Skip to main content
NCBI home page
Springer logoLink to Springer
. 2025 Jan 10;41(1):28. doi: 10.1007/s10565-025-09986-6

SENP1 inhibits aerobic glycolysis in Aβ1-42-incubated astrocytes by promoting PUM2 deSUMOylation

Qianshuo Liu 1,#, Meixi Jiang 1,#, Zhengze Wang 1, Jihong Meng 2, Hui jia 3, Jing Li 1, Jiacai Lin 4,✉, Libin Guo 5,✉, Lianbo Gao 1,✉
PMCID: PMC11723902  PMID: 39794619

Abstract

Alzheimer’s disease (AD), the most prevalent form of dementia in the elderly, involves critical changes such as reduced aerobic glycolysis in astrocytes and increased neuronal apoptosis, both of which are significant in the disease’s pathology. In our study, astrocytes treated with amyloid β1-42 (Aβ1-42) to simulate AD conditions exhibited upregulated expressions of small ubiquitin-like modifier (SUMO)-specific protease 1 (SENP1) and Pumilio RNA Binding Family Member 2 (PUM2), alongside decreased levels of Nuclear factor erythroid 2-related factor 2 (NRF2). SENP1 is notably the most upregulated SUMOylation enzyme in Aβ1-42-exposed astrocytes. Functional assays including Ni2+-Nitrilotriacetic acid (NTA) agarose bead pull-down and co-immunoprecipitation (Co-IP) confirmed SENP1’s role in actively deSUMOylating PUM2, thereby enhancing its stability and expression. The interaction between PUM2 and the 3’ untranslated region (3’UTR) of NRF2 mRNA reduces NRF2 levels, subsequently diminishing the transcriptional activation of critical glycolytic enzymes, Hexokinase 1 (HK1) and Glucose Transporter 1 (GLUT1). These changes contribute to the observed reduction in glycolytic function in astrocytes, exacerbating neuronal apoptosis. Targeted interventions, such as knockdown of Senp1 or Pum2 or overexpression of NRF2 in APPswe/PSEN1dE9 (APP/PS1) transgenic mice, effectively increased HK1 and GLUT1 levels, decreased apoptosis, and alleviated cognitive impairment. These findings highlight the important roles of the SENP1/PUM2/NRF2 pathway in influencing glucose metabolism in astrocytes, presenting new potential therapeutic targets for AD.

Graphical Abstract

1. SENP1 deSUMOylates PUM2 enhancing its expression in Aβ1-42-microenvironment.

2. PUM2 binds to NRF2 mRNA reducing its stability and expression in Aβ1-42-microenvironment.

3. Reduced NRF2 reduces its transcriptional promoting effects on HK1 and GLUT1 in Aβ1-42-microenvironment.

4. Reduced HK1 and GLUT1 decreases aerobic glycolysis in astrocytes, leading to an increase in co-culture neuronal apoptosis in Aβ1-42-microenvironment.

graphic file with name 10565_2025_9986_Figa_HTML.jpg

Supplementary Information

The online version contains supplementary material available at 10.1007/s10565-025-09986-6.

Keywords: Aerobic glycolysis, Alzheimer’s disease, DeSUMOylation, SENP1, PUM2, NRF2

Introduction

Alzheimer’s disease (AD), increasingly prevalent among the elderly, is distinguished by multiple neuropathological transformations including extracellular β-amyloid peptide (Aβ) accumulation, hyperphosphorylated tau protein aggregation, and progressive neuronal degradation (Ossenkoppele et al. 2022). A pivotal metabolic process in astrocytes, aerobic glycolysis, not only facilitates energy creation, transmission, and storage in the brain but also supports neuronal function and survival through lactate production (Le Douce et al. 2020). Significant reductions in aerobic glycolysis in astrocytes, alongside neuronal apoptosis, have been identified as major factors contributing to cognitive decline in AD (Pang et al. 2022; Traxler et al. 2022). Initial stages of AD often reveal diminished aerobic glycolysis and lactate production in astrocytes, which in turn instigates inflammatory neuronal changes, disrupts synaptic operations, and intensifies neuronal degradation and cell death (Chao et al. 2019; Cunnane et al. 2020). In addition, reduced aerobic glycolytic activity in astrocytes may disrupt neurotransmitter metabolism, potentially triggering additional neuronal apoptosis (Yamashita et al. 2021).

Aβ notably impairs aerobic glycolysis in astrocytes, leading to insufficient energy supply, accelerating astrocyte aging, altering metabolic activity to produce toxic byproducts, enhancing ROS generation, and triggering oxidative stress and programmed cell death (Habib et al. 2020). These alterations collectively mimic the pathological environment of AD (Habib et al. 2020; Hampel et al. 2021; Kim et al. 2024b; Mossmann et al. 2014). Given Aβ1–42’s significant neurotoxicity, many studies have established AD cell models by incubating cells with this peptide (Habib et al. 2020). Therefore, this study used Aβ1–42 to incubate astrocytes and neuronal cells to establish AD in vitro cell models.

Hexokinase (HK), essential for aerobic glycolysis, phosphorylates glucose to glucose-6-phosphate, primarily through HK1 in brain astrocytes (Chen et al. 2022). The glucose transporter 1 (GLUT1), integral for glucose crossing in glycolysis, marks the first rate-limiting phase of aerobic glycolysis (Veys et al. 2020). Decreases in HK1 and GLUT1 expressions are observed in early AD, correlating with reduced aerobic glycolysis, neuronal dysfunction, and behavioral deficits (Zhang et al. 2020).

Protein SUMOylation represents a crucial post-translational modification process where proteins are tagged by small ubiquitin-like modifier (SUMO) proteins. This modification is dynamic and reversible, orchestrated by an enzyme cascade with SUMO-specific proteases (SENPs) facilitating the removal of SUMO from substrates (Shangguan et al. 2021). SENP1, an integral member of this protease family, plays a multifaceted role in several physiological and pathological processes including aerobic glycolysis, neurodegeneration, cell death, and inflammation through its ability to deSUMOylate specific targets (Dangoumau et al. 2021; Shangguan et al. 2021; Yan et al. 2022). For instance, SENP1 is known to deSUMOylate HIF-1α, significantly enhancing the protein’s stability (Yan et al. 2022). Elevated SENP1 expression in AD model hippocampus correlates with impaired spatial learning and memory (Chen et al. 2021). Furthermore, SENP1 influences metabolic pathways by modulating the interaction between glycolytic enzymes and mitochondria, leading to metabolic reprogramming that can affect disease states such as prostate cancer (Shangguan et al. 2021).

Pumilio RNA Binding Family Member 2 (PUM2) is a distinctive RNA binding protein within the PUF family, recognized for its significant role in cellular aging (Kopp et al. 2019). PUM2 is crucial in embryonic development and cell differentiation, where it facilitates translation and mRNA degradation (Lin et al. 2023; Uyhazi et al. 2020). Nuclear factor erythroid 2-related factor 2 (NRF2), a member of the basic leucine zipper protein family, serves as a vital transcription factor (Ma 2013). It plays a pivotal role in preventing cellular aging, reducing oxidative stress, and mitigating inflammation (George et al. 2022; Nakano-Kobayashi et al. 2023). Furthermore, NRF2 has been shown to enhance glycolysis in diverse cell types, including acute myeloid leukemia cells, macrophages in chronic obstructive pulmonary disease, and tumor cells (Chu et al. 2023; Ryan et al. 2023; Zhang et al. 2024). Importantly, recent studies suggest that NRF2 possesses a neuroprotective function in AD, enhancing antioxidative defenses, detoxification processes, and autophagy (Bai et al. 2023; Calabrese and Kozumbo 2021; Lane et al. 2021). However, the specific effects of NRF2 on aerobic glycolysis in AD remain to be fully elucidated.

In our study, we investigated the expression levels of SENP1, PUM2, and NRF2 in astrocytes after exposure to Aβ1–42. Our findings indicate that SENP1 stimulates the expression of PUM2 by increasing its deSUMOylation in astrocytes treated with Aβ1–42. Overexpressed PUM2 destabilizes NRF2 mRNA, which in turn diminishes aerobic glycolysis in astrocytes by downregulating the transcription of HK1 and GLUT1. This reduction in aerobic glycolysis exacerbates neuronal apoptosis. Our research identifies a novel therapeutic strategy for AD that involves reprogramming glucose metabolism in astrocytes, potentially offering new avenues for treatment.

Materials and methods

Cell culture procedures

Human astrocytes (HAs) (ZY-1028, Zeye, Shanghai, CN) and human neuroblastoma cell line SH-SY5Y (ZY-C6068H, Zeye, Shanghai, CN) were acquired from ScienCell Research Laboratories, California, USA. HAs were cultured in astrocyte medium containing 10% fetal bovine serum (FBS) (#0500, ScienCell, CA, USA), Astrocyte Growth Supplement, and 1% penicillin–streptomycin solution, all sourced from SienCell. SH-SY5Y cells were maintained in MEM/F12 medium with 15% FBS (#0500, ScienCell, CA, USA). Additionally, human embryonic kidney (HEK)−293 T cells obtained from Shanghai Institutes for Biological Sciences Cell Resource Centre were grown in Dulbecco’s Modified Eagle’s Medium (SH30022.01, HyClone, UT, USA) containing 10% FBS (#0500, ScienCell, CA, USA). Authentication of all cell lines was confirmed through Short Tandem Repeat analysis per the Cell Line Authentication guidelines. Cells were incubated at 37 °C in a humidified atmosphere with 5% CO2. For experimental treatments, HAs were exposed to Aβ1–42 at a concentration of 5 μM for 48 h (Liu et al. 2020).

Experiment design and animal care

Transgenic male mice, APPswe/PSEN1dE9 (APP/PS1) on a C57BL6 background, were sourced from Jackson Laboratory, West Grove, PA, USA. The animal study protocols were approved by Ethics Board for Laboratory Animals at China Medical University (CMU202302073), with a focus on minimizing animal suffering and reducing the number of animals utilized. Mice were maintained under controlled environmental conditions with 50% humidity, temperatures between 22–25 °C, and a 12-h light/dark cycle, and had access to a standard diet and water. We allocated 24 six-month-old male mice randomly into four groups for behavioral studies: APP/PS1 + Vector, APP/PS1 + shSenp1, APP/PS1 + shPum2, and APP/PS1 + Nrf2, with six mice per group. For cellular analyses, including immunohistochemistry fluorescence tests, TdT-mediated dUTP-biotin nick end labeling (TUNEL) assays, and H&E staining, three mice from each group were randomly selected.

5 μM Aβ1–42 preparation

Equilibrate Aβ1–42 (A9810, Sigma-Aldrich, ST. Louis, USA) at room temperature for 30 min, then dissolve in Hexafluoro isopropyl alcohol (HFIP) (#105,228, Sigma-Aldrich, Sigma-Aldrich, ST. Louis, USA) to make a 1 mM solution and incubate for another 30 min. Evaporate HFIP overnight to yield an Aβ1–42 film, centrifuge to remove residue, and store film at −20 °C. Before use, resuspend the film in Dimethyl sulfoxide for a 5 mM stock solution and dilute to 5 μM with physiological saline.

Intracerebroventricular injection

We employed recombinant adeno-associated virus serotype 9 (AAV9) to modulate gene expression in astrocytes of APP/PS1 mice, aiming to either repress or overexpress specific genes. To repress gene expression, short-hairpin RNA (shRNA) targeting mouse Senp1 (NM_001379573.1) and Pum2 (NM_001160219.1) were cloned into vector pAKD-CMV-bGlobin-eYFP-H1-shRNA (Obio Technology, Shanghai, CN). For overexpression purposes, full-length Nrf2 gene (NM_001399226.1) was inserted into AAV9-hSyn-Nrf2-YFP vector (Obio Technology, Shanghai, CN). The target sequences for shSenp1 and shPum2 are detailed in Table A4.

For the injection process, mice were anesthetized with 2,2,2-tribromoethanol (M06376-UYX, Balb, Beijing, CN) at a dosage of 1000 mg/kg, administered intravenously. Surgery was initiated once the mice exhibited no movement or response to toe pinches, ensuring deep anesthesia. Intracerebroventricular injections were precisely performed using a stereotaxic apparatus to deliver the virus into cerebral astrocytes, following established protocols (Liu et al. 2021). Behavioral assessments, cellular apoptosis identification, and the evaluation of HK1 and GLUT1 expressions were conducted four weeks after the viral injections, as depicted in Figure S9d.

The Morris water maze

The Morris water maze (MWM), consisting of spatial training trials and probe trials, evaluated hippocampus-dependent spatial navigation learning and memory (Zhang et al. 2023). Refer to Supplemental Materials and Methods for more information.

Probe trials: less frequency to the platform and longer latency to the platform indicate poor memory.

Real-time PCR assays

RNA from cell lines was isolated using Trizol Reagent (15,596,026, Life Technologies Corporation, Paisley, UK). Refer to Supplemental Materials and Methods for more information. Primers are shown in Table A1.

Plasmids and cell transfections

To knock down gene expression, short-hairpin RNAs (shRNAs) targeting SENP1 (NM_001267594.2), PUM2 (NM_001282752.3), and NRF2 (NM_001145412.3) were respectively ligated into LV10 (U6/GFP/Neo) vectors (GenePharma, Shanghai, CN) to construct the shSENP1, shPUM2, and shNRF2 plasmids. Conversely, full-length cDNAs of SENP1, PUM2, and NRF2 were inserted into pIRES2 vectors (GenScript, NJ, USA) to generate plasmids for overexpression. Non-targeting sequences served as negative controls. Astrocytes were transfected using Lipofectamine 3000 (L3000150, Invitrogen, CA, USA) in an Aβ1–42 environment, following the provided protocol, and selected with G418 (0.4 mg/mL, A1720, Sigma-Aldrich, ST. Louis, USA). Sequences of shSENP1, shPUM2, and shNRF2 are shown in Table A2. Co-transfection experiments involved stable transfection with shSENP1 followed by transient overexpression of SENP1 in Aβ1–42-treated astrocytes. Cells were analyzed 48 h post-transfection (Figure S2).

Site-directed mutagenesis was performed using QuikChange Lightning Site-Directed Mutagenesis Kits (210,519, Stratagene, VA, USA) to create PUM2-mut (K91/985R). V5-PUM2 was obtained by using standard cloning procedures that amplified and subcloned human PUM2 into pcDNA4-TO-V5-PUM2 (Addgene, MA, USA). SENP1C603S was generated by site-directed mutagenesis using primers 5’-GATGAATGGAAGTGACAGTGGGATGTTTGCCTGC-3’ and 5’-GCAGGCAAACATCCCACTGTCACTTCCATTCATC-3’. Flag-tagged SENP1 (WT and C603S mutant) plasmids were generated by subcloning cDNA inserts into pCMV-N-Flag vector (BiovectorPCMVNF, Addgene, MA, USA). His-SUMO1 and His-SUMO2/3 and V5-PUM2 were purchased from Addgene. HEK293T cells were transiently transfected using Lipofectamine 3000 (L3000150, Invitrogen, CA, USA) following manufacturers’ protocol.

Growth inhibition assays

Refer to Supplemental Materials and Methods for more information.

Western blot assays

Astrocytes isolated from transfected APP/PS1 mice were processed using established methodologies previously outlined in reference (Cheng et al. 2023). Additional protocol details can be found in the Supplemental Materials and Methods.

Cell immunofluorescence staining

For full procedural details, refer to the Supplemental Materials and Methods.

Co-cultured neuron apoptosis analysis

Astrocytes (2 × 105 cells/cm2) were grown overnight in culture media with 5 μM Aβ1–42 in the lower compartments of Transwell inserts (#7369, 0.4 μm pore size; Corning, NY, USA). SH-SY5Y cells were then put in the upper compartments and co-cultured two days before harvest. After washing with phosphate buffered saline and centrifuging twice, resuspend cells in membrane-bound protein-V binding buffer. Stain with Annexin V-PE/7AAD and incubate darkly at room temperature for 15 min. Flow cytometry (FACScan, BD Biosciences, NJ, USA) measures neurons apoptosis.

Microarray analysis of the human mRNA expression profile

The expression profiling of human mRNA was performed utilizing services from Arraystar (Aksomics, Shanghai, China).

Extracellular acidification rate (ECAR) measurement

ECAR was determined using the XF24 Extracellular Flux Analyzer (Seahorse Bioscience, MA, USA) following the XF glycolysis stress test kit (103,020–100, Seahorse Bioscience, MA, USA) protocol. The analyzer was provided by China Medical University’s Experimental Teaching Center. Cell seeding was performed in 24-well plates, followed by incubation at 37 °C under 5% CO2. The XF24 sensor cartridge was hydrated overnight in preparation. For glucose deprivation, cells were incubated in XF Base Medium (pH 7.4, 2 mM glutamine, no glucose) at 37 °C in a CO2-free incubator. ECAR was first measured during a non-glycolytic phase (0–17 min). Following this, glucose was introduced at 10 mM to initiate glycolysis, with measurements taken three times during the early phase (26–45 min post-administration). The introduction of oligomycin aimed to inhibit oxidative phosphorylation, compelling cells to rely solely on glycolysis, thus increasing acidity levels and ECAR (52–72 min). The experiment concluded with addition of 2-deoxyglucose (2-DG) to competitively inhibit glycolysis, monitored over 78–96 min. Following the measurement, cell numbers were double-checked. We normalized all measurements using cell numbers. We calculated glycolysis by subtracting the ECAR attributed to non-glycolytic processes from the ECAR measured following glucose supplementation.

Lactate assays

The lactate concentrations in cell samples were measured using the Lactate Assay Kit (A019-2–1, Jiang Cheng, Nanjing, CN) according to the manufacturer’s instructions. Briefly, homogenized cell and mixed with the enzyme reaction buffer and color developer. Incubate the mixture at 37 °C for 10 min and terminated reaction with the termination solution. The OD value was then assessed using the microplate reader at a wavelength of 530 nm. The concentration of the lactate in the sample was calculated based on the standard.

Glucose uptake assays

The 2-dexoyglucose (2-DG) uptake was detected using a Glucose Uptake Kit (ab136955, Abcam, MA, USA) according to the manufacturer’s instructions. Briefly, the cells were starved in glucose-free dulbecco’s modified eagle medium (DMEM) for 16 h followed by three times PBS wash. The cells were then treatment with high glucose DMEM containing 100 µg/mL 2-DG at 37 °C for 20 min. Cells were then lysed in a lysis buffer and freezed. The sample was then heated at 85 °C for 40 min and cool down on ice for 5 min followed by the treatment with the neutralization buffer. Spin down the samples at 500 rpm for 5 min and incubated the supernatant with reaction buffer at 37 °C for 1 h. The fluorescence was then measured with microplate reader (Elx808, BioTek, VT, USA). Results normalized to the cell numbers, taken control group as 1.

Ni2+-Nitrilotriacetic acid (Ni2+-NTA) agarose bead pull-down assays

The SENP1-mediated deSUMOylation of PUM2 was analyzed through the SUMOylation assay in HEK293T. The His-SUMO1 or His-SUMO2/3 was co-transfected with V5-PUM2, and Flag-SENP1 (WT or C603S mutant) into HEK293T cells with Lipofectamine 3000 (L3000150, Invitrogen, CA, USA) following the manufacturer’s instructions. After 24 h, harvested the cells and lysed in lysis buffer (50 mM NaH2PO4/NaOH, 300 mM NaCl, 10 mM imidazole, 0.5% NP-40, and 1 mM phenyl-methylsulfonyl fluoride). The lysates were then sonicated to release the intracellular proteins. Spined down the lysates and incubated the supernatant with pre-washed Ni–NTA agarose bead (30,230, Qiagen, DUS, GER) for 2 h at 4 °C. Discarded the non-binding proteins by centrifuging and rinsed the beads extensively with washing buffer (50 mM NaH2PO4/NaOH, 300 mM NaCl, 250 mM imidazole) to eliminate the non-specific binding proteins. The protein was then eluted with elution buffer (50 mM NaH2PO4/NaOH, 300 mM NaCl, 500 mM imidazole) at room temperature for 20 min. The deSUMOylation of PUM2 was analysis by Western blot using the anti-V5 (R960-25, Invitrogen, CA, USA) antibodies. The expression of V5-PUM2, Flag-SENP1and control β-actin in the input were analysis by anti-V5, anti-Flag (B3111, Cell Signaling Technology, MA, USA), and anti-β-actin (66,009–1-Ig, Proteintech, Hubei, CN) respectively. The assays (Fig. 2a-d) were performed once.

Fig. 2.

Fig. 2

Open in a new tab

SENP1 deSUMOylates PUM2. (a and b) SENP1 deSUMOylates PUM2 in cells. HEK-293 T cells were transfected with V5-PUM2, His-SUMO1 (a), or His-SUMO2/3 (b) with or without Flag-SENP1 for 48 h and treated with MG132 for 6 h. The cells were subjected to pulldown (PD) using Ni2+-NTA bead under denaturing conditions, followed by western blot. (c and d) SENP1 deSUMOylates endogenous PUM2 in cells. HEK-293 T cells were transfected with His-SUMO1 (c) or His-SUMO2/3 (d) with or without Flag-SENP1 for 48 h and treated with MG132 for 6 h. The cells were subjected to pulldown (PD) using a Ni2+-NTA bead under denaturing conditions, followed by western blot. (e) SENP1 de-SUMOylated endogenous PUM2 in HEK293T cells, confirmed by the CO-IP method. HEK293T cells were transfected with indicated plasmids (shNC, shSENP1), then lysed for the CO-IP method with an anti-SUMO1 antibody, an anti-PUM2 antibody, or normal IgG, followed by western blot assays with anti-PUM2 and anti-SUMO1 antibodies

Co-immunoprecipitation assays (Co-IP)

The RIPA lysis buffer (P0013B, Beyotime, Jiangsu, CN) was used to lysis the HEK293T or Aβ1–42 on ice for 60 min. Centrifuge the cell lysate with 15,000 g for 5 min. Incubate the supernatant with the anti-PUM2 (ab92390, Abcam, MA, USA) (or anti-SUMO1(ab32058, Abcam, MA, USA)) conjugated A/G magnetic beads (B23201, Bimake, Shanghai, CN) overnight at 4 °C. Wash the beads with high salt solution and boiled the samples in SDS loading buffer for 10 min before loading into the SDS-PAGE gel. The sample was then ready for Western blot analysis. The assays (Fig. 2e, 3i) were performed once.

Chromatin immunoprecipitation (ChIP) assays

ChIP assays were performed using a Simple-ChIP Enzymatic Chromatin IP Kit (#9003, Cell Signaling Technology, MA, USA) following the manufacturer’s description. Primers used for ChIP PCR are shown in Table A3. Cross-linked the Aβ1–42-incubated astrocytes cell with 1% formaldehyde and terminated the cross-linking with glycine. The corss-linked cell was then lysed with the lysis buffer and digested with the micrococcal nuclease at 37 °C for 20 min to yield DNA fragments. A total of 20ug digested chromatin was incubated with the anti-NRF2 (#12,721, Cell Signaling Technology, MA, USA) immunoprecipiting antibody or the negative control IgG (ab18413, Abcam, MA, USA) with the protein G agarose beads overnight at 4 °C. Pellet the protein G magnetic beads for the immunoprecipitated sample and then eluted the chromatin from the beads. Reverse-crosslinked the chromatin with NaCl and proteinase K at 65 °C for 2 h. DNA was then extracted and verified by PCR.

Reporter vector construction and Luciferase reporter assays

Amplified the HK1 or GLUT1 promoter regions from human genomic DNA and sub-cloning into the pGL3-Basic-Luciferase vector (E1751, Promega, WI, USA) to obtain pGL3-HK1-Luciferase and pGL3-GLUT1-Luciferase. Meanwhile, the full-length of NRF2 was constructed into the pEX3 (pGCMV/MCS/Neo) vector to generate pEX3-NRF2. Co-transfected the pGL3-HK1-Luciferase or pGL3-GLUT1-Luciferase with pEX3-NRF2 into HEK293T cells using Lipofectamine 3000 (L3000150, Invitrogen, CA, USA). 48 h post-transfection, the dual luciferase assays were performed according to the manufacturer’s instruction (LUC1, Sigma-Aldrich, ST. Louis, USA) with firefly luciferase measurements normalized to the transfection control renilla luciferase activity.

RNA immunoprecipitation (RIP) assays

The RNA-IP assay was performed using the Magna RNA-binding protein immunoprecipitation kit (17–700, Millipore, MA, USA) according to the manufactures’ instructions. Briefly, the harvested cells were resuspended with RIP lysis buffer and incubated with 50ul magnetic beads and 5 μg anti-Ago2 antibody (#2897, Cell Signaling Technology, MA, USA) or the negative control mouse IgG antibodies (ab18413, Abcam, MA, USA) overnight at 4 ℃. The immunoprecipitated samples were then washed extensively with wash buffer and eluted with RIP buffer containing proteinase K. Purified the RNA with TRIzol reagent according to the manufacturer’s instructions and reversed transcribe the RNA to synthesis the cDNA with PrimeScript™ RT reagent Kit with gDNA Eraser (Perfect Real Time) kit (RR047A, Takara, Beijing, CN). Samples were then analysis by qRT-PCR.

RNA pull-down assays

RNA pull-down assay was conducted using Pierce Magnetic RNA–Protein Pull-Down Kit (20,164, Thermo Fisher Scientific, USA) according to the manufacturer’s protocols. Briefly, the biotin-labeled RNA probe of NRF2 or IgG (negative control) were first synthesized using the Biotin RNA Labelling Mix (11,685,597,910, Roche, Shanghai, CN). The astrocytes were lysed and incubated with the biotin-labeled NRF2 or IgG probe and streptavidin magnetic beads overnight at 4 °C. The co-precipated sample was then washed with ice-cold wash buffer and boiled in SDS loading buffer. The samples were then identified by Western blot and β-actin served as a loading control.

Immuno-histofluorescence staining

Mice were intraperitoneally anesthetized with 1000 mg/kg 2,2,2-tribromoethanol (M06376-UYX, Balb, Beijing, CN) and perfused from the left ventricle with 30 ml PBS followed by 30 ml 4% paraformaldehyde. The cerebral cortex tissue was then collected and fixed with 4% paraformaldehyde overnight for paraffin slices. The 5-μm slices were subsequently subjected to a heat treatment in EDTA for 15 min antigen retrieval. The slices were permeabilized with 0.3% Triton X-100 and blocked with 8% donkey serum for non-specific binding. After blocking, the samples were incubated with the primary antibodies anti-GFAP (60,190–1-Ig, Proteintech, Hubei, CN), anti-NeuN (ab279297, Abcam, MA, USA), anti-HK1 (ab150423, Abcam, MA, USA), and anti-GLUT1(ab115730, Abcam, MA, USA) overnight at 4 °C. Washed the samples with PBS and incubated with the Alexa Fluor 488-conjugated goat anti-mouse IgG (A0428, Beyotime, Jiangsu, CN) for GFAP expression and Alexa Fluor 555-conjugated donkey anti-rabbit IgG (A0453, Beyotime, Jiangsu, CN) for HK1 or GLUT1 expression. Nuclei were stained with DAPI (C1002, Beyotime, Jiangsu, CN) at room temperature for 20 min. The slides were sealed and viewed with a confocal microscopy (Nikon, TYO, Japan) in 5 random views. The pixel intensity of each image was calculated using ImageJ.

Nascent RNA capture and mRNA stability assays

Refer to Supplemental Materials and Methods.

H&E staining and TUNEL assays

Assay protocols can be found in the Supplemental Materials and Methods.

Statistical analysis

The measurement data were expressed as mean ± SD. Student’s t-test (two-tailed) or one-way ANOVA was used to analysis the data that was normally distributed and homogeneous variance. Welch’s t-test was used to evaluate data appeared to be normally distributed but heterogeneous variance. Mann–Whitney U test was used to evaluate data that was not normally distributed. *P < 0.05 was considered as statistically significant. SPSS26.0 was used for data analysis and GraphPad prism 6.0 was used for graphs.

Results

SENP1 expression was significantly elevated in astrocytes incubated with Aβ1–42, potentially contributing to suppressed glycolysis and enhanced neuronal apoptosis in the Aβ1–42 microenvironment.

We utilized the well-established Aβ-incubated cell model to explore AD by co-culturing HAs and SH-SY5Y cells with Aβ1–42 (Efremova et al. 2015; Mizuno et al. 2005). We observed a positive correlation between the degree of inhibition and both the concentration and duration of Aβ1–42 exposure (Figure S1a). This level of inhibition effectively mimics the neurotoxic effects seen in AD while preserving sufficient cell viability to study response and survival mechanisms under neurotoxic conditions. After treating astrocytes and SH-SY5Y cells with 5 μM Aβ1–42 for 48 h, we measured the ECAR in both normal and pre-incubated cells using the Seahorse XF Analyzer. This measurement indicated variable glycolytic flux, from which we calculated glycolysis (detailed in “Materials and Methods”). Astrocytes treated with Aβ1–42 demonstrated diminished glycolytic capability (Figure S1b), along with decreased lactate production and glucose intake (Figure S1c, d). Known glycolysis regulators, GLUT1 and HK1, were also explored to understand their roles in the metabolic reprogramming in Aβ1–42-exposed astrocytes, revealing reduced expression of both proteins (Figure S1e). These findings underscore the impact of Aβ1–42 exposure on astrocyte glycolysis.

Further investigation into SUMOylation processes involved measuring mRNA levels of enzymes like SENP1 in astrocytes treated with Aβ1–42. RNA microarray analysis consistently showed elevated SENP1 expression (Figs. 1a, b), a finding corroborated by Western blot assays indicating increased SENP1 protein levels (Fig. 1c). The immunofluorescence staining highlighted significantly greater SENP1 fluorescence intensity in the nucleus of Aβ1–42-treated astrocytes compared to controls (Fig. 1d). To investigate SENP1’s role in astrocytes, we generated astrocytes with stable SENP1 knockdown and re-expression of SENP1 in knockdown cells (Figure S2a-c). SENP1 knockdown enhanced glycolytic capacity, lactate production, and glucose uptake, accompanied by increased GLUT1 and HK1 protein levels (Figs. 1e-h, S3a, b). Re-expression of SENP1 reversed these changes, restoring glycolysis suppression and reduced protein expression (Figs. 1e-h, S3a, b). Moreover, SENP1 knockdown in Aβ1–42-treated astrocytes significantly reduced apoptosis in co-cultured neurons, while SENP1 re-expression reversed this protective effect (Figs. 1i, S3c). These results highlight SENP1’s pivotal role in modulating glycolysis and apoptosis within the AD microenvironment, suggesting that increased SENP1 expression in Aβ1–42-pre-incubated astrocytes could hinder glycolysis and accelerate neuronal apoptosis.

Fig. 1.

Fig. 1

Open in a new tab

Increased SENP1 expression in Aβ1–42-incubated astrocytes inhibits glycolysis and induces apoptosis in co-cultured neurons in the Aβ1–42-microenvironment. (a) RNA microarrays of the enzymes related to SUMOylation was performed in Aβ1–42-incubated astrocytes. Green indicates high relative expression, and brown indicates low relative expression. (b) Relative expression levels of SENP1, PIAS3, and RANBP2 in Aβ1–42-incubated astrocytes by qRT-PCR. Data are presented as mean ± SD (n = 5), **P < 0.01 versus HAs group. (c) Relative expression level of SENP1 protein in Aβ1–42-incubated astrocytes by western blot. Data are presented as mean ± SD (n = 3), **P < 0.01 versus HAs group. (d) Immunofluorescence staining analysis of the location of SENP1 (green) in both the nucleus and the cytoplasm of Aβ1–42-incubated astrocytes. The scale bar represents 25 μm. (e) Effects of SENP1 on ECAR in Aβ1–42-incubated astrocytes. ECAR is shown in milli pH per minute (milli pH/min), representing the rate of acidification of the extracellular medium. (f-g) Effects of SENP1 on lactate production (f) and glucose uptake (g) in Aβ1–42-incubated astrocytes. Data are presented as mean ± SD (n = 3, each). **P < 0.01 versus the shNC group. ##P < 0.01 versus the shSENP1 + SENP1-NC group. (h) Effects of SENP1 on protein levels of HK1 and GLUT1 in Aβ1–42-incubated astrocytes by western blot. (i) Effects of SENP1 knockdown or re-expression in astrocytes on apoptosis of co-cultured neurons in the Aβ1–42-microenvironment

SENP1 deSUMOylated PUM2

We investigated whether SENP1 could deSUMOylate PUM2, leveraging earlier findings that indicated PUM2’s susceptibility to SUMOylation in glioma cells(Wang et al. 2020). In our experiments, HEK293T cells were transfected with V5-tagged PUM2, His-tagged SUMO1 or His-SUMO2/3, and Flag-tagged wild-type SENP1 or its inactive mutant SENP1C603S. MG132 is well known as a potent proteasome inhibitor, which plays a crucial role in degrading ubiquitinated proteins. By inhibiting the proteasome, MG132 disrupts protein degradation processes, resulting in the accumulation of regulatory proteins. Following transfection, cells were treated with MG132, and SUMO-conjugates were isolated using Ni2+-NTA purification for Western blot analysis. Our results demonstrated that while wild-type SENP1 reduced the SUMOylation of PUM2 modified by SUMO1, the SENP1C603S mutant did not (Fig. 2a, c). Conversely, SENP1 showed minimal impact on the levels of SUMO2/3-conjugated PUM2 (Fig. 2b, d). Co-immunoprecipitation confirmed that SENP1 decreased endogenous PUM2 SUMOylation. Knockdown of SENP1 via shSENP1 resulted in increased SUMOylation of PUM2 (Fig. 2e), suggesting SENP1’s role in modulating PUM2’s SUMO1 modification.

SENP1 inhibited glycolysis in Aβ1–42 incubated astrocytes via reducing PUM2 SUMOylation

Our investigations revealed a significant increase in PUM2 expression in astrocytes treated with Aβ1–42, consistent with previous findings (Cheng et al. 2023) (Fig. 3a, S4a). Immunofluorescence staining demonstrated elevated nuclear localization of PUM2 in Aβ1–42-incubated astrocytes compared to controls (Fig. 3b). Astrocytes with PUM2 knockdown exhibited enhanced glycolysis (Fig. 3c, S4b), lactate production (Fig. 3d), glucose uptake (Fig. 3e), and increased expression of HK1 and GLUT1 (Fig. 3f, S4c). Additionally, flow cytometry revealed a significant reduction in apoptosis among co-cultured neurons (Fig. 3g, S4d). Re-expression of PUM2 in knockdown cells reversed these metabolic and apoptotic effects, confirming its critical role (Fig. 3c-g, S4b, d).

Fig. 3.

Fig. 3

Open in a new tab

Glycolysis was inhibited by SENP1 through the reduction of PUM2 deSUMOylation in the Aβ1–42-incubated astrocytes. (a) Effects of Aβ1–42 on SENP1 protein level by western blot. (b) Immunofluorescence staining analysis of the location of PUM2 (green) in both the nucleus and the cytoplasm of Aβ1–42-incubated astrocytes. The scale bar represents 25 μm. (c) Effects of PUM2 on ECAR in Aβ1–42-incubated astrocytes. ECAR is shown in milli pH per minute (milli pH/min), representing the rate of acidification of the extracellular medium. (d-e) Effects of PUM2 on lactate production (d) and glucose uptake (e) in Aβ1–42-incubated astrocytes. Data are presented as mean ± SD (n = 3, each). **P < 0.01 versus the shNC group. ##P < 0.01 versus the shPUM2 + PUM2-NC group. (f) Effects of PUM2 on protein levels of HK1 and GLUT1 in Aβ1–42-incubated astrocytes by western blot. (g) Effects of PUM2 knockdown or re-expression in astrocytes on apoptosis of co-cultured neurons in the Aβ1–42-microenvironment. (h) Immunofluorescence staining analysis of the subcellular location of SUMO1 (red) and SENP1 (green) in Aβ1–42-incubated astrocytes. The scale bar represents 25 μm. (i) SENP1 deSUMOylates PUM2 in Aβ1–42-incubated astrocytes, confirmed by the CO-IP method. Aβ1–42-incubated astrocytes were transfected with indicated plasmids (shNC, shSENP1), then lysed for the CO-IP method with an anti-SUMO1 antibody, an anti-PUM2 antibody or normal IgG, followed by western blot assays with anti-PUM2 and anti-SUMO1 antibodies

Our further studies focused on the interaction between SENP1 and PUM2 in Aβ1–42-treated astrocytes. Immunofluorescence staining showed nuclear co-localization of SUMO1 and PUM2 (Fig. 3h), while Co-IP experiments revealed increased SUMO1-conjugated PUM2 levels in SENP1-knockdown astrocytes (Fig. 3i). The protein expression of PUM2 was diminished in the shSENP1 group but restored upon SENP1 re-expression (Figure S4e). These findings suggest that PUM2 undergoes SUMO1-mediated degradation, which is counteracted by SENP1 through deSUMOylation.

Using the GPS-SUMO prediction tool (available at http://sumosp.biocuckoo.org), we identified potential SUMOylation sites in PUM2 and generated corresponding mutants (K91/985R, 2KR) (Figure S4f, g). Proteasome involvement in PUM2 degradation was confirmed as MG132 treatment significantly reduced PUM2 degradation in SENP1-knockdown cells (Figure S4h). Overexpression of PUM2 reversed the glycolytic and metabolic enhancements observed in SENP1-knockdown astrocytes, including reduced lactate production, glucose uptake, and lower HK1 and GLUT1 expression levels in Aβ1–42-treated astrocytes (Figures S5a-d). These results demonstrate that SENP1 enhances PUM2 expression through deSUMOylation, thus suppressing aerobic glycolysis in Aβ1–42 incubated astrocytes.

The expression of NRF2 was reduced in Aβ1–42-incubated astrocytes, and NRF2 overexpression might enhance glycolysis in the astrocytes

Utilizing the RBP map database and RNA microarray analyses, we identified significant interactions between several transcription factors and PUM2, with NRF2 showing the most notable upregulation in PUM2 knockdown astrocytes pre-incubated with Aβ1–42 (Figs. 4a, b). Notably, the protein expression levels of NRF2 were decreased under Aβ1–42 treatment (Fig. 4c). Overexpression of NRF2 in these cells led to increases in glycolysis (Fig. 4d, S6a), lactate production (Fig. 4e), glucose uptake (Fig. 4f), and the expression levels of HK1 and GLUT1 (Figures S6b, 4 g, S6c). Conversely, NRF2 knockdown produced opposite effects, enhancing apoptosis in co-cultured neurons (Figs. 4d-g, S6a-c, 4 h, S6d). Our results above suggest that NRF2 might enhance the glycolysis of astrocytes and reduce apoptosis in co-cultured neurons in Aβ1–42 microenvironment.

Fig. 4.

Fig. 4

Open in a new tab

NRF2 expression was decreased in Aβ1–42-incubated astrocytes, and astrocyte glycolysis may be enhanced by NRF2 overexpression. (a) RNA microarrays were performed in Aβ1–42-incubated HAs treated with shPUM2. Green indicates high relative expression, and yellow indicates low relative expression. (b) Relative expression of NRF2, KLF4 and VSX1 in Aβ1–42-incubated HAs by qRT-PCR. Data are presented as mean ± SD (n = 5), *P < 0.05 versus shNC group, **P < 0.01 versus shNC group. (c) Relative expression levels of NRF2 protein in Aβ1–42-incubated astrocytes by western blot. Data are presented as mean ± SD (n = 3), **P < 0.01 versus HAs group. (d) Effects of NRF2 on ECAR in Aβ1–42-incubated astrocytes. ECAR is shown in milli pH per minute (milli pH/min), representing the rate of acidification of the extracellular medium. (e–f) Effects of NRF2 on lactate production (e) and glucose uptake (f) in Aβ1–42-incubated astrocytes. Data are presented as mean ± SD (n = 3, each). **P < 0.01 versus the NRF2-NC group. ##P < 0.01 versus the shNC group. (g) Effects of NRF2 on protein levels of HK1 and GLUT1 in Aβ1–42-incubated astrocytes by western blot. (h) Effects of NRF2 over-expression or knockdown or in astrocytes on apoptosis of co-cultured neurons in the Aβ1–42-microenvironment

PUM2 might decrease the glycolysis of astrocytes by reducing NRF2 mRNA stability in the Aβ1–42 microenvironment

Further investigations focused on how PUM2 might influence glycolysis through its interaction with NRF2. PUM2 knockdown resulted in increased NRF2 expression levels, an effect that was reversed upon re-expression of PUM2 (Figures S7a, 5a, S7b). PUM2 recognizes and binds to a particular sequence known as the PUM-binding element (PBE), which is characterized by the sequence 5′-UGUAHAUA-3′. In this sequence, ‘H’ can represent adenine (A), cytosine (C), or uracil (U). PBE motifs have been identified within its 3’ UTR (Figure S10). RIP demonstrated that PUM2 interacts with NRF2 mRNA within astrocytes incubated with Aβ1–42, specifically binding to the 3’UTR of NRF2 mRNA (Fig. 5b, c). PUM2 knockdown extended the half-life of NRF2 mRNA, implicating a post-transcriptional regulatory mechanism by PUM2 on NRF2 stability (Fig. 5d). Nascent RNA capture experiments showed no changes in NRF2 mRNA synthesis following PUM2 knockdown, supporting the notion of post-transcriptional regulation (Figure S7c). To further elucidate the role of PUM2 in controlling glycolysis via NRF2 expression, we transfected shNRF2 plasmid into shPUM2 astrocytes pre-incubated with Aβ1–42. NRF2 knockdown mitigated the glycolytic enhancement seen with PUM2 knockdown, affecting lactate production, glucose uptake, and the expression levels of GLUT1 and HK1 (Fig. 5e-h, S7d, e). Additionally, while SENP1 knockdown initially upregulated NRF2 expression, this increase was reversed with PUM2 overexpression (Fig. 5i, S7f). These findings collectively suggest that PUM2 diminishes glycolysis in astrocytes exposed to Aβ1–42 by destabilizing NRF2 mRNA, offering a new understanding of how NRF2 and PUM2 interplay affects metabolic processes in neurodegenerative environments.

Fig. 5.

Fig. 5

Open in a new tab

PUM2 regulated astrocytes glycolysis by decreasing the half-life of NRF2 mRNA in Aβ1–42 microenvironment. (a) Effects of PUM2 on NRF2 protein level in Aβ1–42-incubated astrocytes by western blot. (b) RNA immunoprecipitation confirmed the bonding between PUM2 and NRF2 mRNA in Aβ1–42 microenvironment. Relative enrichment was measured by qRT-PCR. Data are presented as mean ± SD (n = 3, each). **P < 0.01 versus Anti-IgG group. (c) Western blot of the proteins from NRF2 CDS, NRF2 3’UTR and antisense NRF2 pull-down assays in Aβ1–42 microenvironment. (d) The graph shows NRF2 mRNA levels in astrocytes pre-incubated with Aβ1–42 at different times treated with ActD in the Control, shNC and shPUM2 groups by qRT-PCR. Data are presented as mean ± SD (n = 3, each). (e) Effects of PUM2 and NRF2 knockdown on ECAR in Aβ1–42-incubated astrocytes. ECAR is shown in milli pH per minute (milli pH/min), representing the rate of acidification of the extracellular medium. (f-g) Effects of PUM2 and NRF2 knockdown on lactate production (f) and glucose uptake (g) in Aβ1–42-incubated astrocytes. Data are presented as mean ± SD (n = 3, each). **P < 0.01 versus the shPUM2-NC + shNRF2-NC group. ##P < 0.01 versus the shPUM2 + shNRF2-NC group. (h) Effects of PUM2 and NRF2 knockdown on protein levels of HK1 and GLUT1 in Aβ1–42-incubated astrocytes by western blot. (i) Effects of SENP1 knockdown and PUM2 overexpression on NRF2 protein level in Aβ1–42-incubated astrocytes by western blot

NRF2 bound to the promoters of GLUT1 and HK1 and promoted transcription

Importantly, we found NRF2 overexpression increased the mRNA expression of HK1 and GLUT1, and NRF2 knockdown decreased the expression (Figures S6b, 4 g, S6c). Using bioinformatics tools JASPAR (https://jaspar.elixir.no/), we identified potential NRF2 binding sites within the promoters of HK1 and GLUT1. To validate NRF2’s regulatory role on these genes, we conducted dual-luciferase reporter and ChIP assays. Co-transfection with pEX3-NRF2 significantly elevated the promoter activities of both HK1 and GLUT1 (Fig. 6a, b). Further, deletion of presumed NRF2 binding sites at − 1039 in HK1 and − 1016 in GLUT1 significantly reduced their promoter activities, confirming NRF2’s binding and transcriptional enhancement (Fig. 6a-d). ChIP assays substantiated NRF2’s binding to these promoters, solidifying its role in transcriptional regulation within the Aβ1–42 microenvironment.

Fig. 6.

Fig. 6

Open in a new tab

NRF2 increased the promoter activity of HK1 and GLUT1 in astrocytes pre-incubated with Aβ1−42. (a-b) Schematic depiction of the different reporter plasmids and relative luciferase activity in Aβ1–42 microenvironment. The relative luciferase activity was conducted after Aβ1–42-incubated astrocytes co-transfected with HK1 (a) and GLUT1 (b) promoter (−1,500 to 0 bp) (or promoter-deleted putative NRF2 binding sites) with pEX3-NRF2 or pEX3 empty vector. Data represent mean ± SD (n = 3, each). **P < 0.01. (c-d) NRF2 interacted with the promoters of HK1 (c) and GLUT1 (d) in Aβ1–42-incubated astrocytes. Transcription start site (TSS) was designated as + 1. Putative NRF2 binding sites are illustrated. Immunoprecipitated DNA was amplified by PCR. Normal rabbit IgG was used as a negative control

Senp1 knockdown, Pum2 knockdown, or Nrf2 overexpression enhanced aerobic glycolysis in astrocytes and reduced cognitive impairment in APP/PS1 transgenic mice

To explore the in vivo effects of SENP1, PUM2, and NRF2 on astrocyte aerobic glycolysis and cognitive function, shRNA constructs targeting these genes were delivered via intracerebroventricular injections using AAV9 vectors into APP/PS1 transgenic mice. This specific targeting ensured that the knockdown or overexpression effects were astrocyte-restricted (Figure S8). The MWM test assessed the spatial learning in these mice, showing that SENP1 knockdown, PUM2 knockdown, or NRF2 overexpression significantly improved the cognitive abilities compared to the control group. These genetic modifications notably decreased escape latency, increased platform crosses, and reduced the time spent navigating away from the platform location (Fig. 7a-c).

Fig. 7.

Fig. 7

Open in a new tab

Senp1 knockdown, Pum2 knockdown, or Nrf2 overexpression increased astrocyte aerobic glycolysis and decreased cognitive impairment in APP/PS1 transgenic mice. (a-c) Cognition was evaluated in the mice by the Morris water maze (MWM), including the representative swimming track (a), latency to the platform during the spatial training trials of MWM (b), and frequency across the platform and latency to cross the platform of MWM in the probe trials (c). (d) Evaluation of protein levels of HK1 and GLUT1 by immuno-histofluorescence staining in vivo, and immuno-histofluorescence staining assays of frozen slices (10 μm thick) of the frontal association cortex were stained with HK1 and GLUT1 (red), GFAP (green), and DAPI (blue) (scale bar = 100 and 25 μm). (e) Evaluation of protein levels of HK1 and GLUT1 in astrocytes in vivo by western blot. (f-g) Evaluation of frontal association cortex neuron apoptosis in vivo, including H&E-stained (scale bar = 100 and 25 μm) (f) and terminal deoxynucleotidyl transferase (TdT)-mediated dUTP-biotin nick end labeling (TUNEL) assay (scale bar = 50 μm) (g)

Immuno-histofluorescence staining of the frontal association cortex and subsequent Western blot analysis confirmed that SENP1 knockdown, PUM2 knockdown, or NRF2 overexpression led to increased expression of HK1 and GLUT1, suggesting enhanced glycolytic support from astrocytes to neurons (Fig. 7d, e, and S9a, b). This was further evidenced by reduced neuronal apoptosis in the frontal association cortex of the treated groups (Figs. 7f-g, and S9c). These results underline that modulation of SENP1, PUM2, or NRF2 can significantly influence astrocyte metabolism and neuro-supportive functions, ultimately aiding in alleviating cognitive impairments in AD models.

Discussion

This study uniquely elucidates how SENP1, PUM2, and NRF2 expression dynamics in Aβ1–42-incubated astrocytes modulate cellular mechanisms underlying AD. To summarize, our study reveals that SENP1 and PUM2 are significantly over-expressed in Aβ1–42-incubated astrocytes, whereas NRF2 is markedly downregulated. Knocking down SENP1 or PUM2 and overexpressing NRF2 markedly improved aerobic glycolysis and decreased neuronal apoptosis in the Aβ1–42 microenvironment. Our findings indicate that SENP1 deSUMOylates PUM2, thereby preventing its protein degradation. The subsequent increase in PUM2 diminishes NRF2 expression by destabilizing its mRNA. This reduction in NRF2 ultimately suppresses glycolysis by reducing the transcription of HK1 and GLUT1 (Fig. 8).

Fig. 8.

Fig. 8

Open in a new tab

The schematic illustration of interactions between SENP1, PUM2, and NRF2 in Aβ1–42-incubated astrocytes

Astrocytes play a crucial role in maintaining neuronal health by providing metabolic support, regulating glucose metabolism, storing glycogen, and facilitating the astrocyte-neuron lactate shuttle, essential for neuronal function and brain energy metabolism (Theparambil et al. 2024). In AD, dysfunctional astrocytes decrease their metabolic support, exacerbating neuronal damage (Linnerbauer et al. 2020). Additionally, metabolic abnormalities in astrocytes can activate microglia through the secretion of inflammatory cytokines like TNF-α and IL-1β, leading to increased neuroinflammation (Johnson et al. 2020). These abnormalities also disrupt the production of key metabolites, such as glutamine, affecting microglial function and amyloid-beta (Aβ) clearance, ultimately impairing the balance between neuroprotection and neurotoxicity (Lepiarz-Raba et al. 2023; Sadeghdoust et al. 2024). The interactions among these factors significantly contribute to AD progression. Therefore, targeting metabolic abnormalities in astrocytes may provide new therapeutic strategies for treating AD.

The accumulation of misfolded proteins, a hallmark of AD, leads to brain damage and cognitive decline (Kim et al. 2024a). Protein SUMOylation, a reversible post-translational modification, regulates key cellular processes such as DNA repair, RNA processing, and metabolism (Vertegaal 2022). Critically, SUMOylation facilitates the degradation of target proteins through the proteasome pathway, a process that is reversible, allowing proteins to switch between SUMOylated and deSUMOylated states (Li et al. 2024; Yang et al. 2024). Disruption in SUMOylation processes has been implicated in the cognitive impairments observed in AD, as demonstrated by electrophysiological studies on hippocampal neurons (Lee et al. 2014). It is established that SENP1 modulates protein function by deSUMOylation, influencing various cellular outcomes (Qin et al. 2023). For example, SENP1 inhibits SUMOylation on APC/C residues, reducing APC/C degradation and thus delaying mitotic exit in proliferative cells(Liu et al. 2023). SENP1 also mitigates c-Myc degradation through the proteasome system in H1299 cells, highlighting its broad regulatory impacts (Sun et al. 2018). In this study, we extend the understanding of SENP1’s role by demonstrating its upregulation and predominant localization in the nuclei of Aβ1–42-treated astrocytes, which directly contributes to the inhibition of aerobic glycolysis. Immunofluorescence assays reveal the co-localization of SENP1 and PUM2 proteins, suggesting potential functional interactions in this pathological state. Further, we identified that SENP1 specifically removes SUMO-1 from PUM2 based on pull-down and Co-IP assays. The proteasome inhibitor experiments confirm that SENP1 prevents PUM2 degradation via deSUMOylation. This finding demonstrates that SENP1 enhances the expression of PUM2 by deSUMOylating it, which subsequently inhibits its degradation via the proteasome pathway in astrocytes incubated with Aβ1–42.

RNA-binding proteins (RBPs) are key regulators of post-transcriptional gene expression and are increasingly recognized as contributors to neurodegenerative diseases like AD (Antal et al. 2023; Wolozin and Ivanov 2019). Our previous findings suggest that RBPs TRA2A and KHDRBS2 disrupt the blood–brain barrier (BBB) by destabilizing specific mRNAs in the Aβ1–42 microenvironment (Liu et al. 2021, 2020). PUM2 acts as a post-transcriptional repressor by binding to mRNAs that control vital cellular functions such as mitosis and mitochondrial activity, thereby promoting cellular aging (Kopp et al. 2019). It also interacts with mRNAs containing pumilio response elements (PBE), inhibiting their translation and transport, which paradoxically facilitates axon regeneration in neurons (Martinez et al. 2019). Furthermore, PUM2’s regulation of specific mRNAs decreases the population of neural stem cells in the gyrus, increasing perinatal apoptosis and adversely affecting learning and memory capabilities (Zhang et al. 2017). Our results identified a significant increase in PUM2 levels within astrocytes incubated with Aβ1–42. This elevation was associated with a decrease in aerobic glycolysis and an increase in apoptosis in co-cultured neurons. These findings indicate that PUM2 may contribute to the metabolic dysregulation observed in AD pathology.

Utilizing the RBP map database and RNA microarray analyses, we identified significant interactions between several transcription factors and PUM2. Notably, NRF2 showed the most substantial upregulation in PUM2 knockdown astrocytes pre-incubated with Aβ1–42. RIP and RNA pull-down assays revealed that PUM2 directly interacts with the 3’UTR of NRF2 mRNA, reducing its stability and subsequently lowering its protein levels. Extensive research has confirmed that a reduction in NRF2 levels is observed in AD and that it has protective effects against various age-related illnesses(Bai et al. 2023; Lane et al. 2021).

NRF2 enhances mitochondrial activity and lowers reactive oxygen species (ROS) levels, which helps alleviate AD-like symptoms (She et al. 2024). It also lowers levels of amyloid precursor protein (APP) and Aβ, which are crucial in the development of cognitive impairment in aging mice(Zhang et al. 2021). Consistently, our findings demonstrate a reduction in NRF2 levels in Aβ1–42-incubated astrocytes. Moreover, our findings reveal that the suppression of NRF2 expression, facilitated by the increased activity of PUM2, leads to diminished aerobic glycolysis and heightened apoptosis in neurons co-cultured with Aβ1–42-incubated astrocytes. Conversely, NRF2 overexpression mitigates these effects by enhancing glycolysis and decreasing neuronal apoptosis, highlighting its pivotal role in the AD microenvironment.

NRF2, belonging to the Cap-n-Collar (CNC) family and characterized by a basic-region leucine zipper structure, has been shown to bind directly to the promoters of specific genes to enhance their transcriptional activity (Uruno and Yamamoto 2023). This transcription factor not only promotes the transcription of cytoprotective genes in response to cellular stress but also activates the NADPH-generating enzyme ME1, critical for energy metabolism and the restoration of macrophage function in chronic diseases like COPD (Bollong et al. 2018; Ryan et al. 2023). Furthermore, NRF2 has been found to promote glycolysis by increasing the transcription of HK2 in contexts of hepatocellular carcinoma tumorigenesis (Zhao et al. 2022). Our findings demonstrate that NRF2 directly binds to the promoter regions of HK1 and GLUT1, as confirmed by luciferase assays and ChIP experiments. Overexpression of NRF2 significantly increases the mRNA and protein levels of these enzymes, suggesting that NRF2 could potentially slow Alzheimer's disease progression by enhancing the transcription of HK1 and GLUT1 in astrocytes.

Interventions like SENP1 or PUM2 knockdown and NRF2 overexpression have been shown to improve cognitive functions in APP/PS1 mice, a well-established AD model. These interventions notably enhance HK1 and GLUT1 expression in the mice’s brain astrocytes and decrease apoptosis in the frontal association cortex, suggesting potential mitigation of AD symptoms and progression.

This study suggests that SENP1 and NRF2 may serve as potential targets for AD treatment. Gene therapy represents the most direct and promising approach; however, its delivery to the brain faces challenges due to the BBB and the complexity of neuronal networks. Notably, we have identified small molecules such as JCP-666 and VEA-260 for SENP1 inhibition (Wei et al. 2022), as well as Dimethyl fumarate, a small-molecule activator of NRF2, currently in clinical trials for various diseases and approved for treating relapsing–remitting multiple sclerosis and psoriasis—which present promising opportunities for targeting these proteins (Lyu et al. 2024). Further research is needed to evaluate the safety and efficacy of these therapeutic approaches. The limitations of this study include the inability of the APP/PS1 transgenic mouse model to fully replicate human AD in terms of metabolic processes, as well as the absence of human brain tissue samples to demonstrate the relevance of SENP1, PUM2, and NRF2 expression in human AD.

In conclusion, our study reveals novel dynamics in the expression of SENP1, PUM2, and NRF2 within astrocytes preconditioned with Aβ1–42. SENP1 inhibits PUM2 SUMOylation, leading to increased PUM2, which in turn promotes NRF2 mRNA degradation. This decrease in NRF2 diminishes its transcriptional regulation of critical glycolytic genes, notably HK1 and GLUT1, thereby reducing aerobic glycolysis in Aβ1–42-incubated astrocytes and prompting apoptosis in co-cultured neurons. Importantly, in vivo interventions involving Senp1 or Pum2 knockdown, or NRF2 overexpression, significantly enhanced the expression of HK1 and GLUT1 in astrocytes and reduced cell apoptosis in the frontal association cortex of APP/PS1 mice. These results not only demonstrate the crucial role of these molecular interactions in influencing AD pathology but also opens up new avenues for exploring therapeutic strategies aimed at modulating these proteins, potentially offering new ways to combat the effects of AD. Further exploration into how these molecular mechanisms can be effectively targeted in therapeutic contexts is essential for advancing the treatment of AD.

Supplementary Information

Below is the link to the electronic supplementary material.

ESM1 (2.5MB, pdf)

(PDF 2.52 MB)

ESM2 (3.2MB, docx)

(DOCX 3.17 MB)

ESM3 (20.6KB, docx)

(DOCX 20.5 KB)

ESM4 (28.5KB, docx)

(DOCX 28.4 KB)

Acknowledgements

We thank all individuals who take part in this research.

Author contribution

Lianbo G. and Qianshuo L. conceived and designed the study. Qianshuo L., Libin G., Meixi J., Zhengze W., Jihong M. and Jing L. performed, analyzed, and interpreted the studies. Qianshuo L., Miexi J. and Libin G. wrote the article. Lianbo G., Libin G., and Qianshuo L. contributed reagents/materials/analysis tools and revised the article. All authors discussed the data and commented on the manuscript.

Funding

This work is supported by the National Natural Science Foundation Cultivation Project of the Fourth Affiliated Hospital of China Medical University (Grant No. f00049) and National Natural Science Foundation of China (Grant No. 82401820).

Data availability

The data used that support the fndings of this study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval

The animal study was reviewed and approved by the Institution Animal Ethics Committee of China Medical University.

Footnotes

Publisher's note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Qianshuo Liu and Meixi Jiang contributed equally to this work.

Contributor Information

Jiacai Lin, Email: sovereignljc@126.com.

Libin Guo, Email: 13840116491@163.com.

Lianbo Gao, Email: lbgao@cmu.edu.cn.

References

  1. Antal CE, Oh TG, Aigner S, Luo EC, Yee BA, Campos T, Tiriac H, Rothamel KL, Cheng Z, Jiao H, Wang A, Hah N, Lenkiewicz E, Lumibao JC, Truitt ML, Estepa G, Banayo E, Bashi S, Esparza E, Munoz RM, Diedrich JK, Sodir NM, Mueller JR, Fraser CR, Borazanci E, Propper D, Von Hoff DD, Liddle C, Yu RT, Atkins AR, Han H, Lowy AM, Barrett MT, Engle DD, Evan GI, Yeo GW, Downes M, Evans RM. A super-enhancer-regulated RNA-binding protein cascade drives pancreatic cancer. Nat Commun. 2023;14:5195. 10.1038/s41467-023-40798-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bai X, Bian Z, Zhang M. Targeting the Nrf2 signaling pathway using phytochemical ingredients: A novel therapeutic road map to combat neurodegenerative diseases. Phytomedicine. 2023;109:154582. 10.1016/j.phymed.2022.154582. [DOI] [PubMed] [Google Scholar]
  3. Bollong MJ, Lee G, Coukos JS, Yun H, Zambaldo C, Chang JW, Chin EN, Ahmad I, Chatterjee AK, Lairson LL, Schultz PG, Moellering RE. A metabolite-derived protein modification integrates glycolysis with KEAP1-NRF2 signalling. Nature. 2018;562:600–4. 10.1038/s41586-018-0622-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Calabrese EJ, Kozumbo WJ. The hormetic dose-response mechanism: Nrf2 activation. Pharmacol Res. 2021;167:105526. 10.1016/j.phrs.2021.105526. [DOI] [PubMed] [Google Scholar]
  5. Chao CC, Gutierrez-Vazquez C, Rothhammer V, Mayo L, Wheeler MA, Tjon EC, Zandee SEJ, Blain M, de Lima KA, Takenaka MC, Avila-Pacheco J, Hewson P, Liu L, Sanmarco LM, Borucki DM, Lipof GZ, Trauger SA, Clish CB, Antel JP, Prat A, Quintana FJ. Metabolic Control of Astrocyte Pathogenic Activity via cPLA2-MAVS. Cell. 2019;179(1483–98):e22. 10.1016/j.cell.2019.11.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Chen X, Zhang Y, Wang Q, Qin Y, Yang X, Xing Z, Shen Y, Wu H, Qi Y. The function of SUMOylation and its crucial roles in the development of neurological diseases. FASEB J. 2021;35:e21510. 10.1096/fj.202002702R. [DOI] [PubMed] [Google Scholar]
  7. Chen QT, Zhang ZY, Huang QL, Chen HZ, Hong WB, Lin T, Zhao WX, Wang XM, Ju CY, Wu LZ, Huang YY, Hou PP, Wang WJ, Zhou D, Deng X, Wu Q. HK1 from hepatic stellate cell-derived extracellular vesicles promotes progression of hepatocellular carcinoma. Nat Metab. 2022;4:1306–21. 10.1038/s42255-022-00642-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cheng YT, Woo J, Luna-Figueroa E, Maleki E, Harmanci AS, Deneen B. Social deprivation induces astrocytic TRPA1-GABA suppression of hippocampal circuits. Neuron. 2023;111(1301–15):e5. 10.1016/j.neuron.2023.01.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Chu X, Zhong L, Dan W, Wang X, Zhang Z, Liu Z, Lu Y, Shao X, Zhou Z, Chen S, Liu B. DNMT3A R882H mutation promotes acute leukemic cell survival by regulating glycolysis through the NRF2/NQO1 axis. Cell Signal. 2023;105:110626. 10.1016/j.cellsig.2023.110626. [DOI] [PubMed] [Google Scholar]
  10. Cunnane SC, Trushina E, Morland C, Prigione A, Casadesus G, Andrews ZB, Beal MF, Bergersen LH, Brinton RD, de la Monte S, Eckert A, Harvey J, Jeggo R, Jhamandas JH, Kann O, la Cour CM, Martin WF, Mithieux G, Moreira PI, Murphy MP, Nave KA, Nuriel T, Oliet SHR, Saudou F, Mattson MP, Swerdlow RH, Millan MJ. Brain energy rescue: an emerging therapeutic concept for neurodegenerative disorders of ageing. Nat Rev Drug Discov. 2020;19:609–33. 10.1038/s41573-020-0072-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dangoumau A, Marouillat S, Coelho R, Wurmser F, Brulard C, Haouari S, Laumonnier F, Corcia P, Andres CR, Blasco H, Vourc'h P. Dysregulations of Expression of Genes of the Ubiquitin/SUMO Pathways in an In Vitro Model of Amyotrophic Lateral Sclerosis Combining Oxidative Stress and SOD1 Gene Mutation. Int J Mol Sci. 2021;22. 10.3390/ijms22041796 [DOI] [PMC free article] [PubMed]
  12. Efremova L, Schildknecht S, Adam M, Pape R, Gutbier S, Hanf B, Burkle A, Leist M. Prevention of the degeneration of human dopaminergic neurons in an astrocyte co-culture system allowing endogenous drug metabolism. Br J Pharmacol. 2015;172:4119–32. 10.1111/bph.13193. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. George M, Tharakan M, Culberson J, Reddy AP, Reddy PH. Role of Nrf2 in aging, Alzheimer’s and other neurodegenerative diseases. Ageing Res Rev. 2022;82:101756. 10.1016/j.arr.2022.101756. [DOI] [PubMed] [Google Scholar]
  14. Habib N, McCabe C, Medina S, Varshavsky M, Kitsberg D, Dvir-Szternfeld R, Green G, Dionne D, Nguyen L, Marshall JL, Chen F, Zhang F, Kaplan T, Regev A, Schwartz M. Disease-associated astrocytes in Alzheimer’s disease and aging. Nat Neurosci. 2020;23:701–6. 10.1038/s41593-020-0624-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Hampel H, Hardy J, Blennow K, Chen C, Perry G, Kim SH, Villemagne VL, Aisen P, Vendruscolo M, Iwatsubo T, Masters CL, Cho M, Lannfelt L, Cummings JL, Vergallo A. The Amyloid-beta Pathway in Alzheimer’s Disease. Mol Psychiatry. 2021;26:5481–503. 10.1038/s41380-021-01249-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Johnson ECB, Dammer EB, Duong DM, Ping L, Zhou M, Yin L, Higginbotham LA, Guajardo A, White B, Troncoso JC, Thambisetty M, Montine TJ, Lee EB, Trojanowski JQ, Beach TG, Reiman EM, Haroutunian V, Wang M, Schadt E, Zhang B, Dickson DW, Ertekin-Taner N, Golde TE, Petyuk VA, De Jager PL, Bennett DA, Wingo TS, Rangaraju S, Hajjar I, Shulman JM, Lah JJ, Levey AI, Seyfried NT. Large-scale proteomic analysis of Alzheimer’s disease brain and cerebrospinal fluid reveals early changes in energy metabolism associated with microglia and astrocyte activation. Nat Med. 2020;26:769–80. 10.1038/s41591-020-0815-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Kim S, Chun H, Kim Y, Kim Y, Park U, Chu J, Bhalla M, Choi SH, Yousefian-Jazi A, Kim S, Hyeon SJ, Kim S, Kim Y, Ju YH, Lee SE, Lee H, Lee K, Oh SJ, Hwang EM, Lee J, Lee CJ, Ryu H. Astrocytic autophagy plasticity modulates Abeta clearance and cognitive function in Alzheimer’s disease. Mol Neurodegener. 2024b;19:55. 10.1186/s13024-024-00740-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Kim JJ, Vitale D, Otani DV, Lian MM, Heilbron K, andMe Research T, Iwaki H, Lake J, Solsberg CW, Leonard H, Makarious MB, Tan EK, Singleton AB, Bandres-Ciga S, Noyce AJ, Global Parkinson's Genetics P, Blauwendraat C, Nalls MA, Foo JN, Mata I. Multi-ancestry genome-wide association meta-analysis of Parkinson's disease. Nat Genet. 2024a;56:27-36. 10.1038/s41588-023-01584-8 [DOI] [PMC free article] [PubMed]
  19. Kopp F, Elguindy MM, Yalvac ME, Zhang H, Chen B, Gillett FA, Lee S, Sivakumar S, Yu H, Xie Y, Mishra P, Sahenk Z, Mendell JT. PUMILIO hyperactivity drives premature aging of Norad-deficient mice. Elife. 2019;8. 10.7554/eLife.42650 [DOI] [PMC free article] [PubMed]
  20. Lane DJR, Metselaar B, Greenough M, Bush AI, Ayton SJ. Ferroptosis and NRF2: an emerging battlefield in the neurodegeneration of Alzheimer’s disease. Essays Biochem. 2021;65:925–40. 10.1042/EBC20210017. [DOI] [PubMed] [Google Scholar]
  21. Le Douce J, Maugard M, Veran J, Matos M, Jego P, Vigneron PA, Faivre E, Toussay X, Vandenberghe M, Balbastre Y, Piquet J, Guiot E, Tran NT, Taverna M, Marinesco S, Koyanagi A, Furuya S, Gaudin-Guerif M, Goutal S, Ghettas A, Pruvost A, Bemelmans AP, Gaillard MC, Cambon K, Stimmer L, Sazdovitch V, Duyckaerts C, Knott G, Herard AS, Delzescaux T, Hantraye P, Brouillet E, Cauli B, Oliet SHR, Panatier A, Bonvento G. Impairment of Glycolysis-Derived l-Serine Production in Astrocytes Contributes to Cognitive Deficits in Alzheimer’s Disease. Cell Metab. 2020;31(503–17):e8. 10.1016/j.cmet.2020.02.004. [DOI] [PubMed] [Google Scholar]
  22. Lee L, Dale E, Staniszewski A, Zhang H, Saeed F, Sakurai M, Fa M, Orozco I, Michelassi F, Akpan N, Lehrer H, Arancio O. Regulation of synaptic plasticity and cognition by SUMO in normal physiology and Alzheimer’s disease. Sci Rep. 2014;4:7190. 10.1038/srep07190. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Lepiarz-Raba I, Gbadamosi I, Florea R, Paolicelli RC, Jawaid A. Metabolic regulation of microglial phagocytosis: Implications for Alzheimer’s disease therapeutics. Transl Neurodegener. 2023;12:48. 10.1186/s40035-023-00382-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Li SR, Kang NN, Wang RR, Li MD, Chen LH, Zhou P, Xu DX, Zhao H, Fu L. ALKBH5 SUMOylation-mediated FBXW7 m6A modification regulates alveolar cells senescence during 1-nitropyrene-induced pulmonary fibrosis. J Hazard Mater. 2024;468:133704. 10.1016/j.jhazmat.2024.133704. [DOI] [PubMed] [Google Scholar]
  25. Lin Y, Kwok S, Hein AE, Thai BQ, Alabi Y, Ostrowski MS, Wu K, Floor SN. RNA molecular recording with an engineered RNA deaminase. Nat Methods. 2023;20:1887–99. 10.1038/s41592-023-02046-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Linnerbauer M, Wheeler MA, Quintana FJ. Astrocyte Crosstalk in CNS Inflammation. Neuron. 2020;108:608–22. 10.1016/j.neuron.2020.08.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Liu Q, Zhu L, Liu X, Zheng J, Liu Y, Ruan X, Cao S, Cai H, Li Z, Xue Y. TRA2A-induced upregulation of LINC00662 regulates blood-brain barrier permeability by affecting ELK4 mRNA stability in Alzheimer’s microenvironment. RNA Biol. 2020;17:1293–308. 10.1080/15476286.2020.1756055. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Liu Q, Liu X, Zhao D, Ruan X, Su R, Shang X, Wang D, Yang C, Xue Y. Pseudogene ACTBP2 increases blood-brain barrier permeability by promoting KHDRBS2 transcription through recruitment of KMT2D/WDR5 in Abeta(1-)(42) microenvironment. Cell Death Discov. 2021;7:142. 10.1038/s41420-021-00531-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Liu W, Wang Y, Bozi LHM, Fischer PD, Jedrychowski MP, Xiao H, Wu T, Darabedian N, He X, Mills EL, Burger N, Shin S, Reddy A, Sprenger HG, Tran N, Winther S, Hinshaw SM, Shen J, Seo HS, Song K, Xu AZ, Sebastian L, Zhao JJ, Dhe-Paganon S, Che J, Gygi SP, Arthanari H, Chouchani ET. Lactate regulates cell cycle by remodelling the anaphase promoting complex. Nature. 2023;616:790–7. 10.1038/s41586-023-05939-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Lyu F, Gong H, Wu X, Liu X, Lu Y, Wei X, Liu C, Shen Y, Wang Y, Lei L, Chen J, Ma S, Sun H, Yu D, Han J, Xu Y, Wu D. Dimethyl fumarate ameliorates chronic graft-versus-host disease by inhibiting Tfh differentiation via Nrf2. Leukemia. 2024. 10.1038/s41375-024-02475-5. [DOI] [PubMed] [Google Scholar]
  31. Ma Q. Role of nrf2 in oxidative stress and toxicity. Annu Rev Pharmacol Toxicol. 2013;53:401–26. 10.1146/annurev-pharmtox-011112-140320. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Martinez JC, Randolph LK, Iascone DM, Pernice HF, Polleux F, Hengst U. Pum2 Shapes the Transcriptome in Developing Axons through Retention of Target mRNAs in the Cell Body. Neuron. 2019;104(931–46):e5. 10.1016/j.neuron.2019.08.035. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Mizuno T, Kuno R, Nitta A, Nabeshima T, Zhang G, Kawanokuchi J, Wang J, Jin S, Takeuchi H, Suzumura A. Protective effects of nicergoline against neuronal cell death induced by activated microglia and astrocytes. Brain Res. 2005;1066:78–85. 10.1016/j.brainres.2005.10.050. [DOI] [PubMed] [Google Scholar]
  34. Mossmann D, Vogtle FN, Taskin AA, Teixeira PF, Ring J, Burkhart JM, Burger N, Pinho CM, Tadic J, Loreth D, Graff C, Metzger F, Sickmann A, Kretz O, Wiedemann N, Zahedi RP, Madeo F, Glaser E, Meisinger C. Amyloid-beta peptide induces mitochondrial dysfunction by inhibition of preprotein maturation. Cell Metab. 2014;20:662–9. 10.1016/j.cmet.2014.07.024. [DOI] [PubMed] [Google Scholar]
  35. Nakano-Kobayashi A, Canela A, Yoshihara T, Hagiwara M. Astrocyte-targeting therapy rescues cognitive impairment caused by neuroinflammation via the Nrf2 pathway. Proc Natl Acad Sci U S A. 2023;120:e2303809120. 10.1073/pnas.2303809120. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Ossenkoppele R, van der Kant R, Hansson O. Tau biomarkers in Alzheimer’s disease: towards implementation in clinical practice and trials. Lancet Neurol. 2022;21:726–34. 10.1016/S1474-4422(22)00168-5. [DOI] [PubMed] [Google Scholar]
  37. Pang K, Jiang R, Zhang W, Yang Z, Li LL, Shimozawa M, Tambaro S, Mayer J, Zhang B, Li M, Wang J, Liu H, Yang A, Chen X, Liu J, Winblad B, Han H, Jiang T, Wang W, Nilsson P, Guo W, Lu B. An App knock-in rat model for Alzheimer’s disease exhibiting Abeta and tau pathologies, neuronal death and cognitive impairments. Cell Res. 2022;32:157–75. 10.1038/s41422-021-00582-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Qin Y, Zhang X, Song Y, Zhong B, Liu L, Wang D, Zhang Y, Lu W, Zhao X, Jia Z, Li M, Zhang L, Qing G. A highly sensitive nanochannel device for the detection of SUMO1 peptides. Chem Sci. 2023;14:8360–8. 10.1039/d3sc02140h. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Ryan EM, Sadiku P, Coelho P, Watts ER, Zhang A, Howden AJM, Sanchez-Garcia MA, Bewley M, Cole J, McHugh BJ, Vermaelen W, Ghesquiere B, Carmeliet P, Rodriguez Blanco G, Von Kriegsheim A, Sanchez Y, Rumsey W, Callahan JF, Cooper G, Parkinson N, Baillie K, Cantrell DA, McCafferty J, Choudhury G, Singh D, Dockrell DH, Whyte MKB, Walmsley SR. NRF2 Activation Reprograms Defects in Oxidative Metabolism to Restore Macrophage Function in Chronic Obstructive Pulmonary Disease. Am J Respir Crit Care Med. 2023;207:998–1011. 10.1164/rccm.202203-0482OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Sadeghdoust M, Das A, Kaushik DK. Fueling neurodegeneration: metabolic insights into microglia functions. J Neuroinflammation. 2024;21:300. 10.1186/s12974-024-03296-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Shangguan X, He J, Ma Z, Zhang W, Ji Y, Shen K, Yue Z, Li W, Xin Z, Zheng Q, Cao Y, Pan J, Dong B, Cheng J, Wang Q, Xue W. SUMOylation controls the binding of hexokinase 2 to mitochondria and protects against prostate cancer tumorigenesis. Nat Commun. 2021;12:1812. 10.1038/s41467-021-22163-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. She L, Sun J, Xiong L, Li A, Li L, Wu H, Ren J, Wang W, Liang G, Zhao X. Ginsenoside RK1 improves cognitive impairments and pathological changes in Alzheimer’s disease via stimulation of the AMPK/Nrf2 signaling pathway. Phytomedicine. 2024;122:155168. 10.1016/j.phymed.2023.155168. [DOI] [PubMed] [Google Scholar]
  43. Sun XX, Chen Y, Su Y, Wang X, Chauhan KM, Liang J, Daniel CJ, Sears RC, Dai MS. SUMO protease SENP1 deSUMOylates and stabilizes c-Myc. Proc Natl Acad Sci U S A. 2018;115:10983–8. 10.1073/pnas.1802932115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Theparambil SM, Kopach O, Braga A, Nizari S, Hosford PS, Sagi-Kiss V, Hadjihambi A, Konstantinou C, Esteras N, Gutierrez Del Arroyo A, Ackland GL, Teschemacher AG, Dale N, Eckle T, Andrikopoulos P, Rusakov DA, Kasparov S, Gourine AV. Adenosine signalling to astrocytes coordinates brain metabolism and function. Nature. 2024;632:139–46. 10.1038/s41586-024-07611-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Traxler L, Herdy JR, Stefanoni D, Eichhorner S, Pelucchi S, Szucs A, Santagostino A, Kim Y, Agarwal RK, Schlachetzki JCM, Glass CK, Lagerwall J, Galasko D, Gage FH, D’Alessandro A, Mertens J. Warburg-like metabolic transformation underlies neuronal degeneration in sporadic Alzheimer’s disease. Cell Metab. 2022;34(1248–63):e6. 10.1016/j.cmet.2022.07.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Uruno A, Yamamoto M. The KEAP1-NRF2 System and Neurodegenerative Diseases. Antioxid Redox Signal. 2023;38:974–88. 10.1089/ars.2023.0234. [DOI] [PubMed] [Google Scholar]
  47. Uyhazi KE, Yang Y, Liu N, Qi H, Huang XA, Mak W, Weatherbee SD, de Prisco N, Gennarino VA, Song X, Lin H. Pumilio proteins utilize distinct regulatory mechanisms to achieve complementary functions required for pluripotency and embryogenesis. Proc Natl Acad Sci U S A. 2020;117:7851–62. 10.1073/pnas.1916471117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Vertegaal ACO. Signalling mechanisms and cellular functions of SUMO. Nat Rev Mol Cell Biol. 2022;23:715–31. 10.1038/s41580-022-00500-y. [DOI] [PubMed] [Google Scholar]
  49. Veys K, Fan Z, Ghobrial M, Bouche A, Garcia-Caballero M, Vriens K, Conchinha NV, Seuwen A, Schlegel F, Gorski T, Crabbe M, Gilardoni P, Ardicoglu R, Schaffenrath J, Casteels C, De Smet G, Smolders I, Van Laere K, Abel ED, Fendt SM, Schroeter A, Kalucka J, Cantelmo AR, Walchli T, Keller A, Carmeliet P, De Bock K. Role of the GLUT1 Glucose Transporter in Postnatal CNS Angiogenesis and Blood-Brain Barrier Integrity. Circ Res. 2020;127:466–82. 10.1161/CIRCRESAHA.119.316463. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Wang D, Ruan X, Liu X, Xue Y, Shao L, Yang C, Zhu L, Yang Y, Li Z, Yu B, Feng T, Liu Y. SUMOylation of PUM2 promotes the vasculogenic mimicry of glioma cells via regulating CEBPD. Clin Transl Med. 2020;10:e168. 10.1002/ctm2.168. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Wei J, Wang H, Zheng Q, Zhang J, Chen Z, Wang J, Ouyang L, Wang Y. Recent research and development of inhibitors targeting sentrin-specific protease 1 for the treatment of cancers. Eur J Med Chem. 2022;241:114650. 10.1016/j.ejmech.2022.114650. [DOI] [PubMed] [Google Scholar]
  52. Wolozin B, Ivanov P. Stress granules and neurodegeneration. Nat Rev Neurosci. 2019;20:649–66. 10.1038/s41583-019-0222-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Yamashita R, Long J, Longacre T, Peng L, Berry G, Martin B, Higgins J, Rubin DL, Shen J. Deep learning model for the prediction of microsatellite instability in colorectal cancer: a diagnostic study. Lancet Oncol. 2021;22:132–41. 10.1016/S1470-2045(20)30535-0. [DOI] [PubMed] [Google Scholar]
  54. Yan L, Zhang T, Wang K, Chen Z, Yang Y, Shan B, Sun Q, Zhang M, Zhang Y, Zhong Y, Liu N, Gu J, Xu D. SENP1 prevents steatohepatitis by suppressing RIPK1-driven apoptosis and inflammation. Nat Commun. 2022;13:7153. 10.1038/s41467-022-34993-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Yang Y, Huang L, Zhang N, Deng YN, Cao X, Liang Y, Hou H, Luo Y, Yang Y, Li Q, Liang S. SUMOylation of annexin A6 retards cell migration and tumor growth by suppressing RHOU/AKT1-involved EMT in hepatocellular carcinoma. Cell Commun Signal. 2024;22:206. 10.1186/s12964-024-01573-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Zhang M, Chen D, Xia J, Han W, Cui X, Neuenkirchen N, Hermes G, Sestan N, Lin H. Post-transcriptional regulation of mouse neurogenesis by Pumilio proteins. Genes Dev. 2017;31:1354–69. 10.1101/gad.298752.117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Zhang W, Feng C, Jiang H. Novel target for treating Alzheimer’s Diseases: Crosstalk between the Nrf2 pathway and autophagy. Ageing Res Rev. 2021;65:101207. 10.1016/j.arr.2020.101207. [DOI] [PubMed] [Google Scholar]
  58. Zhang D, Zhang Y, Pan J, Cao J, Sun X, Li X, Zhang L, Qin C. Degradation of NLRP3 by p62-dependent-autophagy improves cognitive function in Alzheimer’s disease by maintaining the phagocytic function of microglia. CNS Neurosci Ther. 2023;29:2826–42. 10.1111/cns.14219. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Zhang L, Zhang J, Ye ZW, Muhammad A, Li L, Culpepper JW, Townsend DM, Tew KD. Adaptive changes in tumor cells in response to reductive stress. Biochem Pharmacol. 2024;219:115929. 10.1016/j.bcp.2023.115929. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Zhang X, Wang R, Hu D, Sun X, Fujioka H, Lundberg K, Chan ER, Wang Q, Xu R, Flanagan ME, Pieper AA, Qi X. Oligodendroglial glycolytic stress triggers inflammasome activation and neuropathology in Alzheimer's disease. Sci Adv. 2020;6. 10.1126/sciadv.abb8680 [DOI] [PMC free article] [PubMed]
  61. Zhao L, Kang M, Liu X, Wang Z, Wang Y, Chen H, Liu W, Liu S, Li B, Li C, Chang A, Tang B. UBR7 inhibits HCC tumorigenesis by targeting Keap1/Nrf2/Bach1/HK2 and glycolysis. J Exp Clin Cancer Res. 2022;41:330. 10.1186/s13046-022-02528-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

ESM1 (2.5MB, pdf)

(PDF 2.52 MB)

ESM2 (3.2MB, docx)

(DOCX 3.17 MB)

ESM3 (20.6KB, docx)

(DOCX 20.5 KB)

ESM4 (28.5KB, docx)

(DOCX 28.4 KB)

Data Availability Statement

The data used that support the fndings of this study are available from the corresponding author on reasonable request.

Articles from Cell Biology and Toxicology are provided here courtesy of Springer

RESOURCES