGTGAA) against the GAA gene (PubMed Gene ID: 14387) was constructed by GeneChem Ltd. Successful lentivirus-mediated silencing of GAA expression was confirmed by immunoblotting.

2.32. GAA knockdown mouse model

AAVs with an astrocyte-specific GFAP promoter used to drive short hairpin RNA (shRNA) containing 21 nucleotides (CAAGAGCGTTGTGCAACAATA) specific for the GAA gene (PubMed Gene ID: 14387) followed by a FLAG tag were purchased from GeneChem Ltd. Two microliters of AAV (5 × 10¹² v.g./mL) were individually injected into the lateral ventricle and striatum three weeks before MCAO. The stereotactic injection was performed according to our previous studies [18]. The stereotaxic apparatus coordinates were as follows: 1.4 mm deep, 1.0 mm to the left of the midline and 0.22 mm posterior to the bregma for intracerebroventricular injection. 3.0 mm deep, 2 mm to the right of midline and 0.3 mm anterior to the bregma for right striatum injection. Two microliters of AAV were infused into the brain over 20 min using a Hamilton syringe. Successful AAV-mediated silencing of GAA expression was confirmed using immunofluorescence.

2.33. In vitro γ-GT knockdown models

A lentivirus with a siRNA (TCCCTCAATCATCCTGGATAA) against the γ-GT gene (PubMed Gene ID: 14598) was constructed by GeneChem Ltd. Successful lentivirus-mediated silencing of γ-GT expression was confirmed by immunoblotting.

2.34. In vitro specificity protein 1 (SP1) amino acid mutant model and TATA binding protein (TBP) amino acid mutant model

The SP1 amino acid mutant models (S493^A, S613^A, T641^A, S642^A, S699^A, and S703^A) and TBP amino acid mutant models (T113^A and S238^A) were constructed in two steps. First, lentiviruses targeting the CRISPR-Cas9 system with a specific sgRNA against SP1 (GGCACCCTGTGAAAGTTGTG, PubMed Gene ID: 20683) and a specific sgRNA against TBP (GGAGTAAGTCCTGTGCCGTA, PubMed Gene ID: 21374) were constructed by GeneChem Ltd, which were added to cultured astrocytes at an MOI of 100 for 48 h to knock out the SP1 and TBP genes, respectively. Next, lentiviruses driving the expression of mutant SP1 (S493^A, S613^A, T641^A, S642^A, S699^A, and S703^A) and mutant TBP (T113^A and S238^A) were constructed by GeneChem Ltd, which were added to the culture media of SP1-knockout astrocytes and TBP-knockout astrocytes, respectively, for 48 h at an MOI of 100.

2.35. In vitro rGAA eukaryotic expression system

The lysosome-guided signal peptide of GAA (Gene ID: 14387) was replaced with an extracellular-guided signal peptide (base sequence: ATGGAGACAGACACACTCCTGCTATGGGTACTGCTGCTCTGGGTTCCAGGTTCCACTGGTGAC) followed by a His tag, which was constructed using the GV658 plasmid by GeneChem Ltd. The rGAA plasmid was transfected into 293F suspension cells, and rGAA in serum-free culture medium was collected and purified using the Dynabeads™ His-Tag Isolation and Pulldown Kit (Thermo Fisher Scientific, 10104D) seven days after transfection. The purified proteins were confirmed by SDS-PAGE and immunoblotting. The activity of purified rGAA was confirmed by a GAA activity assay kit.

Lysosomes were obtained from cultured astrocytes using a Lysosome Enrichment Kit (Thermo Fisher Scientific, 89839). Lysosomes were lysed with 30 % KOH, boiled for 10 min, and centrifuged at 12,000 rpm for 10 min at 4 °C to remove insoluble materials. The pH of the supernatant was adjusted to 4.0 to mimic the acidic conditions of the lysosome. rGAA was added to the supernatant to degrade glycogen in the lysosome, which was achieved by incubation at 37 °C for 1 h. The reaction was stopped by boiling for 10 min, and the glucose and glucosamine levels were determined using a Glucose Assay Kit and a Glucosamine Assay Kit, respectively. The glycophagy-derived glucose and glucosamine levels were calculated by subtracting the glucose and glucosamine levels in the boiled rGAA-treated astrocytes from the glucose and glucosamine levels in the rGAA-treated astrocytes.

2.36. Dual-luciferase reporter system

The binding domain between the promoter of GABARAPL1 and the transcription factor SP1 or TBP was predicted using the JASPAR database (https://jaspar.genereg.net/). The GV238 plasmid containing the whole GABARAPL1 promoter sequence (PubMed Gene ID: 57436) or its fragments, followed by firefly luciferase activity, was constructed by GeneChem Ltd. The pcDNA3.1 plasmid overexpressing SP1 (PubMed Gene ID: 20683) or TBP (PubMed Gene ID: 21374) was constructed by GeneChem Ltd. In addition, the pRL-TK plasmid driving Renilla luciferase expression was obtained from Promega (E2241). The above three plasmids were cotransfected into 293T cells for 72 h. We then used the Dual-Luciferase Reporter Assay System (Promega, E1910) to detect firefly luciferase and renilla luciferase activities.

2.37. Cell survival analysis

An LDH Cytotoxicity Assay Kit (Thermo Fisher Scientific, C20301) was used to detect lactate dehydrogenase (LDH) release in cultured astrocytes. A Cell Counting Kit-8 (CCK-8, Sigma-Aldrich, 96992) was used to measure the viability of cocultured neurons. The incubation time was 1 h for the cocultured neurons in the CCK-8 analysis.

2.38. Antibodies

The primary antibodies were as follows: Rabbit anti-MAP2 (Abcam, ab183830, 1:200 for ICC), Rabbit anti-γ-GT (Abcam, ab109427, 1:1000 for IB), Rabbit anti-Lamin B1 (Abcam, ab133741, 1:1000 for IB), Rabbit anti-Flag (Abcam, ab205606, 1:100 for IF), Mouse anti-ALDH1L1 (Abcam, ab56777, 1:100 for IF), Mouse anti-S100β (Proteintech, 66616-1-Ig, 1:100 for IF), Goat anti-α-tubulin (Abcam, ab289875, 1:100 for ICC), Rabbit anti-Akt (Abcam, ab179463, 1:200 for IF, 1:1000 for IB), Rabbit anti-phosphorylated Akt (Abcam, ab192623, 1:200 for IF, 1:1000 for IB), Mouse anti-β-actin (Cell Signaling Technology, 3700, 1:1000 for IB), Rabbit anti-phosphorylated PI3K (Abcam, ab182651, 1:100 for IF, 1:500 for IB), Rabbit anti-PI3K (Abcam, ab191606, 1:200 for IF, 1:1000 for IB), Rabbit anti-β-tubulin (Cell Signaling Technology, 2146, 1:1000 for IB), Rabbit anti-His-tag (Cell Signaling Technology, 12698, 1:1000 for IB), Rabbit anti-STBD1 (Proteintech, 11842-1-AP, 1:1000 for IB, 1:100 for IF), Rabbit anti-GABARAPL1 (Proteintech, 11010-1-AP, 1:1000 for IB, 1:100 for IF), Rabbit anti-GAA (Proteintech, 14367-1-AP, 1:1000 for IB, 1:100 for IF), Rabbit anti-SP1 (Proteintech, 21962-1-AP, 1:1000 for IB, 1:100 for ICC, 4 μg for IP), Rabbit anti-TBP (Proteintech, 22006-1-AP, 1:1000 for IB, 1:100 for ICC, 4 μg for IP), Rabbit anti-BRCA1 (Proteintech, 22362-1-AP, 1:1000 for IB), Rabbit anti-RXRA (Proteintech, 21218-1-AP, 1:2000 for IB), Rabbit anti-CREB1 (Proteintech, 12208-1-AP, 1:2000 for IB), Rabbit anti-c-MYC (Proteintech, 10828-1-AP, 1:2000 for IB), Rabbit anti-FoxO1 (Abcam, ab179450, 1:1000 for IB), Rabbit anti-O-GlcNAcase (Proteintech, 61425, 1:1000 for IB), Rabbit anti-O-GlcNAc transferase (Proteintech, 11576-2-AP, 1:2000 for IB) and Mouse anti-O-Linked N-Acetylglucosamine (PTM Bio, PTM-952, 1:1000 for IB, 1:50 for ICC). IF represents immunofluorescence. ICC represents immunocytochemistry. IB represents immunoblotting. IP represents immunoprecipitation.

2.39. Drug

The PI3K inhibitor LY294002 was obtained from Sigma-Aldrich (440202). The LY294002 was added to the culture medium at a final concentration of 40 μM immediately post-OGD/R.

2.40. Statistics

All the data are presented as the means ± standard deviations (SDs) and were analyzed using IBM SPSS 20.0 software. The data were obtained from at least three replicates, and two-sided statistical tests were performed. Paired sample t-tests were used to analyze differences between paired groups. Independent t-tests were used to analyze the differences between the two groups. One-way analysis of variance (ANOVA) was performed to analyze differences among multiple groups. Repeated-measures analysis was used to assess differences between the groups in neurobehavioural tests. Post hoc comparisons were conducted based on the test results for equality of variance. Statistical significance was set at P < 0.05.

3. Results

3.1. Astrocytic glycophagy is dysfunctional due to GABARAPL1 downregulation during cerebral reperfusion

We first focused on determining the glycophagic flux in the penumbra region in the mouse MCAO/R model. Using electron microscopy, we found that the number of glycophagy-related vacuoles, including glycophagosomes and glycophagosome-lysosome fusion vacuoles, started to decrease at 6 h, reached its lowest at 12 h, and remained stable for at least 48 h post-MCAO/R (Fig. 1A). We then detected the glycogen level in the lysosome to reflect the lysosome-mediated glycogen degradation flux in the penumbra region and observed that the glycogen level in the lysosome reached its lowest level at 12 h post-MCAO/R (Fig. 1B). Along with the decrease in glycophagic flux, brain glycogen accumulated and reached its highest level at 12 h post-MCAO (Fig. 1C). Next, we adopted an in vitro OGD/R model to further confirm astrocytic glycophagic dysfunction in I/R. Using electron microscopy, we observed that the number of astrocytic glycophagy-related vacuoles decreased post-OGD/R treatment (Fig. 1D). In addition, the glycogen level in the lysosomes of cultured astrocytes decreased post-OGD/R treatment (Fig. 1E). The glycogen level in the cultured astrocytes increased post-OGD/R treatment (Fig. 1F).

Fig. 1 — Astrocytic glycophagy is blocked due to GABARAPL1 downregulation during MCAO/R and OGD/R. (A) Electron microscopic analysis of the average number of glycophagy-related vacuoles in astrocytes in the penumbra region of mice following cerebral I/R (n = 8). The average number of glycophagy-related vacuoles in the mouse was determined as the average number of glycophagy-related vacuoles in the ten astrocytes per mouse. The arrows indicate glycophagy-related vacuoles, including glycophagosomes and glycophagosome-lysosome fusion vacuoles. Cyto represents the cytoplasm. Nu represents the nucleus. Scale bars are 2 μm. (B) Glycogen levels in the lysosomes of the penumbra region in mice following I/R (n = 8), which were normalized to the brain tissue protein levels. (C) Glycogen levels in the penumbra region of mice following I/R (n = 8), which were normalized to protein levels in brain tissue. (D) Electron microscopic analysis of the average number of glycophagy-related vacuoles in cultured astrocytes post-OGD/R (n = 8). The average number of glycophagy-related vacuoles in the cultured astrocytes was determined as the average number of glycophagy-related vacuoles in the ten astrocytes per sample. The arrows represent glycophagy-related vacuoles. Scale bars are 2 μm. (E) Glycogen levels in the lysosomes of cultured astrocytes post-OGD/R (n = 8), which were normalized to protein levels in cultured astrocytes. (F) Glycogen levels in cultured astrocytes post-OGD/R (n = 8), which were normalized to the protein level in cultured astrocytes. (G–I) Immunofluorescence analysis of STBD1 (G), GABARAPL1 (H), and GAA (I) expression in astrocytes in the penumbra region of mice following cerebral I/R (n = 8). Astrocytes are marked using S100β. Scale bars are 50 μm. (J) Astrocytic GAA activity in the penumbra region in mice following cerebral I/R (n = 8). Astrocytes in the penumbra region are separated from the mouse brain after stroke using MACS toward the ASCA-2. (K–M) Immunoblotting analysis of STBD1 (K), GABARAPL1 (L), and GAA (M) expression in cultured astrocytes post-OGD/R (n = 8). (N) GAA activity in cultured astrocytes post-OGD/R (n = 8). The data are presented as the means ± SDs. Statistical analyses were performed using one-way ANOVA followed by the LSD post hoc test (A-C and F-N) and the Donnett T3 post hoc test (D and E). *P < 0.05, **P < 0.01, ***P < 0.001.

Previous studies suggested that starch-binding domain-containing protein 1 (STBD1) guides the transport of glycogen granules into glycophagosomes by binding to GABARAPL1, and glycophagosomes combine with lysosomes to decompose glycogen using GAA [6,26]. Subsequently, we performed immunofluorescence analysis and observed that STBD1 and GAA expression remained unchanged, whereas GABARAPL1 was substantially downregulated and reached its lowest level at 12 h post-MCAO/R (Fig. 1G–I and Figs. S1A and B). We also used MACS to sort astrocytes (marked by ACSA-2 [24]) from the penumbra and discovered that astrocytic GAA activity remained unchanged post-MCAO/R (Fig. 1J). In the OGD/R model, immunoblotting (Fig. 1K-M) and RT-qPCR (Figs. S1C–E) revealed that GABARAPL1 expression also decreased and reached its lowest level at 12 h post-OGD/R treatment, whereas STBD1 and GAA expression remained unaffected. GAA activity also did not change after OGD/R treatment (Fig. 1N).

We also obtained brain tissue from 10 patients with ischemic stroke who received thrombolytic treatment within 6 h after the onset of stroke and whose intervals between thrombolysis and death were approximately 12 h (Fig. S2). We used electron microscopy and observed that the number of glycophagy-related vacuoles in the penumbra region decreased at approximately 12 h after cerebral reperfusion in patients with ischemic stroke (Fig. 2A). In addition, the glycogen level in the lysosome in the penumbra region was decreased at approximately 12 h after cerebral reperfusion in stroke patients (Fig. 2B). Along with the decrease in glycophagy, the glycogen level in the penumbra region was upregulated in stroke patients (Fig. 2C). We discovered that GABARAPL1 expression decreased and that STBD1 and GAA expression did not change at approximately 12 h after cerebral reperfusion in stroke patients (Fig. 2D–F). Moreover, the astrocytic GAA activity in the penumbra region was not changed in stroke patients (Fig. 2G).

Fig. 2 — Astrocytic glycophagy is blocked due to GABARAPL1 downregulation in the penumbra region of the human brain during cerebral I/R. (A) Electron microscopic analysis of the average number of glycophagy-related vacuoles in astrocytes in the penumbra region of the human brain at approximately 12 h following cerebral I/R (n = 10). The average number of glycophagy-related vacuoles in humans was determined as the average number of glycophagy-related vacuoles in the ten astrocytes per human. The arrows indicate glycophagy-related vacuoles. Scale bars are 2 μm. “Contra” represents the contralateral hemisphere, whereas “Ipsi” represents the ipsilateral hemisphere. (B) Glycogen levels in the lysosomes of the penumbra region in the human brain at approximately 12 h following cerebral I/R (n = 10), which were normalized to the brain tissue protein levels. (C) Glycogen levels in the penumbra region of the human brain at approximately 12 h following cerebral I/R (n = 10), which were normalized to protein levels in brain tissue. (D–F) Immunofluorescence analysis of STBD1 (D), GABARAPL1 (E), and GAA (F) expression in astrocytes in the penumbra region of the human brain at approximately 12 h following cerebral I/R (n = 10). Astrocytes are marked using S100β. Scale bars are 50 μm. (G) Astrocytic GAA activity in the penumbra region of the human brain at approximately 12 h following cerebral I/R (n = 10). Astrocytes in the penumbra region are separated from the human brain using MACS toward the ASCA-2. The data are presented as the means ± SDs. Statistical analyses were performed using paired t-tests (A–G). ***P < 0.001.

3.2. Astrocytic PI3K-Akt pathway activation participates in driving GABARAPL1 downregulation during cerebral reperfusion

Next, we explored the mechanism underlying I/R-induced GABARAPL1 downregulation. At present, the precise TFs for GABARAPL1 in mouse astrocytes are enigmatic. Therefore, we performed reverse ChiP-MS to investigate the binding interactions between the GABARAPL1 promoter and its TFs (Fig. S3) and discovered 103 GABARAPL1-related TFs in cultured mouse astrocytes (Fig. 3A). Among them, 55 TFs remained unchanged, 19 TFs were upregulated, and 29 TFs were downregulated after OGD/R treatment (Fig. 3A). There are many upstream pathways of glycophagy, such as the cAMP-protein kinase A pathway, PI3K-Akt pathway, insulin-mTOR pathway, Notch pathway, and calcium signaling pathway [6,26]. To reveal the upstream pathways of GABARAPL1 downregulation after OGD/R, RNA sequencing was performed to detect alterations in KEGG pathways in cultured astrocytes (Fig. S4). We observed that the PI3K-Akt pathway, which was activated, had the highest gene ratio among the upregulated KEGG pathways (Fig. 3B and S4). Accordingly, we focused on the PI3K-Akt pathway and immunoblotting analysis was conducted to confirm the ratios of phosphorylated PI3K/PI3K and phosphorylated Akt/Akt, which were both upregulated post-OGD/R treatment in cultured astrocytes (Fig. 3C and D). Combined with the results from reverse ChiP-MS (Fig. 3A) and the PI3K-Akt KEGG pathway graph in GenomeNet (https://www.genome.jp/), only breast cancer 1 protein (BRCA1), forkhead Box O1 (FoxO1), retinoic acid X receptor α (RXRA), cAMP response element-binding protein 1 (CREB1) and Myc proto-oncogene protein (c-MYC) simultaneously work as the TFs for GABARAPL1 and the key downstream nodes for the PI3K-Akt pathway. Accordingly, we performed reverse-ChiP immunoblotting and observed that the binding of BRCA1 to GABARAPL1 promoter was enhanced after OGD/R treatment (Fig. 3E). For FoxO1, CREB1 and c-MYC, the combination decreased, and RXRA remained unchanged after OGD/R treatment (Fig. 3F–I). The PI3K inhibitor LY294002 (LY) can impact the interaction between GABARAPL1 promoter and BRCA1, FoxO1, RXRA, CREB1 and c-MYC (Fig. 3E–I) and block the decrease in GABARAPL1 after OGD/R treatment (Fig. 3J), which indicates that BRCA1, FoxO1, RXRA, CREB1 and c-MYC act as downstream nodes for PI3K-Akt activation-mediated GABARAPL1 downregulation during cerebral I/R.

Fig. 3 — PI3K-Akt pathway activation participates in GABARAPL1 downregulation during MCAO/R and OGD/R. (A) Reverse ChiP-MS analysis of TF binding to the GABARAPL1 promoter in cultured astrocytes at 12 h post-OGD/R (n = 4). Seven reverse ChiP probes were designed based on the GABARAPL1 promoter nucleotide sequence. The probe length is approximately 40 bps. The eluted TF proteins were analyzed using mass spectrometry. (B) RNA sequencing analysis of the PI3K-Akt pathway in cultured astrocytes at 12 h post-OGD/R (n = 4). (C) Immunoblotting analysis of the ratio of phosphorylated PI3K (p-PI3K) normalized to PI3K in cultured astrocytes at 12 h post-OGD/R (n = 6). LY294002 (LY) was added to the culture medium at a final concentration of 40 μM immediately after OGD/R. (D) Immunoblotting analysis of the ratio of phosphorylated Akt (p-Akt) normalized to Akt in cultured astrocytes at 12 h post-OGD/R (n = 6). (E–I) Reverse ChiP-immunoblotting analysis of BRCA1 (E), FoxO1 (F), RXRA (G), CREB1 (H) and c-MYC (I) binding to the GABARAPL1 promoter in cultured astrocytes at 12 h post-OGD/R (n = 6). (J) Immunoblotting analysis of GABARAPL1 expression in cultured astrocytes at 12 h post-OGD/R (n = 6). (K and L) Immunofluorescence analysis of p-PI3K (K) and PI3K (L) expression in astrocytes in the penumbra regions of mice at 12 h following cerebral I/R (n = 8). Astrocytes are marked using S100β. Scale bars are 50 μm. (M and N) Immunofluorescence analysis of p-Akt (M) and Akt (N) expression in astrocytes of the penumbra regions in mice at 12 h following cerebral I/R (n = 8). Scale bars are 50 μm. The data are presented as the means ± SDs. Statistical analyses were performed using one-way ANOVA followed by an LSD post hoc test (F and H–J) and the Donnett T3 post hoc test (C-E and G), and independent t-tests (A and K–N). ns represents no significance. *P < 0.05, **P < 0.01, ***P < 0.001.

Immunofluorescence was used to explore whether the PI3K-Akt pathway was also activated in the mouse MCAO/R model. We observed that the level of phosphorylated PI3K was increased and that the level of PI3K was not changed in astrocytes in the penumbra region after MCAO/R (Fig. 3K and L). The level of phosphorylated Akt was increased, but the level of Akt was unchanged in astrocytes after MCAO/R (Fig. 3M and N).

3.3. Decreases in glucosamine levels mediated by glycophagic dysfunction aggravate GABARAPL1 downregulation in astrocytes during cerebral reperfusion

Current knowledge suggests that astrocytic glycophagy decomposes glycogen into glucose [11,26]. We speculated that astrocytic glycophagy might degrade brain glycogen into glucose and glucosamine. To confirm this hypothesis, we replaced the lysosome-guided signal peptide of the GAA enzyme with an extracellular-guided signal peptide to separate the resynthesized GAA (rGAA) from the intracellular GAA using an eukaryotic expression system (Fig. 4A and Figs. S5A–C). We then digested lysosomal glycogen in astrocytes with rGAA. Glucose and glucosamine levels were both upregulated in rGAA-treated astrocytes than in boiled rGAA-treated astrocytes (Fig. 4A). The ratio of glycophagy-derived glucose to glucosamine was approximately 3:1 (Fig. 4A). Accordingly, brain glycophagy can degrade glycogen into glucose and glucosamine simultaneously. We then detected the astrocytic protein O-GlcNAcylation level post-OGD/R, and the O-GlcNAcylation levels of GABARAPL1-related TFs were selected for further analysis (Fig. 4B). Among them, 10 TFs contained O-GlcNAcylated sites in their amino acid sequences, and the O-GlcNAcylation levels of six TFs decreased significantly post-OGD/R treatment (Fig. 4B). Combined with the results from reverse ChiP-MS (Fig. 3A), only SP1 and TBP exhibited changes in both O-GlcNAcylation and GABARAPL1-related binding conditions after OGD/R treatment.

Next, we wanted to clarify whether there is an unsuspected feedback loop in which glycophagy dysfunction-mediated glucosamine decline can inhibit SP1 and TBP nuclear translocation by suppressing their O-GlcNAcylation levels, and this feedback exacerbates GABARAPL1 downregulation. To answer this question, we first constructed cultured GABARAPL1-overexpressing astrocytes (Ve-GA, Fig. S5D). Previous studies suggested that GABARAPL1 might not be specific for glycophagy and might also participate in other types of autophagy, such as mitophagy [27,28]. Accordingly, we further downregulated GAA expression in cultured Ve-GA astrocytes (Ve-GA+sh-GAA, Fig. S5E) since GAA is reported to be specific for glycophagy [9,29]. We performed electron microscopy and observed that the number of glycophagy-related vacuoles increased with GABARAPL1 overexpression (Fig. 4C). This effect was blocked when GAA was knocked down post-OGD/R treatment (Fig. 4C). In addition, the glycogen level in astrocytic lysosomes was also increased in cultured Ve-GA astrocytes and decreased in cultured Ve-GA+sh-GAA astrocytes post-OGD/R treatment (Fig. 4D). As a result, the accumulation of glycogen in astrocytes decreased with GABARAPL1 overexpression and was further enhanced when GAA was inhibited post-OGD/R treatment (Fig. 4E). The decrease in glucosamine levels was enhanced with GABARAPL1 overexpression (Fig. 4F). This effect was blocked when GAA was further downregulated post-ODG/R treatment (Fig. 4F). The level of UDP-GlcNAc, the substrate for O-GlcNAcylation, was also increased in cultured Ve-GA astrocytes but decreased in cultured Ve-GA+sh-GAA astrocytes post-OGD/R treatment (Fig. 4G). Notably, the expression of O-GlcNAcase and O-GlcNAc transferase remained unchanged with GABARAPL1 overexpression or GAA knockdown in cultured astrocytes post-OGD/R treatment (Figs. S5F and G). Coimmunoprecipitation was performed to pull down astrocytic SP1 and TBP, and we observed that the O-GlcNAcylation levels of SP1 and TBP increased with GABARAPL1 overexpression after OGD/R, which further decreased when GAA was knocked down (Fig. 4H and I). Triple-label immunocytochemistry was used to explore the expression of O-GlcNAcylated SP1 and O-GlcNAcylated TBP in the nuclei of cultured astrocytes, and both were enhanced in cultured Ve-GA astrocytes and decreased in cultured Ve-GA+sh-GAA astrocytes after OGD/R treatment (Fig. 4J and K). Along with the alterations in O-GlcNAcylated SP1 and TBP, GABARAPL1 overexpression increased SP1 and TBP nuclear translocation, and this effect was blocked by GAA knockdown after OGD/R treatment (Fig. 4L and M). Accordingly, the interaction between the GABARAPL1 promoter and SP1 and TBP, analyzed by reverse ChiP-immunoblotting, increased with GABARAPL1 overexpression and further decreased when GAA was knocked down (Fig. 4N and O).

There are 6 O-GlcNAcylated amino acid sites for SP1 and 2 O-GlcNAcylated amino acid sites for TBP in mouse astrocytes (Fig. 4B). We then investigated which O-GlcNAcylation site determines the nuclear translocation of SP1 and TBP. We used NLS Mapper to predict nuclear localization signals (NLSs) in the amino acid sequences of SP1 and TBP and found that S613, S699, and S703 are localized in the NLS of SP1, while S238 is localized in the NLS of TBP (Fig. 4P and Q). In addition, we constructed SP1 amino acid mutant cultured astrocytes (S493^A, S613^A, T641^A, S642^A, S699^A, and S703^A) and TBP amino acid mutant cultured astrocytes (T113^A and S238^A) to change threonine or serine to alanine, respectively [30]. We found that the SP1 S613^A mutant, but not other mutants, significantly impacted the nuclear translocation of SP1 (Fig. 4P). The S238^A mutation, but not the T113^A mutation, determines the nuclear translocation of TBP (Fig. 4Q). To further determine the accurate binding domain between the GABARAPL1 promoter and the SP1 and TBP, we used the JASPAR database to predict the binding domain. SP1 binding domain may be located in two regions (−2000 to −448 and −447 to +40) and the TBP binding domain may be located in four regions (−2000 to −1504, −1503 to −1179, −1178 to −462, and −461 to +40). Accordingly, we truncated the GABARAPL1 promoter into two fragments for SP1 and four fragments for TBP. We observed that SP1 was located at the −2000 to −448 region (Fig. 4R), as determined by a dual-luciferase reporter system, and TBP was located at the −461 to +40 region (Fig. 4S). Based on these findings, we propose that astrocyte exists a positive feedback loop to aggravate glycophagic dysfunction during cerebral I/R. Specifically, astrocytic glycophagy dysfunction suppresses glycogen-derived glucosamine levels, which inhibits SP1 and TBP nuclear translocation via O-GlcNAcylation, and feedback exacerbates GABARAPL1 downregulation.

3.4. Restoration of astrocytic glycophagy ameliorates reperfusion injury after ischemic stroke

We aimed to determine whether the restoration of astrocytic glycophagy could improve the in vivo outcomes of cerebral reperfusion injury by constructing a tamoxifen-induced astrocyte-specific GABARAPL1 overexpression transgenic mouse model (KI-GA, Figs. S6A–C). Seven-week-old offspring of GABARAPL1-loxP mice and ALDH1L1-CreERT2 mice received intraperitoneal injections of tamoxifen (30 mg/kg) daily for one week to conditionally overexpress astrocytic GABARAPL1 post-MCAO/R (Fig. 5A). To confirm the specificity of glycophagy restoration towards brain I/R, we conditionally knocked down astrocytic GAA expression in the KI-GA transgenic model by injection of AAV with shRNA targeting the GAA gene into lateral ventricular and right striatum (KI-GA+KD-GAA, Fig. S6D). We first conducted electron microscopy and observed that the glycophagic flux in KI-GA transgenic mice was enhanced and further inhibited by GAA knockdown post-MCAO/R (Fig. 5B). The glycogen level in the lysosome was also upregulated in KI-GA transgenic mice and was further suppressed by GAA knockdown post-MCAO/R (Fig. 5C). In addition, we analyzed the brain glycogen level and found that it decreased in KI-GA transgenic mice but increased when astrocytic GAA was further blocked post-MCAO/R insult (Fig. 5D).

Along with the increase in glycophagic flux in astrocytes, the infarct volume in KI-GA transgenic mice decreased (Fig. 5E). This effect was blocked when astrocytic GAA was inhibited at three days post-MCAO/R (Fig. 5E). Neurobehavioral tests, including the corner test, grid-walking test, and adhesive removal test, were performed to assess sensorimotor function in KI-GA and KI-GA+KD-GAA transgenic mice. The total number of steps increased, and the foot fault ratio decreased with astrocytic GABARAPL1 overexpression, which was blocked when astrocytic GAA was further inhibited (Fig. 5F). The time to contact and time to remove in the adhesive removal test decreased in KI-GA transgenic mice and increased in KI-GA+KD-GAA transgenic mice post-MCAO/R (Fig. 5G). The number of right turns in the corner test decreased in KI-GA transgenic mice and increased with GAA suppression (Fig. 5H). In addition, we speculated that astrocytic GABARAPL1 overexpression-induced neurobehavioral recovery may be associated with neuronal synaptic regeneration. Thus, we used Golgi staining to analyze the neuronal synaptic structure in the penumbra region and discovered that the number of dendritic spines and the number of synaptic branches were increased in KI-GA transgenic mice but decreased with GAA knockdown at 14 days post-MCAO/R (Fig. 5I).

3.5. Glycophagy restoration reduces oxidative injury and DNA damage in astrocytes during cerebral reperfusion

What is the mechanism underlying astrocytic glycophagy restoration-mediated neuroprotection against brain I/R insult? We observed that the level of glucose in astrocytes increased with GABARAPL1 overexpression and further decreased when GAA was downregulated after OGD/R treatment (Fig. 6A). We then performed metabolomic analysis of central carbon metabolism, including glycolysis, the PPP, and the TCA cycle, in cultured astrocytes. We observed that the levels of metabolites in the PPP but not in glycolysis or the TCA cycle were increased in cultured Ve-GA astrocytes and further decreased in cultured Ve-GA+sh-GAA astrocytes after OGD/R treatment (Fig. 6B). In addition, we measured ¹³C-labeled metabolic flux and discovered that ¹³C-labeled glucose enters the PPP more than into glycolysis or the TCA cycle (Fig. 6C). The levels of ¹³C-labeled metabolites in the PPP, including ribulose-5-phosphate, 6-phosphogluconolactone, ribose-5-phosphate, 6-phosphogluconate and xylulose-5-phosphate, increased with GABARAPL1 overexpression and further decreased when GAA was further knocked down post-OGD/R treatment (Fig. 6C).

Fig. 6 — GABARAPL1 overexpression decreases oxidative stress and repairs DNA damage in astrocytes during OGD/R. (A) Glucose levels in cultured astrocytes at 12 h post-OGD/R (n = 8). (B) Metabolomic analysis of the levels of metabolites involved in glycolysis, the TCA cycle, and the PPP in cultured astrocytes at 12 h post-OGD/R (n = 4). (C) Metabolic flux analysis using ¹³C-labeled glucose to detect the metabolic flux of glycolysis, TCA cycle and the PPP in cultured astrocytes at 12 h post-OGD/R (n = 4). The left panel is a schematic diagram presenting ¹³C-labeled metabolites in glycolysis, TCA cycle, and the PPP from ¹³C-labeled glucose. The black and red circles represent unlabeled and ¹³C-labeled carbon atoms, respectively. (D) G6P levels in cultured astrocytes at 12 h post-OGD/R (n = 8). (E) NADPH levels in cultured astrocytes at 12 h post-OGD/R (n = 8). (F) GSH levels in cultured astrocytes at 12 h post-OGD/R (n = 8). (G) ROS levels in cultured astrocytes at 12 h post-OGD/R (n = 8). (H) PRPP levels in cultured astrocytes at 12 h post-OGD/R (n = 8). (I) Comet assay to detect DNA damage in cultured astrocytes at 12 h post-OGD/R (n = 8). DNA damage was quantified as the percentage of tail DNA intensity divided by the whole-cell DNA intensity. Scale bars are 100 μm. (J) LDH release in cultured astrocytes at 24 h post-OGD/R (n = 8). The data are presented as the means ± SDs. Statistical analyses were performed using one-way ANOVA followed by the LSD post hoc test (A and C-J). *P < 0.05, **P < 0.01, ***P < 0.001. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Next, we aimed to clarify the influence of GABARAPL1 overexpression-induced PPP enhancement on astrocyte after OGD/R treatment. We discovered that G6P was increased in cultured Ve-GA astrocytes and decreased in cultured Ve-GA+sh-GAA astrocytes post-OGD/R treatment (Fig. 6D). The levels of NADPH and GSH were increased with GABARAPL1 overexpression and helped combat oxidative stress damage in cultured astrocytes, which was blocked when GAA was further blocked (Fig. 6E–G). In addition, the PPP acts as a crucial pathway to produce molecules such as ribose-5-phosphate to repair DNA injury [31]. We observed that the level of PRPP, which is derived from ribose-5-phosphate, was increased in cultured Ve-GA astrocytes and decreased in cultured Ve-GA+sh-GAA astrocytes after OGD/R treatment (Fig. 6H). We used a comet assay to detect DNA damage and observed that DNA damage was alleviated in cultured Ve-GA astrocytes and further aggravated in cultured Ve-GA+sh-GAA astrocytes after OGD/R treatment (Fig. 6I). Accordingly, the damage to cultured Ve-GA astrocytes, reflected by LDH release, decreased with GABARAPL1 overexpression and further increased with GAA knockdown post-OGD/R treatment (Fig. 6J).

3.6. Restoration of astrocytic glycophagy accelerates neighboring neuron recovery via astrocyte-neuron interactions between GSH and glucosamine during cerebral reperfusion

We used an astrocyte-neuron coculture system (Fig. 7A) and discovered that neuron survival increased when cocultured with Ve-GA astrocytes and decreased when cocultured with Ve-GA+sh-GAA astrocytes after OGD/R treatment (Fig. 7B). We detected three metabolic end products including GSH, PRPP, and glucosamine, whose levels increased with the restoration of glycophagy in astrocytes during I/R. As reported in previous studies, several metabolic interactions occur between astrocytes and neurons [17], but which interaction exerts the predominant impact on astrocytic glycophagy restoration-mediated protection of surrounding neurons remains uncertain. We first focused on the astrocyte-neuron GSH interaction, in which astrocytes can provide GSH to surrounding neurons via the astrocytic membrane γ-GT [32]. We then detected the GSH levels in the coculture medium and cocultured neurons and observed that the GSH levels increased with astrocytic GABARAPL1 overexpression and further decreased with astrocytic GAA knockdown after OGD/R treatment (Fig. 7C and D). The level of neuronal NADPH increased and the level of ROS decreased when cocultured with Ve-GA astrocytes post-OGD/R, and these effects were further blocked when cocultured with Ve-GA+sh-GAA astrocytes (Fig. 7E and F). Next, we detected the neuronal PRPP level and discovered that it did not change with astrocytic GABARAPL1 overexpression or GAA knockdown post-OGD/R treatment (Fig. 7G), suggesting that astrocyte-derived PRPP cannot be transported into surrounding neurons. In addition to those of GSH and PRPP, glucosamine levels in the coculture medium increased with astrocytic GABARAPL1 overexpression but further decreased with astrocytic GAA knockdown post-OGD/R treatment (Fig. 7H). Neuronal glucosamine levels were also increased when cocultured with Ve-GA astrocytes and decreased when cocultured with Ve-GA+sh-GAA astrocytes post-OGD/R treatment (Fig. 7I). We then adopted the immunocytochemistry to analyze the neuronal synaptic structure. The number of neuronal synaptic branches increased after coculture with Ve-GA astrocytes and decreased after coculture with Ve-GA+sh-GAA astrocytes post-OGD/R treatment (Fig. 7J).

To further determine the different effects of astrocyte-derived GSH and glucosamine on neurons, we knocked down γ-GT in Ve-GA astrocytes to conditionally suppress the astrocyte-neuron GSH interaction (Fig. S7). We observed that neuronal viability was partially decreased with γ-GT knockdown in Ve-GA astrocytes but was still increased compared with that in OGD-treated astrocytes (Fig. 7K). In addition, the antioxidant effects induced by astrocytic GABARAPL1 overexpression in surrounding neurons were blocked when astrocytic γ-GT was further knocked down (Fig. 7L-N). However, astrocytic GABARAPL1 overexpression-induced neuronal synaptic regeneration was not affected when astrocytic γ-GT was further knocked down (Fig. 7O). Based on the above evidences, we considered that astrocyte-derived GSH predominantly strengthened neuronal antioxidant abilities and that astrocyte-derived glucosamine mainly promoted neuronal synaptic regeneration during astrocytic glycophagy restoration after cerebral reperfusion.

3.7. A hypocaloric diet enhances astrocytic glycophagy and is neuroprotective against reperfusion injury after ischemic stroke

We then speculated that energy resupply during the acute phase of brain reperfusion should begin at a low level and gradually change to a normal state, which might improve the prognosis for patients with ischemic stroke. To verify this hypothesis, we changed the reperfusion medium to low glucose (2.5 mM) for cultured astrocytes post-OGD/R and observed that low glucose treatment increased the number of glycophagy-related vacuoles in cultured astrocytes, and this effect was blocked by GAA knockdown post-OGD/R treatment (Fig. 8A). In addition, the glycogen level in lysosomes was increased in low glucose-treated cultured astrocytes and further decreased with GAA knockdown post-OGD/R treatment (Fig. 8B). The damage to cultured astrocytes was alleviated by low glucose treatment and further exacerbated by GAA knockdown post-OGD/R treatment (Fig. 8C). In addition, the survival of cocultured neurons was enhanced when the reperfusion medium was changed to low glucose, and this effect was blocked by GAA knockdown post-OGD/R treatment (Fig. 8D).

Fig. 8 — A hypocaloric diet alleviates brain damage during MCAO/R and OGD/R. (A) Electron microscopic analysis of the average number of glycophagy-related vacuoles in cultured astrocytes at 12 h post-OGD/R (n = 8). The arrows represent glycophagy-related vacuoles. The cultured astrocytes were treated with a low-glucose medium immediately after OGD/R. Scale bars are 2 μm. (B) Glycogen levels in the lysosomes of cultured astrocytes at 12 h post-OGD/R (n = 8). (C) LDH release in cultured astrocytes at 24 h post-OGD/R (n = 8). (D) CCK-8 analysis to detect cocultured neuronal viability at 24 h post-OGD/R (n = 8). The cocultured astrocytes and cocultured neurons were treated with a low-glucose medium immediately after OGD/R. (E) Timeline depicting the diet strategy and morphological and neurobehavioral analysis of mice following cerebral I/R. (F) Electron microscopic analysis of the average number of glycophagy-related vacuoles in the astrocytes of the penumbra region in mice at 12 h following cerebral I/R (n = 8). The arrows represent glycophagy-related vacuoles. Scale bars are 2 μm. (G) Glycogen levels in the lysosomes of the penumbra region in mice at 12 h following I/R (n = 8). (H) TTC staining showing infarct volumes in mice at 72 h following cerebral I/R (n = 8). Scale bars are 1 mm. (I) The total steps (left) and foot fault ratios (right) before and after reperfusion in mice, as detected by the grid-walking test (n = 8). (J) The time to contact (left) and time to remove (right) before and after reperfusion in mice, as detected by the adhesive removal test (n = 8). (K) The number of right turns in 10 trials before and after reperfusion in mice, as detected by the corner test (n = 8). (L) Golgi staining to detect the neuronal dendritic spine numbers and branches in the penumbra regions of mice at 14 days following cerebral I/R (60 astrocytes from six animals/group). The data are presented as the means ± SDs. Statistical analyses were performed using one-way ANOVA followed by the LSD post hoc test (A-D, F-H and L (left panel)) and the Donnett T3 post hoc test (L (right panel)) and repeated-measures analysis (I–K). *P < 0.05, **P < 0.01, ***P < 0.001.

Next, we considered that the energy supply in the diet should also begin with hypocaloric conditions and gradually change to normocaloric conditions during the acute phase of cerebral reperfusion. Accordingly, we fed the mice a hypocaloric diet for the first week and changed to a normal diet in the second-week post-MCAO/R (Fig. 8E). We observed that a hypocaloric diet increased the number of astrocytic glycophagy-related vacuoles post-MCAO/R, which further decreased with astrocytic GAA inhibition (Fig. 8F). In addition, the glycogen level in lysosomes increased with a hypocaloric diet and decreased with astrocytic GAA suppression post-MCAO/R (Fig. 8G). The infarct volumes decreased in the hypocaloric diet-treated mice and were further augmented with astrocytic GAA knockdown at three days post-MCAO/R (Fig. 8H). Neurobehavioral changes, as measured by the grid-walking test, adhesive removal test, and corner test, were alleviated in hypocaloric diet-treated mice, and these effects were further blocked by astrocytic GAA knockdown post-MCAO/R (Fig. 8I–K). In addition, we discovered that dendritic spine numbers and synaptic branches were upregulated in hypocaloric diet-treated mice than in normocaloric mice and further decreased with astrocytic GAA inhibition at 14 days post-MCAO/R (Fig. 8L).

4. Discussion

In this study, we found that astrocytic glycophagy is dysfunctional due to GABARAPL1 downregulation in brain, which is partially caused by PI3K-Akt pathway activation. Besides, astrocytic glycophagy-derived glucosamine deficiency suppresses GABARAPL1 expression by regulating its TF nuclear translocation through O-GlcNAcylation during cerebral I/R. Restoring glycophagy decreases oxidative injury and DNA damage in astrocytes and improves the survival of neighboring neurons by enhancing antioxidative abilities and synaptic regeneration during cerebral I/R. A hypocaloric diet in the acute phase works as a neuroprotective strategy against reperfusion injury after ischemic stroke.

Glycophagy, a type of glycogen-specific autophagy, is traditionally considered an alternative glycogen degradation pathway [33]. Recent studies have shown that glycophagy has evolved into a multifaceted pivot to regulate cellular metabolic hemostasis [7]. For instance, in skeletal muscle, a short-term decrease in glycophagy enhances fatty acid oxidation, increases glutamine utilization and reduces the NAD content [9]. Lipid droplets depend on glycophagy to provide essential metabolites to adipocytes [10]. Glycophagic dysfunction is involved in many diseases, such as Lafora disease and Pompe disease [11,12], and aggravates diabetic cardiomyopathy damage [13,14]. However, research on brain glycophagy is at an early stage, and the relationship between brain glycophagy and neurological diseases is unknown. Here, we discovered that brain glycophagy can determine the outcomes of reperfusion injury after ischemic stroke. The understanding of brain glycophagy deserves further attention to broaden the field of glycogen metabolism and deepen our understanding of energy metabolism.

In this study, we observed an unexpected glycophagy-oriented feedback loop in the brain in which astrocytic glycophagy can degrade glycogen into glucosamine and feedback control GABARAPL1 expression via O-GlcNAcylation. There might be other glycophagy-oriented control systems in the brain. For instance, glycophagy-derived glucosamine can regulate glycogenesis and glycogenolysis flux in glycogen metabolism via glycosylation. This regulation can be achieved in two ways. First, rate-limiting enzyme expression may be altered in glycogenesis or glycogenolysis by regulating its TF function, which was proven feasible in this study because many TFs can be modified by O-GlcNAcylation. Second, the activities of essential enzymes in glycogenesis or glycogenolysis could be altered via glycosylation. We investigated the O-GlcNAcylation sites of the primary enzymes involved in glycogenesis and glycogenolysis [34,35] and observed that none of them were modified by O-GlcNAcylation (Fig. S8A). Glycosylation can occur in various forms, including O-, N-, P-, C-, and S-glycosylation [36,37]. Whether the key enzymes involved in glycogen metabolism can be modified by other forms of glycosylation, except for O-GlcNAcylation, and the impact of glycosylation on the activity of these enzymes still needs to be investigated. In addition, glucosamine is an abundant constituent of brain glycogen, accounting for approximately 25 % of its composition, whereas it is not found in muscle, heart, or liver glycogen [15]. In the muscle, heart, and liver, glycogen is reportedly degraded into G6P, but not glucosamine-6-phosphate, via glycogenolysis [15]. Consequently, the glycophagy-mediated control system is tissue-specific and mainly occurs in the brain.

Currently, the impact of glycosylation on TF function has not been fully clarified. Some studies indicate that glycosylation can regulate TF nuclear translocation, binding with promoters, stability, and interactions with other cofactors [38]. In this study, we identified 10 GABARAPL1-related TFs with O-GlcNAcylation sites, including NCOR1, FoxO1, FOS, SAP130, EP400, EGR1, SP1, TBP, SREBF2, and FoxA1. At present, some studies suggest that S493 O-GlcNAcylation determines the stability of SP1, and T641, S642, S699, and S703 O-GlcNAcylation determines its transcriptional activity [39,40]. T113 O-GlcNAcylation is crucial for the interaction of TBP with other cofactors [41]. However, the key O-GlcNAcylation sites involved in the nuclear translocation of SP1 and TBP are unknown; the sites identified in this study were S613 for SP1 and S238 for TBP. In addition to SP1 and TBP, O-GlcNAcylation reportedly enhances FoxO1 transcriptional activity in liver cells [42]. O-GlcNAcylation upregulates FOS transcriptional activity and increases its stability in the brain [43]. The stability of FoxA1 is also promoted by O-GlcNAcylation in breast cancer cells [44]. The relationship between O-GlcNAcylation and the functions of NCOR1, SAP130, EP400, EGR1, and SREBF2 still needs to be explored.

The metabolic fate of glucose released from astrocytic glycophagy following GABARAPL1 overexpression during reperfusion remains unclear. We observed that glycophagy-derived glucose enters the PPP more significantly than it enters glycolysis and the TCA cycle, possibly because the key enzyme in glycolysis is suppressed in astrocytes during recanalization following ischemic stroke. One study indicated that many glycolytic enzymes are repressed in astrocytes treated with I/R [45]. Novel drugs aimed at enhancing astrocytic glycolysis appear to be neuroprotective against reperfusion injury after ischemic stroke [46,47]. The development of new approaches to restore glycolytic flux may amplify the neuroprotective effects of glycophagy-liberated glucose on astrocytes by transforming astrocytic glucose metabolism to a more balanced state.

Many metabolic interactions exist between astrocytes and neurons in the brain [17,48]. Regarding the GSH shuttle, astrocytes can transfer GSH into the extracellular space and convert it into reduced substances with the assistance of astrocytic γ-GT and the neuronal ectopeptidase aminopeptidase N [32,49]. The surrounding neurons can utilize these reduced substances to synthesize reduced GSH and fight against oxidative stress [32] because neurons depend on aerobic oxidation to satisfy their high energy demands [50,51]. Restoring astrocytic glycophagy alleviates oxidative stress in surrounding neurons via the GSH shuttle. Compared with the literature on classical GSH shuttling, no studies have investigated the astrocyte-neuron glucosamine interaction, but researchers have provided clues that brain glucose transporters can transport glucosamine into the extracellular matrix [52] and that increasing neuronal O-GlcNAcylation protects neurobehavioral outcomes in young mice with ischemic stroke [53]. Additionally, glycosylation is crucial for synaptogenesis in neuronal synapses [54]. Increasing neuronal glycosylation levels by overexpressing O-GlcNAc transferase can accelerate associative fear memory formation in young adult mice [55]. One study detected the O-GlcNAcylation sites of mouse synapses and found >1750 O-GlcNAcylation sites in proteins in synaptosomes [56]. Another study suggested that elevated O-GlcNAcylation accelerates dopamine neuron synaptic transmission in Parkinson's disease patients [57]. In the present study, we observed that glucosamine-induced restoration of astrocytic glycophagy can promote neuronal synaptic regeneration during I/R. Thus, restoring glycophagy accelerates the resistance of astrocytes and neighboring neurons rather than solely enhancing astrocytic viability to fight reperfusion injury.

Notably, diet therapy after endovascular thrombolysis or thrombectomy is not widely used in clinical practice for stroke patients, especially when stroke patients do not suffer from primary diseases such as diabetes, hyperlipoidaemia or hypertension. In a mouse study, a hypocaloric diet was reported to be neuroprotective against ischemic stroke and beneficial for peri-infarct brain remodeling [58]. Another study revealed that intermittent fasting for four months can attenuate brain damage in I/R by decreasing inflammation in mice [59]. In this study, we aimed to identify a simple and effective therapeutic strategy to restore glycolytic flux. We speculated that a hypocaloric diet is more suitable for stroke patients in the acute phase after reperfusion since stroke patients reportedly have lower caloric requirements during the first five days [60]. Compared with modified full enteral nutrition, diet therapy is changed to a normocaloric diet in the second week after reperfusion because one clinical study reported that long-term hypocaloric enteral nutrition increases mortality in severe stroke patients [61]. Accordingly, we believe that a hypocaloric diet during the first week of reperfusion can promote glycophagic flux and alleviate brain damage after ischemic stroke, suggesting that a hypocaloric diet is not only beneficial for stroke patients with metabolic diseases but also suitable for all stroke patients.

This study had several limitations. Firstly, I/R-induced GABARAPL1 downregulation is caused by multiple TF dysfunction. One TF can be regulated by many signaling pathways, and one signaling pathway can modulate multiple TFs [62]. Here, we focused only on the PI3K-Akt signaling pathway and its downstream TFs. However, other signaling pathways might also participate in modulating glycophagic dysfunction during I/R stress. Additionally, although we did not find SP1 or TBP to be the direct downstream nodes for the PI3K-Akt signaling pathway based on the GenomeNet KEGG website, a previous study suggested that SP1 functions as a downstream effector of the PI3K-Akt signaling pathway in the mammary epithelial cells of cows [63]. The relationship between PI3K-Akt signaling pathway activation during I/R and SP1 nuclear translocation should be confirmed in future research. Finally, one study suggested that GABARAPL1 knockdown inhibits mitophagy in breast cancer cells [27]. Another study revealed that viral interferon regulatory factor 1 of human herpesvirus 8 can interact with GABARAPL1 to activate mitophagy in infected cells [28]. Here, we observed that GABARAPL1 overexpression had little impact on mitophagy flux in astrocytes (Fig. S8B). The role of GABARAPL1 in mitophagy is mainly based on evidences from peripheral tissues. Whether GABARAPL1 can act as the key node to alter mitophagy in the brain, especially in astrocytes, needs to be confirmed in the future.

In conclusion, our findings revealed that astrocytic glycophagy is dysfunctional during I/R and that restoring glycophagy is a novel therapeutic target for reperfusion damage after ischemic stroke. In addition, we proposed the existence of a brain glycophagy-oriented feedback loop, which deepens our understanding of brain glycophagy and establishes a direct connection linking autophagy, metabolism, and epigenetics together.

CRediT authorship contribution statement

Haiyun Guo: Methodology, Investigation. Yumeng Li: Methodology, Investigation. Shiquan Wang: Methodology, Investigation. Yongheng Yang: Investigation. Tiantian Xu: Investigation. Jianshuai Zhao: Investigation. Jin Wang: Investigation. Wenqiang Zuo: Investigation. Pengju Wang: Investigation. Guangchao Zhao: Investigation. Huaning Wang: Investigation. Wugang Hou: Writing – review & editing, Conceptualization. Hailong Dong: Writing – review & editing, Conceptualization. Yanhui Cai: Writing – review & editing, Writing – original draft, Conceptualization.

Declaration of competing interest

The authors declare that they have no competing interests.

Acknowledgments

This study was financially supported by the National Natural Science Foundation of China (Nos. 82271508, 82271316, 82030038, 82221001 and 82230037), the Shaanxi Provincial Natural Science Foundation for Distinguished Young Scholars (No. 2021JC-33) and the Assistance Program of Xijing Hospital (No. XJZT24CY25).

Footnotes

^{Appendix A}

Supplementary data to this article can be found online at https://doi.org/10.1016/j.redox.2024.103234.

Contributor Information

Wugang Hou, Email: gangwuhou@163.com.

Hailong Dong, Email: hldong6@hotmail.com.

Yanhui Cai, Email: MD_CAI@163.com.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Multimedia component 1

mmc1.docx^{(3.4MB, docx)}

Data availability

Data will be made available on request.

References

1.Global, regional, and national burden of stroke and its risk factors, 1990-2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet Neurol. 2021;20(10):795–820. doi: 10.1016/s1474-4422(21)00252-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Mendelson S.J., Prabhakaran S. Diagnosis and management of transient ischemic attack and acute ischemic stroke: a review. JAMA. 2021;325(11):1088–1098. doi: 10.1001/jama.2020.26867. [DOI] [PubMed] [Google Scholar]
3.Bar B., Biller J. Select hyperacute complications of ischemic stroke: cerebral edema, hemorrhagic transformation, and orolingual angioedema secondary to intravenous Alteplase. Expert Rev. Neurother. 2018;18(10):749–759. doi: 10.1080/14737175.2018.1521723. [DOI] [PubMed] [Google Scholar]
4.Pundik S., Xu K., Sundararajan S. Reperfusion brain injury: focus on cellular bioenergetics. Neurology. 2012;79(13 Suppl 1):S44–S51. doi: 10.1212/WNL.0b013e3182695a14. [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Liu M., et al. TIGAR alleviates oxidative stress in brain with extended ischemia via a pentose phosphate pathway-independent manner. Redox Biol. 2022;53 doi: 10.1016/j.redox.2022.102323. [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Koutsifeli P., et al. Glycogen-autophagy: molecular machinery and cellular mechanisms of glycophagy. J. Biol. Chem. 2022;298(7) doi: 10.1016/j.jbc.2022.102093. [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Delbridge L.M., et al. Myocardial autophagic energy stress responses--macroautophagy, mitophagy, and glycophagy. Am. J. Physiol. Heart Circ. Physiol. 2015;308(10):H1194–H1204. doi: 10.1152/ajpheart.00002.2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Chandramouli C., et al. Myocardial glycogen dynamics: new perspectives on disease mechanisms. Clin. Exp. Pharmacol. Physiol. 2015;42(4):415–425. doi: 10.1111/1440-1681.12370. [DOI] [PubMed] [Google Scholar]
9.Heden T.D., et al. Regulation and role of glycophagy in skeletal muscle energy metabolism. Autophagy. 2022;18(5):1078–1089. doi: 10.1080/15548627.2021.1969633. [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Mayeuf-Louchart A., et al. Glycogen dynamics drives lipid droplet biogenesis during Brown adipocyte differentiation. Cell Rep. 2019;29(6):1410–1418.e6. doi: 10.1016/j.celrep.2019.09.073. [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Brewer M.K., Gentry M.S. Brain glycogen structure and its associated proteins: past, present and future. Adv Neurobiol. 2019;23:17–81. doi: 10.1007/978-3-030-27480-1_2. [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Palhegyi A.M., et al. Biomedical implications of autophagy in macromolecule storage disorders. Front. Cell Dev. Biol. 2019;7:179. doi: 10.3389/fcell.2019.00179. [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Reichelt M.E., et al. Myocardial glycophagy - a specific glycogen handling response to metabolic stress is accentuated in the female heart. J. Mol. Cell. Cardiol. 2013;65:67–75. doi: 10.1016/j.yjmcc.2013.09.014. [DOI] [PubMed] [Google Scholar]
14.Mellor K.M., et al. Cardiomyocyte glycophagy is regulated by insulin and exposure to high extracellular glucose. Am. J. Physiol. Heart Circ. Physiol. 2014;306(8):H1240–H1245. doi: 10.1152/ajpheart.00059.2014. [DOI] [PubMed] [Google Scholar]
15.Sun R.C., et al. Brain glycogen serves as a critical glucosamine cache required for protein glycosylation. Cell Metabol. 2021;33(7):1404–1417.e9. doi: 10.1016/j.cmet.2021.05.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Oe Y., Akther S., Hirase H. Regional distribution of glycogen in the mouse brain visualized by immunohistochemistry. Adv Neurobiol. 2019;23:147–168. doi: 10.1007/978-3-030-27480-1_5. [DOI] [PubMed] [Google Scholar]
17.Bélanger M., Allaman I., Magistretti P.J. Brain energy metabolism: focus on astrocyte-neuron metabolic cooperation. Cell Metabol. 2011;14(6):724–738. doi: 10.1016/j.cmet.2011.08.016. [DOI] [PubMed] [Google Scholar]
18.Cai Y., et al. Glycogenolysis is crucial for astrocytic glycogen accumulation and brain damage after reperfusion in ischemic stroke. iScience. 2020;23(5) doi: 10.1016/j.isci.2020.101136. [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Guo H., et al. Astrocytic glycogen mobilization participates in salvianolic acid B-mediated neuroprotection against reperfusion injury after ischemic stroke. Exp. Neurol. 2022;349 doi: 10.1016/j.expneurol.2021.113966. [DOI] [PubMed] [Google Scholar]
20.Kondomerkos D.J., et al. Glycogen autophagy in the liver and heart of newborn rats. The effects of glucagon, adrenalin or rapamycin. Histol. Histopathol. 2005;20(3):689–696. doi: 10.14670/hh-20.689. [DOI] [PubMed] [Google Scholar]
21.Mayeuf-Louchart A. Uncovering the role of glycogen in Brown adipose tissue. Pharm. Res. (N. Y.) 2021;38(1):9–14. doi: 10.1007/s11095-020-02979-6. [DOI] [PubMed] [Google Scholar]
22.Bobryshev Y.V., et al. Changes of lysosomes in the earliest stages of the development of atherosclerosis. J. Cell Mol. Med. 2013;17(5):626–635. doi: 10.1111/jcmm.12042. [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Liu Z., et al. XBP1 deficiency promotes hepatocyte pyroptosis by impairing mitophagy to activate mtDNA-cGAS-STING signaling in macrophages during acute liver injury. Redox Biol. 2022;52 doi: 10.1016/j.redox.2022.102305. [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Kantzer C.G., et al. Anti-ACSA-2 defines a novel monoclonal antibody for prospective isolation of living neonatal and adult astrocytes. Glia. 2017;65(6):990–1004. doi: 10.1002/glia.23140. [DOI] [PubMed] [Google Scholar]
25.Zhang C., et al. Near-infrared oxidative phosphorylation inhibitor integrates acute myeloid leukemia-targeted imaging and therapy. Sci. Adv. 2021;7(1) doi: 10.1126/sciadv.abb6104. [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Zhao H., et al. Glycophagy: an emerging target in pathology. Clin. Chim. Acta. 2018;484:298–303. doi: 10.1016/j.cca.2018.06.014. [DOI] [PubMed] [Google Scholar]
27.Boyer-Guittaut M., et al. The role of GABARAPL1/GEC1 in autophagic flux and mitochondrial quality control in MDA-MB-436 breast cancer cells. Autophagy. 2014;10(6):986–1003. doi: 10.4161/auto.28390. [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Vo M.T., Choi C.Y., Choi Y.B. The mitophagy receptor NIX induces vIRF-1 oligomerization and interaction with GABARAPL1 for the promotion of HHV-8 reactivation-induced mitophagy. PLoS Pathog. 2023;19(7) doi: 10.1371/journal.ppat.1011548. [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Mandl J., Bánhegyi G. The er - glycogen particle - phagophore triangle: a hub connecting glycogenolysis and glycophagy? Pathol. Oncol. Res. 2018;24(4):821–826. doi: 10.1007/s12253-018-0446-0. [DOI] [PubMed] [Google Scholar]
30.Shi Q., et al. Increased glucose metabolism in TAMs fuels O-GlcNAcylation of lysosomal Cathepsin B to promote cancer metastasis and chemoresistance. Cancer Cell. 2022;40(10):1207–1222.e10. doi: 10.1016/j.ccell.2022.08.012. [DOI] [PubMed] [Google Scholar]
31.Jung S., et al. Overexpression of phosphoribosyl pyrophosphate synthase enhances resistance of chlamydomonas to ionizing radiation. Front. Plant Sci. 2021;12 doi: 10.3389/fpls.2021.719846. [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Dringen R. Metabolism and functions of glutathione in brain. Prog. Neurobiol. 2000;62(6):649–671. doi: 10.1016/s0301-0082(99)00060-x. [DOI] [PubMed] [Google Scholar]
33.Aman Y., et al. Autophagy in healthy aging and disease. Nat Aging. 2021;1(8):634–650. doi: 10.1038/s43587-021-00098-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Pederson B.A. Structure and regulation of glycogen synthase in the brain. Adv Neurobiol. 2019;23:83–123. doi: 10.1007/978-3-030-27480-1_3. [DOI] [PubMed] [Google Scholar]
35.Mathieu C., Dupret J.M., Rodrigues-Lima F. The structure and the regulation of glycogen phosphorylases in brain. Adv Neurobiol. 2019;23:125–145. doi: 10.1007/978-3-030-27480-1_4. [DOI] [PubMed] [Google Scholar]
36.Eichler J. Protein glycosylation. Curr. Biol. 2019;29(7):R229–r231. doi: 10.1016/j.cub.2019.01.003. [DOI] [PubMed] [Google Scholar]
37.Varone E., et al. ERO1 alpha deficiency impairs angiogenesis by increasing N-glycosylation of a proangiogenic VEGFA. Redox Biol. 2022;56 doi: 10.1016/j.redox.2022.102455. [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Ozcan S., Andrali S.S., Cantrell J.E. Modulation of transcription factor function by O-GlcNAc modification. Biochim. Biophys. Acta. 2010;1799(5–6):353–364. doi: 10.1016/j.bbagrm.2010.02.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Zhang H., et al. O-GlcNAc modification mediates aquaporin 3 to coordinate endometrial cell glycolysis and affects embryo implantation. J. Adv. Res. 2022;37:119–131. doi: 10.1016/j.jare.2021.06.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Chung S.S., et al. Activation of PPARgamma negatively regulates O-GlcNAcylation of Sp1. Biochem. Biophys. Res. Commun. 2008;372(4):713–718. doi: 10.1016/j.bbrc.2008.05.096. [DOI] [PubMed] [Google Scholar]
41.Hardivillé S., et al. TATA-box binding protein O-GlcNAcylation at T114 regulates formation of the B-tfiid complex and is critical for metabolic gene regulation. Mol. Cell. 2020;77(5):1143–1152.e7. doi: 10.1016/j.molcel.2019.11.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Kuo M., et al. O-glycosylation of FoxO1 increases its transcriptional activity towards the glucose 6-phosphatase gene. FEBS Lett. 2008;582(5):829–834. doi: 10.1016/j.febslet.2008.02.010. [DOI] [PubMed] [Google Scholar]
43.Choi H., et al. Amyloid β-induced elevation of O-GlcNAcylated c-Fos promotes neuronal cell death. Aging Cell. 2019;18(1) doi: 10.1111/acel.12872. [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Liu Y., et al. Resistance to bortezomib in breast cancer cells that downregulate Bim through FOXA1 O-GlcNAcylation. J. Cell. Physiol. 2019;234(10):17527–17537. doi: 10.1002/jcp.28376. [DOI] [PubMed] [Google Scholar]
45.Ge X.L., et al. Inhibition of miR-19a protects neurons against ischemic stroke through modulating glucose metabolism and neuronal apoptosis. Cell. Mol. Biol. Lett. 2019;24:37. doi: 10.1186/s11658-019-0160-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Takahashi S. Neuroprotective function of high glycolytic activity in astrocytes: common roles in stroke and neurodegenerative diseases. Int. J. Mol. Sci. 2021;22(12) doi: 10.3390/ijms22126568. [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Qiang S.J., et al. The discovery of novel PGK1 activators as apoptotic inhibiting and neuroprotective agents. Front. Pharmacol. 2022;13 doi: 10.3389/fphar.2022.877706. [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Sun T., et al. Emerging role of metabolic reprogramming in hyperoxia-associated neonatal diseases. Redox Biol. 2023;66 doi: 10.1016/j.redox.2023.102865. [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Dringen R., et al. Aminopeptidase N mediates the utilization of the GSH precursor CysGly by cultured neurons. J. Neurosci. Res. 2001;66(5):1003–1008. doi: 10.1002/jnr.10042. [DOI] [PubMed] [Google Scholar]
50.Itoh Y., et al. Dichloroacetate effects on glucose and lactate oxidation by neurons and astroglia in vitro and on glucose utilization by brain in vivo. Proc. Natl. Acad. Sci. U.S.A. 2003;100(8):4879–4884. doi: 10.1073/pnas.0831078100. [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Bouzier-Sore A.K., et al. Competition between glucose and lactate as oxidative energy substrates in both neurons and astrocytes: a comparative NMR study. Eur. J. Neurosci. 2006;24(6):1687–1694. doi: 10.1111/j.1460-9568.2006.05056.x. [DOI] [PubMed] [Google Scholar]
52.Koepsell H. Glucose transporters in brain in health and disease. Pflügers Archiv. 2020;472(9):1299–1343. doi: 10.1007/s00424-020-02441-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Wang Z., et al. Increasing O-GlcNAcylation is neuroprotective in young and aged brains after ischemic stroke. Exp. Neurol. 2021;339 doi: 10.1016/j.expneurol.2021.113646. [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Parkinson W., et al. N-glycosylation requirements in neuromuscular synaptogenesis. Development. 2013;140(24):4970–4981. doi: 10.1242/dev.099192. [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Wheatley E.G., et al. Neuronal O-GlcNAcylation improves cognitive function in the aged mouse brain. Curr. Biol. 2019;29(20):3359–3369.e4. doi: 10.1016/j.cub.2019.08.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Trinidad J.C., et al. Global identification and characterization of both O-GlcNAcylation and phosphorylation at the murine synapse. Mol. Cell. Proteomics. 2012;11(8):215–229. doi: 10.1074/mcp.O112.018366. [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Lee B.E., et al. O-GlcNAcylation regulates dopamine neuron function, survival and degeneration in Parkinson disease. Brain. 2020;143(12):3699–3716. doi: 10.1093/brain/awaa320. [DOI] [PMC free article] [PubMed] [Google Scholar]
58.de Carvalho T.S., et al. Hypocaloric diet initiated post-ischemia provides long-term neuroprotection and promotes peri-infarct brain remodeling by regulating metabolic and survival-promoting proteins. Mol. Neurobiol. 2021;58(4):1491–1503. doi: 10.1007/s12035-020-02207-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
59.Fann D.Y., et al. Intermittent fasting attenuates inflammasome activity in ischemic stroke. Exp. Neurol. 2014;257:114–119. doi: 10.1016/j.expneurol.2014.04.017. [DOI] [PubMed] [Google Scholar]
60.Magnuson B., Peppard A., Auer Flomenhoft D. Hypocaloric considerations in patients with potentially hypometabolic disease States. Nutr. Clin. Pract. 2011;26(3):253–260. doi: 10.1177/0884533611405673. [DOI] [PubMed] [Google Scholar]
61.Zhao J., et al. Safety and efficacy of three enteral feeding strategies in patients with severe stroke in China (OPENS): a multicentre, prospective, randomised, open-label, blinded-endpoint trial. Lancet Neurol. 2022;21(4):319–328. doi: 10.1016/s1474-4422(22)00010-2. [DOI] [PubMed] [Google Scholar]
62.Hu S., et al. Nrf1 is an indispensable redox-determining factor for mitochondrial homeostasis by integrating multi-hierarchical regulatory networks. Redox Biol. 2022;57 doi: 10.1016/j.redox.2022.102470. [DOI] [PMC free article] [PubMed] [Google Scholar]
63.Yang Y., et al. Acetate upregulates GPR43 expression and function via PI3K-AKT-SP1 signaling in mammary epithelial cells during milk fat synthesis. J. Agric. Food Chem. 2023;71(43):16003–16015. doi: 10.1021/acs.jafc.3c00965. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Multimedia component 1

mmc1.docx^{(3.4MB, docx)}

Data Availability Statement

Data will be made available on request.

[bib1] 1.Global, regional, and national burden of stroke and its risk factors, 1990-2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet Neurol. 2021;20(10):795–820. doi: 10.1016/s1474-4422(21)00252-0. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib2] 2.Mendelson S.J., Prabhakaran S. Diagnosis and management of transient ischemic attack and acute ischemic stroke: a review. JAMA. 2021;325(11):1088–1098. doi: 10.1001/jama.2020.26867. [DOI] [PubMed] [Google Scholar]

[bib3] 3.Bar B., Biller J. Select hyperacute complications of ischemic stroke: cerebral edema, hemorrhagic transformation, and orolingual angioedema secondary to intravenous Alteplase. Expert Rev. Neurother. 2018;18(10):749–759. doi: 10.1080/14737175.2018.1521723. [DOI] [PubMed] [Google Scholar]

[bib4] 4.Pundik S., Xu K., Sundararajan S. Reperfusion brain injury: focus on cellular bioenergetics. Neurology. 2012;79(13 Suppl 1):S44–S51. doi: 10.1212/WNL.0b013e3182695a14. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib5] 5.Liu M., et al. TIGAR alleviates oxidative stress in brain with extended ischemia via a pentose phosphate pathway-independent manner. Redox Biol. 2022;53 doi: 10.1016/j.redox.2022.102323. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib6] 6.Koutsifeli P., et al. Glycogen-autophagy: molecular machinery and cellular mechanisms of glycophagy. J. Biol. Chem. 2022;298(7) doi: 10.1016/j.jbc.2022.102093. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib7] 7.Delbridge L.M., et al. Myocardial autophagic energy stress responses--macroautophagy, mitophagy, and glycophagy. Am. J. Physiol. Heart Circ. Physiol. 2015;308(10):H1194–H1204. doi: 10.1152/ajpheart.00002.2015. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib8] 8.Chandramouli C., et al. Myocardial glycogen dynamics: new perspectives on disease mechanisms. Clin. Exp. Pharmacol. Physiol. 2015;42(4):415–425. doi: 10.1111/1440-1681.12370. [DOI] [PubMed] [Google Scholar]

[bib9] 9.Heden T.D., et al. Regulation and role of glycophagy in skeletal muscle energy metabolism. Autophagy. 2022;18(5):1078–1089. doi: 10.1080/15548627.2021.1969633. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib10] 10.Mayeuf-Louchart A., et al. Glycogen dynamics drives lipid droplet biogenesis during Brown adipocyte differentiation. Cell Rep. 2019;29(6):1410–1418.e6. doi: 10.1016/j.celrep.2019.09.073. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib11] 11.Brewer M.K., Gentry M.S. Brain glycogen structure and its associated proteins: past, present and future. Adv Neurobiol. 2019;23:17–81. doi: 10.1007/978-3-030-27480-1_2. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib12] 12.Palhegyi A.M., et al. Biomedical implications of autophagy in macromolecule storage disorders. Front. Cell Dev. Biol. 2019;7:179. doi: 10.3389/fcell.2019.00179. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib13] 13.Reichelt M.E., et al. Myocardial glycophagy - a specific glycogen handling response to metabolic stress is accentuated in the female heart. J. Mol. Cell. Cardiol. 2013;65:67–75. doi: 10.1016/j.yjmcc.2013.09.014. [DOI] [PubMed] [Google Scholar]

[bib14] 14.Mellor K.M., et al. Cardiomyocyte glycophagy is regulated by insulin and exposure to high extracellular glucose. Am. J. Physiol. Heart Circ. Physiol. 2014;306(8):H1240–H1245. doi: 10.1152/ajpheart.00059.2014. [DOI] [PubMed] [Google Scholar]

[bib15] 15.Sun R.C., et al. Brain glycogen serves as a critical glucosamine cache required for protein glycosylation. Cell Metabol. 2021;33(7):1404–1417.e9. doi: 10.1016/j.cmet.2021.05.003. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib16] 16.Oe Y., Akther S., Hirase H. Regional distribution of glycogen in the mouse brain visualized by immunohistochemistry. Adv Neurobiol. 2019;23:147–168. doi: 10.1007/978-3-030-27480-1_5. [DOI] [PubMed] [Google Scholar]

[bib17] 17.Bélanger M., Allaman I., Magistretti P.J. Brain energy metabolism: focus on astrocyte-neuron metabolic cooperation. Cell Metabol. 2011;14(6):724–738. doi: 10.1016/j.cmet.2011.08.016. [DOI] [PubMed] [Google Scholar]

[bib18] 18.Cai Y., et al. Glycogenolysis is crucial for astrocytic glycogen accumulation and brain damage after reperfusion in ischemic stroke. iScience. 2020;23(5) doi: 10.1016/j.isci.2020.101136. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib19] 19.Guo H., et al. Astrocytic glycogen mobilization participates in salvianolic acid B-mediated neuroprotection against reperfusion injury after ischemic stroke. Exp. Neurol. 2022;349 doi: 10.1016/j.expneurol.2021.113966. [DOI] [PubMed] [Google Scholar]

[bib20] 20.Kondomerkos D.J., et al. Glycogen autophagy in the liver and heart of newborn rats. The effects of glucagon, adrenalin or rapamycin. Histol. Histopathol. 2005;20(3):689–696. doi: 10.14670/hh-20.689. [DOI] [PubMed] [Google Scholar]

[bib21] 21.Mayeuf-Louchart A. Uncovering the role of glycogen in Brown adipose tissue. Pharm. Res. (N. Y.) 2021;38(1):9–14. doi: 10.1007/s11095-020-02979-6. [DOI] [PubMed] [Google Scholar]

[bib22] 22.Bobryshev Y.V., et al. Changes of lysosomes in the earliest stages of the development of atherosclerosis. J. Cell Mol. Med. 2013;17(5):626–635. doi: 10.1111/jcmm.12042. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib23] 23.Liu Z., et al. XBP1 deficiency promotes hepatocyte pyroptosis by impairing mitophagy to activate mtDNA-cGAS-STING signaling in macrophages during acute liver injury. Redox Biol. 2022;52 doi: 10.1016/j.redox.2022.102305. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib24] 24.Kantzer C.G., et al. Anti-ACSA-2 defines a novel monoclonal antibody for prospective isolation of living neonatal and adult astrocytes. Glia. 2017;65(6):990–1004. doi: 10.1002/glia.23140. [DOI] [PubMed] [Google Scholar]

[bib25] 25.Zhang C., et al. Near-infrared oxidative phosphorylation inhibitor integrates acute myeloid leukemia-targeted imaging and therapy. Sci. Adv. 2021;7(1) doi: 10.1126/sciadv.abb6104. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib26] 26.Zhao H., et al. Glycophagy: an emerging target in pathology. Clin. Chim. Acta. 2018;484:298–303. doi: 10.1016/j.cca.2018.06.014. [DOI] [PubMed] [Google Scholar]

[bib27] 27.Boyer-Guittaut M., et al. The role of GABARAPL1/GEC1 in autophagic flux and mitochondrial quality control in MDA-MB-436 breast cancer cells. Autophagy. 2014;10(6):986–1003. doi: 10.4161/auto.28390. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib28] 28.Vo M.T., Choi C.Y., Choi Y.B. The mitophagy receptor NIX induces vIRF-1 oligomerization and interaction with GABARAPL1 for the promotion of HHV-8 reactivation-induced mitophagy. PLoS Pathog. 2023;19(7) doi: 10.1371/journal.ppat.1011548. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib29] 29.Mandl J., Bánhegyi G. The er - glycogen particle - phagophore triangle: a hub connecting glycogenolysis and glycophagy? Pathol. Oncol. Res. 2018;24(4):821–826. doi: 10.1007/s12253-018-0446-0. [DOI] [PubMed] [Google Scholar]

[bib30] 30.Shi Q., et al. Increased glucose metabolism in TAMs fuels O-GlcNAcylation of lysosomal Cathepsin B to promote cancer metastasis and chemoresistance. Cancer Cell. 2022;40(10):1207–1222.e10. doi: 10.1016/j.ccell.2022.08.012. [DOI] [PubMed] [Google Scholar]

[bib31] 31.Jung S., et al. Overexpression of phosphoribosyl pyrophosphate synthase enhances resistance of chlamydomonas to ionizing radiation. Front. Plant Sci. 2021;12 doi: 10.3389/fpls.2021.719846. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib32] 32.Dringen R. Metabolism and functions of glutathione in brain. Prog. Neurobiol. 2000;62(6):649–671. doi: 10.1016/s0301-0082(99)00060-x. [DOI] [PubMed] [Google Scholar]

[bib33] 33.Aman Y., et al. Autophagy in healthy aging and disease. Nat Aging. 2021;1(8):634–650. doi: 10.1038/s43587-021-00098-4. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib34] 34.Pederson B.A. Structure and regulation of glycogen synthase in the brain. Adv Neurobiol. 2019;23:83–123. doi: 10.1007/978-3-030-27480-1_3. [DOI] [PubMed] [Google Scholar]

[bib35] 35.Mathieu C., Dupret J.M., Rodrigues-Lima F. The structure and the regulation of glycogen phosphorylases in brain. Adv Neurobiol. 2019;23:125–145. doi: 10.1007/978-3-030-27480-1_4. [DOI] [PubMed] [Google Scholar]

[bib36] 36.Eichler J. Protein glycosylation. Curr. Biol. 2019;29(7):R229–r231. doi: 10.1016/j.cub.2019.01.003. [DOI] [PubMed] [Google Scholar]

[bib37] 37.Varone E., et al. ERO1 alpha deficiency impairs angiogenesis by increasing N-glycosylation of a proangiogenic VEGFA. Redox Biol. 2022;56 doi: 10.1016/j.redox.2022.102455. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib38] 38.Ozcan S., Andrali S.S., Cantrell J.E. Modulation of transcription factor function by O-GlcNAc modification. Biochim. Biophys. Acta. 2010;1799(5–6):353–364. doi: 10.1016/j.bbagrm.2010.02.005. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib39] 39.Zhang H., et al. O-GlcNAc modification mediates aquaporin 3 to coordinate endometrial cell glycolysis and affects embryo implantation. J. Adv. Res. 2022;37:119–131. doi: 10.1016/j.jare.2021.06.022. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib40] 40.Chung S.S., et al. Activation of PPARgamma negatively regulates O-GlcNAcylation of Sp1. Biochem. Biophys. Res. Commun. 2008;372(4):713–718. doi: 10.1016/j.bbrc.2008.05.096. [DOI] [PubMed] [Google Scholar]

[bib41] 41.Hardivillé S., et al. TATA-box binding protein O-GlcNAcylation at T114 regulates formation of the B-tfiid complex and is critical for metabolic gene regulation. Mol. Cell. 2020;77(5):1143–1152.e7. doi: 10.1016/j.molcel.2019.11.022. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib42] 42.Kuo M., et al. O-glycosylation of FoxO1 increases its transcriptional activity towards the glucose 6-phosphatase gene. FEBS Lett. 2008;582(5):829–834. doi: 10.1016/j.febslet.2008.02.010. [DOI] [PubMed] [Google Scholar]

[bib43] 43.Choi H., et al. Amyloid β-induced elevation of O-GlcNAcylated c-Fos promotes neuronal cell death. Aging Cell. 2019;18(1) doi: 10.1111/acel.12872. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib44] 44.Liu Y., et al. Resistance to bortezomib in breast cancer cells that downregulate Bim through FOXA1 O-GlcNAcylation. J. Cell. Physiol. 2019;234(10):17527–17537. doi: 10.1002/jcp.28376. [DOI] [PubMed] [Google Scholar]

[bib45] 45.Ge X.L., et al. Inhibition of miR-19a protects neurons against ischemic stroke through modulating glucose metabolism and neuronal apoptosis. Cell. Mol. Biol. Lett. 2019;24:37. doi: 10.1186/s11658-019-0160-2. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib46] 46.Takahashi S. Neuroprotective function of high glycolytic activity in astrocytes: common roles in stroke and neurodegenerative diseases. Int. J. Mol. Sci. 2021;22(12) doi: 10.3390/ijms22126568. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib47] 47.Qiang S.J., et al. The discovery of novel PGK1 activators as apoptotic inhibiting and neuroprotective agents. Front. Pharmacol. 2022;13 doi: 10.3389/fphar.2022.877706. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib48] 48.Sun T., et al. Emerging role of metabolic reprogramming in hyperoxia-associated neonatal diseases. Redox Biol. 2023;66 doi: 10.1016/j.redox.2023.102865. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib49] 49.Dringen R., et al. Aminopeptidase N mediates the utilization of the GSH precursor CysGly by cultured neurons. J. Neurosci. Res. 2001;66(5):1003–1008. doi: 10.1002/jnr.10042. [DOI] [PubMed] [Google Scholar]

[bib50] 50.Itoh Y., et al. Dichloroacetate effects on glucose and lactate oxidation by neurons and astroglia in vitro and on glucose utilization by brain in vivo. Proc. Natl. Acad. Sci. U.S.A. 2003;100(8):4879–4884. doi: 10.1073/pnas.0831078100. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib51] 51.Bouzier-Sore A.K., et al. Competition between glucose and lactate as oxidative energy substrates in both neurons and astrocytes: a comparative NMR study. Eur. J. Neurosci. 2006;24(6):1687–1694. doi: 10.1111/j.1460-9568.2006.05056.x. [DOI] [PubMed] [Google Scholar]

[bib52] 52.Koepsell H. Glucose transporters in brain in health and disease. Pflügers Archiv. 2020;472(9):1299–1343. doi: 10.1007/s00424-020-02441-x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib53] 53.Wang Z., et al. Increasing O-GlcNAcylation is neuroprotective in young and aged brains after ischemic stroke. Exp. Neurol. 2021;339 doi: 10.1016/j.expneurol.2021.113646. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib54] 54.Parkinson W., et al. N-glycosylation requirements in neuromuscular synaptogenesis. Development. 2013;140(24):4970–4981. doi: 10.1242/dev.099192. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib55] 55.Wheatley E.G., et al. Neuronal O-GlcNAcylation improves cognitive function in the aged mouse brain. Curr. Biol. 2019;29(20):3359–3369.e4. doi: 10.1016/j.cub.2019.08.003. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib56] 56.Trinidad J.C., et al. Global identification and characterization of both O-GlcNAcylation and phosphorylation at the murine synapse. Mol. Cell. Proteomics. 2012;11(8):215–229. doi: 10.1074/mcp.O112.018366. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib57] 57.Lee B.E., et al. O-GlcNAcylation regulates dopamine neuron function, survival and degeneration in Parkinson disease. Brain. 2020;143(12):3699–3716. doi: 10.1093/brain/awaa320. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib58] 58.de Carvalho T.S., et al. Hypocaloric diet initiated post-ischemia provides long-term neuroprotection and promotes peri-infarct brain remodeling by regulating metabolic and survival-promoting proteins. Mol. Neurobiol. 2021;58(4):1491–1503. doi: 10.1007/s12035-020-02207-7. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib59] 59.Fann D.Y., et al. Intermittent fasting attenuates inflammasome activity in ischemic stroke. Exp. Neurol. 2014;257:114–119. doi: 10.1016/j.expneurol.2014.04.017. [DOI] [PubMed] [Google Scholar]

[bib60] 60.Magnuson B., Peppard A., Auer Flomenhoft D. Hypocaloric considerations in patients with potentially hypometabolic disease States. Nutr. Clin. Pract. 2011;26(3):253–260. doi: 10.1177/0884533611405673. [DOI] [PubMed] [Google Scholar]

[bib61] 61.Zhao J., et al. Safety and efficacy of three enteral feeding strategies in patients with severe stroke in China (OPENS): a multicentre, prospective, randomised, open-label, blinded-endpoint trial. Lancet Neurol. 2022;21(4):319–328. doi: 10.1016/s1474-4422(22)00010-2. [DOI] [PubMed] [Google Scholar]

[bib62] 62.Hu S., et al. Nrf1 is an indispensable redox-determining factor for mitochondrial homeostasis by integrating multi-hierarchical regulatory networks. Redox Biol. 2022;57 doi: 10.1016/j.redox.2022.102470. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib63] 63.Yang Y., et al. Acetate upregulates GPR43 expression and function via PI3K-AKT-SP1 signaling in mammary epithelial cells during milk fat synthesis. J. Agric. Food Chem. 2023;71(43):16003–16015. doi: 10.1021/acs.jafc.3c00965. [DOI] [PubMed] [Google Scholar]

PERMALINK

Dysfunction of astrocytic glycophagy exacerbates reperfusion injury in ischemic stroke

Haiyun Guo

Yumeng Li

Shiquan Wang

Yongheng Yang

Tiantian Xu

Jianshuai Zhao

Jin Wang

Wenqiang Zuo

Pengju Wang

Guangchao Zhao

Huaning Wang

Wugang Hou

Hailong Dong

Yanhui Cai

Abstract

Graphical abstract

Highlights

1. Introduction

2. Materials and methods

2.1. Human samples

2.2. Mouse

2.3. Hypocaloric diet

2.4. Primary astrocyte and neuron culture

2.5. Astrocyte–neuron coculture system

2.6. Low-glucose culture

2.7. I/R model in mice

2.8. In vitro I/R model

2.9. Neurobehavioral analysis

2.10. Immunofluorescence and immunocytochemistry staining

2.11. Immunoblotting

2.12. RT-qPCR

2.13. Triphenyl tetrazolium chloride (TTC) staining

2.14. Golgi staining

2.15. Electron microscopy analysis

2.16. Metabolite analysis

2.17. Glycogen levels in total cells

2.18. Glycogen levels in the lysosome

2.19. Enzyme activity analysis

2.20. Nuclear and cytoplasmic protein extraction

2.21. Immunoprecipitation

2.22. Magnetic-activated cell sorting (MACS)

2.23. O-GlcNAcylation analysis

2.24. Comet assay

2.25. Reverse chromatin immunoprecipitation (ChiP)-mass spectrometry (MS)

2.26. Metabolomics analysis

2.27. 13C-labeled glucose metabolic flux analysis

2.28. RNA sequencing

2.29. GABARAPL1 overexpression model in vitro

2.30. GABARAPL1 knockdown model in vitro

2.31. GAA knockdown model in vitro

2.32. GAA knockdown mouse model

2.33. In vitro γ-GT knockdown models

2.34. In vitro specificity protein 1 (SP1) amino acid mutant model and TATA binding protein (TBP) amino acid mutant model

2.35. In vitro rGAA eukaryotic expression system

2.36. Dual-luciferase reporter system

2.37. Cell survival analysis

2.38. Antibodies

2.39. Drug

2.40. Statistics

3. Results

3.1. Astrocytic glycophagy is dysfunctional due to GABARAPL1 downregulation during cerebral reperfusion

Fig. 1.

Fig. 2.

3.2. Astrocytic PI3K-Akt pathway activation participates in driving GABARAPL1 downregulation during cerebral reperfusion

Fig. 3.

3.3. Decreases in glucosamine levels mediated by glycophagic dysfunction aggravate GABARAPL1 downregulation in astrocytes during cerebral reperfusion

Fig. 4.

3.4. Restoration of astrocytic glycophagy ameliorates reperfusion injury after ischemic stroke

Fig. 5.

3.5. Glycophagy restoration reduces oxidative injury and DNA damage in astrocytes during cerebral reperfusion

Fig. 6.

3.6. Restoration of astrocytic glycophagy accelerates neighboring neuron recovery via astrocyte-neuron interactions between GSH and glucosamine during cerebral reperfusion

Fig. 7.

3.7. A hypocaloric diet enhances astrocytic glycophagy and is neuroprotective against reperfusion injury after ischemic stroke

Fig. 8.

4. Discussion

CRediT authorship contribution statement

Declaration of competing interest

2.27. ¹³C-labeled glucose metabolic flux analysis