Abstract
Endoplasmic reticulum (ER) stress is crucial in cerebral ischemia/reperfusion injury by triggering cellular apoptosis and exacerbating neuronal damage. This study elucidates the dynamics of TP53-induced glycolysis and apoptosis regulator (TIGAR) translocation and its role in regulating neural fate during cerebral ischemia-induced ER stress, specifically in male mice. We found enhanced nuclear localization of TIGAR in neurons after transient middle cerebral artery occlusion/reperfusion (tMCAO/R) in male mice, as well as oxygen glucose deprivation/reperfusion (OGD/R) and treatment with ER stress inducer (tunicamycin and thapsigargin) in neuronal cells. Conditional neuronal knockdown of Tigar aggravated the injury following ischemia-reperfusion, whereas overexpression of Tigar attenuated cerebral ischemic injury and ameliorated intraneuronal ER stress. Additionally, TIGAR overexpression reduced the elevation of ATF4 target genes and attenuated ER stress-induced cell death. Notably, TIGAR colocalized and interacted with ATF4 in the nucleus, inhibiting its downstream proapoptotic gene transcription, consequently protecting against ischemic injury. In vitro and in vivo experiments revealed that ATF4 overexpression reversed the protective effects of TIGAR against cerebral ischemic injury. Intriguingly, our study identified the Q141/K145 residues of TIGAR, crucial for its nuclear translocation and interaction with ATF4, highlighting a novel aspect of TIGAR's function distinct from its known phosphatase activity or mitochondrial localization domains. These findings reveal a novel neuroprotective mechanism of TIGAR in regulating ER stress through ATF4-mediated signaling pathways. These insights may guide targeted therapeutic strategies to protect neuronal function and alleviate the deleterious effects of cerebral ischemic injury.
- ATF4
- cerebral ischemia/reperfusion
- endoplasmic reticulum stress
- neuroprotection
- nuclear translocation
- TIGAR
Significance Statement
TIGAR (TP53-induced glycolysis and apoptosis regulator) is one of the downstream target genes of p53, and its encoded protein exerts Fru-2, 6-BPase activity to promote glucose metabolic flux to pentose phosphate pathway. However, the nonenzymatic function of TIGAR has been gradually discovered. Here, we demonstrate that TIGAR translocates to the nucleus to interact with ATF4 in neurons after cerebral ischemia/reperfusion-induced ER stress via its Q141/K145 residues. Then TIGAR inhibits ATF4's downstream proapoptotic genes expression, reducing ER stress-dependent apoptosis, consequently alleviating neuronal damage. This study uncovered a novel neuroprotective mechanism of TIGAR by regulating ER stress via ATF4-mediated signaling pathway. The Q141/K145 residues of TIGAR are critical for its interaction with ATF4 and inhibition of ATF4 target genes.
Introduction
Stroke, characterized by high disability and mortality rates, ranks among the foremost global causes of death. Ischemic stroke, accounting for 70–80% of all strokes, arises from cerebrovascular occlusion due to thrombosis (Donkor, 2018; Fan et al., 2023; Tu et al., 2023). Current therapeutic strategies include pharmacotherapy, interventional procedures, and interventions aimed at restoring cerebral blood flow. Given the complex pathological mechanisms involved, there is a pressing need to investigate more efficacious approaches for both treatment and prognosis. Ischemic stroke disrupts endoplasmic reticulum (ER) function, initiating ER stress and activating the unfolded protein response (UPR). The subsequent cellular adaptations dictate the cell's fate—either survival or death—following ischemia (Wang et al., 2020; L. Wang et al., 2022). Specifically, brain ischemia/reperfusion (I/R)-induced glucose metabolism disorder, inflammation, calcium dysregulation, and increased reactive oxygen species (ROS) all contribute to ER dysfunction. This disruption leads to the accumulation of unfolded and misfolded proteins in the ER lumen, exacerbating ischemic injury (Gong et al., 2017; P. Wang et al., 2022; Guo et al., 2023). Additionally, I/R triggers PERK/ATF4/CHOP-related apoptotic pathways, evidenced by increased expression of p-eIF2α, ATF4, CHOP, and cleaved-caspase12. CHOP has been recognized as a pivotal mediator of ER stress-induced neuronal apoptosis following ischemia/reperfusion injury (Li et al., 2020; Han et al., 2021). Blocking PERK-CHOP-mediated apoptosis has demonstrated protective effects against cerebral ischemia (Fei et al., 2021). A comprehensive understanding of the ER stress mechanisms during cerebral ischemia holds the promise of developing more effective treatment strategies.
The TP53-induced glycolysis and apoptosis regulator (TIGAR) plays a significant role in regulating cellular glycolysis and metabolic homeostasis by inhibiting glycolysis and promoting glucose metabolism to the pentose phosphate pathway (PPP; Bensaad et al., 2006). This increases the production of nicotinamide adenine dinucleotide phosphate (NADPH), thereby reducing intracellular ROS and protecting cells from oxidative damage and apoptosis (Green and Chipuk, 2006; Li et al., 2014; Blacker and Duchen, 2016). Our laboratory previously reported TIGAR's neuroprotective effects in cerebral ischemia/reperfusion, and cerebral preconditioning is attributed to its upregulation of NADPH levels. Downregulation of TIGAR expression aggravates cerebral ischemia/reperfusion-induced neuronal damage and abolishes the neuroprotective effects of cerebral preconditioning, whereas TIGAR overexpression confers protection against cerebral ischemia/reperfusion injury, ultimately improving survival rate and the restoration of motor and cognitive function in ischemic stroke animal models (Li et al., 2014; Zhou et al., 2016). Furthermore, TIGAR plays a crucial role in preventing brain endothelial tight junction injury and inhibiting NF-κB-mediated astrocyte inflammation (Wang et al., 2017; Chen et al., 2018). Clinical evidence suggests that higher serum TIGAR mRNA levels correlate with favorable outcomes in ischemic stroke patients (Changli, 2020). Despite these findings, an unresolved question remains: does TIGAR's neuroprotection stem solely from its phosphatase activity, or are there additional, yet undiscovered mechanisms at play?
TIGAR is widely distributed in various cellular compartments, including mitochondria, ER, cytoplasm and nucleus (Cheung et al., 2012; Li et al., 2014; Yu et al., 2015; Geng et al., 2019). Intriguingly, our preliminary experiments have revealed that TIGAR can translocate to the nucleus during cerebral ischemia in mice or ER stress in neuronal cells. During ER stress, nuclear transcription factors, such as ATF4, CHOP, and X-box binding protein 1 (XBP-1) regulate downstream signaling pathways associated ER stress (Qi et al., 2023). Therefore, it is plausible to hypothesize that TIGAR may translocate to the nucleus during cerebral ischemia or ER stress, where it acts upon certain transcription factors to regulate ER stress, thereby protecting against ischemic neuronal injury.
In this study, we established in vivo and in vitro models of I/R or ER stress to investigate the nuclear translocation of TIGAR during cerebral ischemia. Furthermore, we explored how nuclear translocation of TIGAR regulates ER stress via ATF4/CHOP pathway. The findings expand our understanding of the neuroprotective mechanisms by which TIGAR attenuates cerebral ischemic injury.
Materials and Methods
Materials
Dulbecco's modified Eagle's medium (DMEM), fetal bovine serum (FBS), Neurobasal medium (NBM), B27, and Glutathione magnetic agarose were purchased from Thermo Fisher Scientific. Poly d-lysine and puromycin were purchased from Merck Sigma-Aldrich. Tunicamycin (TM, 12819) was purchased from Cell Signaling Technology. Thapsigargin (Tg, 1138) was purchased from Tocris. RNAiso Plus, PrimeScript RT Master Mix, and SyBR Green were purchased from Takara. Incubator chamber (MC-101) was purchase from Billups-Rothenberg.
Transient middle cerebral artery occlusion model in mice
ICR mice (male, 8–10 weeks, 25–30 g) and ICR pregnant mice (∼17 d of gestation) were obtained from the Experimental Animal Center of Soochow University (Certificate No. 20020008, Grade II) under controlled conditions: temperature (20–25°C), humidity (40–70%), 12 h artificial day/night. Tigar transgenic mice and matched WT mice were generated by Model Animal Research Center of Nanjing University. Tigar transgenic mice were created through pronuclear injection of a pcDNA3.1 vector cloned with TIGAR DNA (provided by Karen H. Vousden, Beatson Cancer Institute) in C57BL/6 zygotes. The control mice are born in the same litter as the transgenic mice but do not contain the TIGAR overexpression gene mentioned above. Tigar-flox mice (Strain NO. T014798) were purchased from GemPharmatech. They were generated by inserting loxP sites on both sides of exon 3 of the Tigar gene. These floxed mice were mated with Nestin-Cre (003771, from The Jackson Laboratory; donated by Dr. Ji Geng) tool mice to obtain TigarNestin-conditional knock-out (TigarNestin-CKO) mice (Geng et al., 2019). The experimental genotypes were neuron-specific Tigar-knock-out mice (TigarNestin-CKO) and floxed control (Tigarfl/fl). The littermates were used for experiments. Mice genotype was identified by PCR. The following primers were used for PCR: TG-Tigar: forward: 5′-ATG GCT CGC TTC GCT CTG-3′; reverse: 5′-CTT CCC TGG CTG CTT TGG-3′; Tigar-flox: Tigar 1: forward: 5′-AAG TGC TGG GAT TAA AGG TGT GTG-3′; reverse: 5′-GCA CGC AAT GAT ATT ACT CCT CTT CC-3′; Tigar 2: forward: 5′-AGA ACC CAG ATG GGC ACA TGA-3′; reverse: 5′-CCC TGT GGC TTC TTA TGT GTG TG-3′; Nestin-Cre: cre-up: 5′-GCC TGC ATT ACC GGT CGA TGC-3′; cre-low: 5′-CAG GGT GTT ATA AGC AAT CCC-3′. All animal experiments adhered to the guidelines approved by the Ethical Committee of Soochow University (Approval No.: SUDA20230925A05) and were in complete compliance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. For in vivo experiments, groups were randomized and procedures were performed in a blinded manner.
The cerebral ischemia/reperfusion model included 2 h of transient middle cerebral artery occlusion (tMCAO) and subsequent restoration of blood flow, conducted as described previously (Clark et al., 1997; Chen et al., 2020). Briefly, the origin of right middle cerebral artery was occluded by a 6-0 silicone-coated suture (Doccol), and the suture was removed 2 h later to restore blood flow. Cerebral blood flow was monitored using a Laser Doppler blood flow meter (LDFML191). Mice were maintained at their normal body temperature with a thermostatic blanket. Sham-operated mice underwent the same surgery without suture insertion. After cerebral ischemia/reperfusion, mice were killed at different time points (1, 3, 6, 12 or 24 h).
Neurological deficit scoring and triphenyltetrazolium chloride staining
The Longa Score Scale was used to assess the neurological behaviors of the mice 24 h after reperfusion. The rating scale used was as described previously (Longa et al., 1989). The assessment was conducted by an investigator blinded to the experimental treatments. Mice were killed 24 h after tMCAO/reperfusion. The mouse brains were rapidly frozen at −20°C for 15 min, sectioned into five coronal slices (2 mm thickness), and immersed in a 1% triphenyltetrazolium chloride (TTC) solution at 42°C for 10 min. Following color development, the brain slices were fixed in 4% paraformaldehyde for subsequent scanning and imaging.
Cell culture
Primary culture cortical neurons were obtained from cerebral cortex of ICR or C57 embryos at 17 d of gestation. In brief, fetal cortices were dissected under aseptic conditions, quickly placed in precooled PBS, and minced. Digestion was started with 2.5% trypsin and terminated with DMEM (Invitrogen) containing 10% FBS (Invitrogen) after 15 min reaction at 37°C. Then the cell suspension was incubated with DNase I (Solarbio), blown for 3 min, and centrifuged for 5 min at 1,500 rpm. The pellets were resuspended in NBM containing 2% B27, 1% penicillin G and streptomycin, 0.5 mM ʟ-glutamine, 25 μM ʟ-glutamate. Cells were seeded on culture dishes coated with poly d-lysine after filtered through a 40 μm cell strainer. When the neurons grew out of antennae, they were cultured in NBM containing 0.5 mM ʟ-glutamine, 2% B27, 1% penicillin G and streptomycin. Half of the medium was replaced every 3 d. Generally, primary neurons were cultured in vitro for 7–10 d for experiments (Sheng et al., 2012, 2014).
HT22 mouse hippocampal neuronal cells (CVCL_0321) and human embryonic kidney 293 Tet-on (HEK293T) cells were purchased from Shanghai Institute of Cell Biology. Cells were cultured in DMEM containing 10% FBS, 100 U/ml streptomycin, 100 U/ml penicillin in 5% CO2/95% air at 37°C.
Oxygen and glucose deprivation–reperfusion and ER stress models
For oxygen and glucose deprivation–reperfusion (OGD/R), the cells were transferred to a medium containing HEPES-balanced salt solution (HBSS; in mM: 10 HEPES, 140 NaCl, 3.5 KCl, 12 MgSO4, 5 NaHCO3, 1.7 CaCl2, 0.4 KH2PO4). Then the cells were placed in an incubator chamber aerated with 5% CO2/95% N2 and cultured at 37°C for 4 h (primary neurons) or 6 h (HT22). To restore oxygen and glucose, cells were replaced with normal medium and transferred from hypoxia chamber to incubator containing 5% CO2/95% air (Sheng et al., 2012).
HT22 cells were treated with tunicamycin (TM) 12 μM or thapsigargin (Tg) 1 μM to induce ER stress (van Vliet et al., 2017; Jeong et al., 2019). Cells were harvested at the indicated time. HEK293T cells were treated with TM 12 μM to induce ER stress. Cells were harvested at 12 h after treatment. TM or Tg was initially prepared with dimethylsulfoxide (DMSO) and diluted with medium before use (the final DMSO concentration < 0.1%).
Cell viability assay
After OGD/R or TM treatment, the medium was replaced with Cell Counting Kit-8 (CCK-8, Dojindo Laboratories) diluent (CCK-8: DMEM, 1:10). Following a 2 h incubation at 37°C, absorbance at 450 nm was measured using a microplate reader (ELX800, Bio-Tek).
Lentivirus, plasmids, or adeno-associated virus infection
Lentiviruses-TIGAR (Ubi-MCS-3FLAG-CBh-gcGFP-IRES-Puro, 2 × 108 TU/ml, Gene ID: 319801, NM_177003) and lentiviruses-ATF4 (Ubi-MCS-SV40-EGFP-IRES-Puro, 1 × 108 TU/ml, Gene ID: 11911, NM_009716) were synthesized by GeneChem. The culture medium containing LV-TIGAR or LV-vector lentivirus was incubated with HT22 or HEK293T cells for 24 h. The virus-stably-infected cells were established by puromycin (2 μg/ml) treatment for 2 weeks and determined with GFP immunofluorescence and immunoblotting (Chen et al., 2020). To overexpress exogenous ATF4, the culture medium containing LV-ATF4 or LV-vector lentivirus was incubated with control or LV-TIGAR-HT22 stably transfected cells or primary cortical neurons from WT and TG mice for 24 h. After an additional 48 h for cell lines or 4 d for primary neurons of culture in normal medium, GFP immunofluorescence and immunoblotting were used to identified the expression of ATF4.
The cDNA of WT- Tigar (NM_020375, 60-872, 813 bp), Tigar Δtm (loss of Fru 2, 6-BPase activity; Bensaad et al., 2006), Tigar Δ258-261 (unable to locate mitochondria; Cheung et al., 2012), Tigar ΔN, Tigar ΔC, Tigar 59-63 AAAAA, and Tigar Q141A, K145A mutants were cloned into p3×FLAG plasmids. The plasmids of WT-TIGAR, Δtm, and Δ258-261 were kindly provided by Prof. Xiangnan Zhang (Zhejiang University). The plasmids of ΔN, ΔC, 59-63 AAAAA, and Q141A, K145A mutants were constructed by Sangon Biotech. The transfections of HT22 or HEK293T cells with these palsmids were performed using jetOPTIMUS (Polyplus) for 24 h.
The 2,188 bp region of the promoter and the first exon of Atf4 was cloned into the pGL3-basic vector (Genewiz) to create the luciferase reporter gene plasmid p-ATF4-luc. HEK293T were cotransfected with p-ATF4-luc and vector, WT, Δtm, or Δ258-261 TIGAR mutants for 24 h using jetOPTIMUS (Jeong et al., 2019). Details of TIGAR mutants and luciferase plasmids are as follows: ATF4-luciferase: pGL3-basic, NC_000022, the promoter and the first exon (2,188 bp); TIGAR WT: p3×Flag, NM_020375, 60-872 (813 bp); TIGAR Δtm: p3×Flag, H11A, E102A, H198A mutation; TIGAR Δ258-261: p3×Flag, 258-261 deletion; TIGAR Q141A, K145A: p3×Flag, Q141A, K145A mutation.
For in vivo ATF4 overexpression, stereotactic injection of adeno-associated virus (AAV) was performed as previously described (Chen et al., 2020). AAV9-ATF4 (CMV bGlobin-MCS-EGFP-3FLAG-WPRE-hGH polyA, Gene ID: 11911, NM_009716) was made by GeneChem. AAV9-ATF4 or AAV9-negative control containing 2 × 1013 viral genome (vg) was injected into the lateral ventricle and striatum, respectively. Mice were killed after 31 d, and viral expression was verified by immunofluorescence and immunoblotting.
Flow cytometry
Twelve hours after TM treatment, cells were washed with precooled PBS, harvested, and suspended in Binding Buffer from Annexin V-PE/7-AAD Apoptosis Detection Kit (Vazyme). Subsequent incubation with Annexin V-PE and 7-AAD reagent in the dark at 25°C for 15 min was followed by a fivefold dilution. Cell apoptosis was analyzed using analytical flow cytometry (FACSCalibur).
Western blot analysis
Lysis of ipsilateral cortex or in vitro cultured cells was performed using a buffer containing 10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, 5 mM EDTA, and protease inhibitor. Protein concentration was determined using the BCA assay. Polyacrylamide gels and nitrocellulose membranes were utilized for protein separation and transfer. Immunoblotting was carried out to detect TIGAR (166290, Santa Cruz), ATF4 (11815, Cell Signaling Technology), CHOP (2895, Cell Signaling Technology), GRP78 (1050, Santa Cruz), cleaved-caspase-12 (ab62484, Abcam) levels. β-actin (A5441, Sigma-Aldrich), GAPDH (60004-1-ig, Proteintech), and Lamin B (sc-6217, Santa Cruz) were used as normalization controls for the assay of whole-cell proteins, cytoplasmic proteins, and nuclear proteins, respectively.
Immunofluorescence
For immunofluorescence in frozen mouse brain sections after ischemia/reperfusion, brains were perfused with PBS and 4% paraformaldehyde (Sheng et al., 2010; Li et al., 2014). After fixation and sucrose dehydration, 20-μm-thick coronal sections were cut with a cryostat. After 1 h of blocking in 5% BSA and 0.4% Triton X-100, sections were incubated with ATF4 (1:100), TIGAR (1:100), or FLAG (1:100, 14793, Cell Signaling Technology) for 48 h, followed by incubation with Coralite488-conjugated anti-Mouse IgG (1:500, A7028, Beyotime) and Coralite594-conjugated anti-Rabbit IgG (1:500, A7016, Beyotime) for 3 h. For immunofluorescence in vitro, HT22 cells were fixed with 4% paraformaldehyde for 15 min after 12 h of TM treatment. The cells were then permeated with 0.1% Triton X-100 for 10 min and blocked with 1% BSA for 1 h, followed by incubation with TIGAR, ATF4, or FLAG antibodies at 4°C overnight, and with Coralite488-conjugated anti-Mouse IgG (1:1,000, SA00013-1, Proteintech) and Coralite594-conjugated anti-Rabbit IgG (1:1000, SA00013-4, Proteintech) for 2 h. Finally, a confocal microscope (ZEISS LSM710) was used to observe immunofluorescence of the samples after DAPI (1:10,000, D9564, Sigma) staining.
Nuclear extraction
The ipsilateral cortex was ground by glass homogenizer with buffer A containing 320 mM sucrose, 3 mM CaCl2, 2 mM MgAc, 0.14 mM EDTA, 1 mM DTT, 0.5% NP40, protease inhibitor. HT22 cells were harvested and lysed in buffer A. The cytoplasmic (supernatant) and nuclear (pellets) fractions were obtained by centrifuging at 600 × g for 15 min. To purify the nuclear fraction, the pellets were resuspended in buffer A and centrifuged at 600 × g for 10 min (Arguelles et al., 2013). Lamin B and GAPDH were used as nuclear and cytoplasmic markers.
Transmission electron microscope and immunoelectron microscopy
Six hours after cerebral ischemia/reperfusion, male C57 mice were killed. One cubic millimeter of the parietal cortex in the ischemic center was rapidly dissected, fixed in 2.5% glutaraldehyde 0.1 M phosphate buffer, pH 7.0–7.5, for 30 min (Song et al., 2017). For transmission electron microscopy, the samples were processed by Servicebio. A JEOL JEM-1230 electron microscope was used to observe the ultrastructure of ER and mitochondria in cortical neurons of mice. For immunoelectron microscopy, the sample was placed in 4°C following rinse with 0.2 mol/L (68.5 g/L) sucrose solution. Tissues were dehydrated with graded alcohol, embedded in resin, and sectioned. Sections were incubated with TIGAR (1:10) overnight at 4°C, followed by incubation with 10 nm colloidal gold secondary antibody (1:50) for 20 min at RT and 1 h at 37°C. A JEOL JEM-1230 electron microscope was used to observe black 10 nm gold particles in cortical neurons, which represent positive expression. Cortical neurons are distinguished by their characteristics, which are characterized by a large, round, and light nucleus with an obvious nucleolus; they often contain scattered RNA particles and dispersed ER profiles and have neural filaments.
RNA analysis
Tissue was isolated from mice subjected to cerebral ischemia/reperfusion, and HT22 cells were harvested at 12 h after TM treatment. Total RNAs were extracted by RNAiso Plus. Then PrimeScript RT Master Mix was used to synthesize cDNA. For RT-qPCR reaction, DNA template, primers, and SyBR Green were used on an ABI Prism 7500 Sequence Detection system (Applied Biosystems). The mRNA levels were analyzed with the comparative ΔΔCt method. The following primers were used for RT-PCR: Atf4: forward: 5′-CCT ATA AAG GCT TGC GGC CA-3′; reverse: 5′-GAT TTC GTG AAG AGC GCC AT-3′; Chop: forward: 5′-TTG AAG ATG AGC GGG TGG CAG-3′; reverse: 5′-CAC GTG GAC CAG GTT CTC TCT C-3′; β-actin: forward: 5′-GGC TGT ATT CCC CTC CAT CG-3′; reverse: 5′-CCA GTT GGT AAC AAT GCC ATG T-3′.
Co-immunoprecipitation and GST pull-down
Cortices were dissected 6 h after reperfusion and lysed in IP buffer (50 mM Tris-HCl, pH 7.4; 150 mM NaCl; 1 mM EDTA; 0.1% SDS; 1% Triton X-100; protease inhibitor). HEK293T cells were harvested and lysed in IP buffer. To eliminate nonspecifically bound proteins, the lysates were precleaned with agarose for 1–3 h. Subsequently, the lysates containing 600 μg of protein, were incubated with antibodies against ATF4 (CST), TIGAR (Santa Cruz), FLAG (Santa Cruz), or IgG overnight and then incubated with protein G-agarose for 4–8 h (Sheng et al., 2014). A portion of the lysates, containing 60 µg of protein, was used as input. Immunoblotting was performed with antibodies against ATF4, TIGAR, or FLAG to analyze the immunoprecipitates.
TIGAR GST fusion protein (Proteintech) and GST (Flarebio) were incubated with prewashed Glutathione magnetic agarose beads for 1 h at 4°C. Afterward, the GST protein beads complex washed away unbound proteins were incubated with tissue lysates overnight. Elution was performed using 50 mM Tris, pH 8.0, containing 20 mM GSH. The eluted proteins were analyzed by immunoblotting using anti-TIGAR or anti-GST antibody (1:1,000; Ren et al., 2012).
Chromatin immunoprecipitation
Simple ChIP enzymatic chromatin IP kit (Cell Signaling Technology) with agarose beads was applied to perform the chromatin immunoprecipitation (ChIP)-RT-qPCR. Twenty million LV-TIGAR-HEK293T cells were lysed and incubated with the FLAG antibody for generating each histone modification ChIP. Control for nonspecific immunoprecipitation of DNA was produced by rabbit IgG. For RT-qPCR assays, ChIP DNA was amplified for Chop and Atf3 primers (Chop: forward: 5′-GAG AAT GAA AGG AAA GTG GCA C-3′; reverse: 5′-ATT CAC CAT TCG GTC AAT CAG A-3′; Atf3: forward: 5′-TAG CCC CTG AAG AAG ATG AAA G-3′; reverse: 5′-CTT CTT CTT GTT TCG GCA CTT T-3′).
Luciferase assay
Transfected HEK293T cells were induced ER stress by TM and harvested at 12 h after treatment. Cells were lysed with the luciferase assay detection kit (Beyotime), and the supernatant was combined with an equal volume of luciferase assay solution. Fluorescence intensity of the mixture was evaluated at 560 nm with a Multiskan Spectrum (TECAN Infinite M1000 Pro).
Molecular docking
The Crystal Structure of TIGAR protein from Homo sapiens (PDB No. 3DCY) and ATF4 (PDB No. 1CI6) were referenced from the Protein Data Bank (http://www.rcsb.org). Protein docking of TIGAR and ATF4 was performed by the ZDOCK SERVER (https://zdock.umassmed.edu/, version ZDOCK 3.0.2). The structure of protein docking was visualized in PyMOL program (version 1.8; Hou et al., 2019; Ren et al., 2020).
Statistical analysis
The manuscript complied with JN's recommendations and requirements on experimental design and analysis (Li et al., 2022). Data were expressed as mean ± SD. All in vitro data represented three independent neuronal cultures with 2–3 replicates. In all animal experiments, n for statistical analysis was the number of mice for each condition. GraphPad Prism 9 was used to perform statistical analysis. The significance between two groups was compared with Student's t test, while multiple groups was determined with one-way or two-way repeated-measures ANOVA. The post hoc analysis was performed by Tukey's test. p value of <0.05 was considered as statistically significant.
Results
Neuronal TIGAR protein translocates to the nucleus in response to cerebral ischemia/reperfusion or ER stress
In preliminary experiment, we observed the nuclear translocation of TIGAR during cerebral ischemia in mice or ER stress in neuronal cells. To delineate the temporal profile of TIGAR and ER stress following cerebral ischemia/reperfusion, we assessed the expression of TIGAR and ER stress-related proteins at distinct intervals after OGD/R in primary cortical neurons and HT22 neuronal cells and tMCAO/reperfusion (I/R) in mice. In both primary neurons (Extended Data Fig. 1-1a) and HT22 cells (Extended Data Fig. 1-1b), a substantial upregulation of TIGAR and ER stress-related proteins GRP78, c-caspase12, and CHOP was evident at 3–6 h post-OGD/R treatment. Similarly, in the ipsilateral cortex of mice subjected to I/R injury, the protein levels of TIGAR and GRP78, c-caspase12, CHOP, and ATF4 were markedly elevated at 3 and 6 h postreperfusion (Extended Data Fig. 1-1c). Transcriptomic analysis of the ipsilateral cortex at 6 h post-I/R revealed a parallel increase in the expression of Tigar and ER stress-related genes (Grp78, Atf4, Chop, Gadd34, Atf3, and Trib3; Extended Data Fig. 1-1d). Tunicamycin (TM) inhibited protein glycosylation, while thapsigargin (Tg) blocks the sarcoplasmic/ER calcium ATPase resulting in calcium depletion in ER. Both agents thus cause UPR and ER stress response (Banerjee et al., 2017). HT22 cells treated with ER stress inducers, TM (12 μM) or Tg (1 μM), exhibited elevated levels of ER stress proteins and TIGAR. Notably, there was an upregulation of TIGAR observed at 6 h after TM treatment (Extended Data Fig. 1-1e). Similarly, Tg treatment also induced the expression of TIGAR and ER stress-related proteins (Extended Data Fig. 1-1f). The above data indicate that cerebral ischemia/reperfusion simultaneously upregulates the expression of TIGAR and activates ER stress in neurons.
To determine the nuclear localization of TIGAR after cerebral ischemia/reperfusion-induced ER stress, cellular fractions were extracted and the protein levels of TIGAR in the cytoplasm and nucleus were analyzed with Western blotting. Compared with the untreated group, TIGAR expression was markedly upregulated in the nucleus of the ischemic cortex and HT22 cells exposed to TM-induced ER stress (Fig. 1a,b). Immunofluorescence assay further substantiated these findings, demonstrating higher TIGAR intensity in the ischemic cortex as well as in primary cortical neurons and HT22 cells subjected to OGD/R. Intriguingly, the nuclear distribution of TIGAR was significantly increased in neurons exposed to ischemia–reperfusion, OGD/R, or ER stress (Fig. 1c–e; Extended Data Fig. 1-2). Immunoelectron microscopy assay further confirmed a higher presence of TIGAR gold particles in the nucleus of neurons in the ischemic cortex of I/R mice compared with sham animals (Fig. 1f). Collectively, these data suggest that TIGAR undergoes dynamic nuclear translocation during I/R or ER stress.
Neuronal TIGAR protein translocates to the nucleus in response to cerebral ischemia/reperfusion or ER stress. a, Mice were subjected to tMCAO/R (I/R) and cortices were dissected 6 h after reperfusion. b, HT22 cells were treated with tunicamycin (TM, 12 μM) to induce ER stress and handled at 6 h after treatment. The temporal profile of TIGAR and ER stress following cerebral ischemia/reperfusion is presented in Extended Data Figure 1-1. a, b, Nucleus and cytoplasm were extracted and subjected to Western blot analysis. Uncropped blots are shown in Extended Data Figure 1-3. c–e, The brain sections (c) or primary neurons (d) were labeled with the anti-TIGAR antibody (red), anti-NeuN antibody (green), and DAPI (blue). The immunofluorescence was examined in the peri-infarct region of cortex. e, HT22 cells undergoing OGD/R were labeled with the anti-TIGAR antibody (red) and DAPI (blue). Insets show the enlarged cells with obvious nuclear translocation of TIGAR. TIGAR fluorescence intensity fold change in the nucleus was calculated from these images. The middle line in the right graph is the mean. The analysis was based on the green, red, and blue channels of single-cell images. N = 40–60 cells per group from three independent experiments or three mice. The mean value for each mouse or independent experiment were statistically analyzed and presented the data in Extended Data Figure 1-2. Scale bar, 20 μm. The original images of all channels are shown in Extended Data Figure 1-2. f, Mice brain cortical tissues were stained with immunogold and examined with electron microscopy. Microphotographs showed immunoreactive elements (colloidal gold particles, red arrowheads) in the nucleus (N) in neurons of the mouse cerebral cortex. Scale bar, 200 nm. n = 3–6 mice or three independent neuronal cell cultures with 2–3 replicates each. Error bar represents mean ± SD. a, b, Two-way analysis of ANOVA with Sidak's post hoc test. c–e, Unpaired Student's t test.
Figure 1-1
ER stress-related proteins and TIGAR protein are increased synchronously after cerebral ischemia/reperfusion (I/R) or ER stress. Primary neurons (a) and HT22 neuronal cells (b) were subjected to oxygen glucose deprivation and reperfusion (OGD/R). Mice (c) were subjected to tMCAO/R (I/R). The cells or cortex were collected at indicated time after reperfusion. (a-c) Protein levels were measured by Western blotting. (d) Heatmap showing the Log2 (fold change) in transcript abundance for genes involved in ER stress related ATF4-CHOP pathway using the RNA-sequencing data of mice underwent tMCAO/R. Relative gene expression is indicated by color: high-expression (red), median-expression (white) and low-expression (green). (e, f) HT22 cells were treated with tunicamycin (TM 12 μM) or thapsigargin (Tg 1 μM) to induce ER stress. Cells were harvested at indicated time. Protein levels were measured by Western blotting. Uncropped blots are shown in Extended Data Figure 1-4. N = 3-6 mice or 3 independent neuronal cell cultures with 2-3 replicates each in vitro. Protein level was normalized with β-actin. Bar represents mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001 vs control group. Two-way analysis of ANOVA with Bonferroni's post hoc test. Download Figure 1-1, TIF file.
Figure 1-2
Neuronal TIGAR translocates to the nucleus during cerebral ischemia/reperfusion or ER stress. Panel a-c are the original magnified images of all channels in Figure. 1c, d and e, respectively. (a) Mice were subjected to tMCAO/R (I/R) and cortex were dissected 6 h after reperfusion. (b) Primary cortical neurons and HT22 cells were subjected to oxygen glucose deprivation and reperfusion (OGD/R). The brain sections (a) or primary neurons (b) were labeled with the anti-TIGAR antibody (red), anti-NeuN antibody (green) and DAPI (blue). HT22 cells (c) were labeled with the anti-TIGAR antibody (red) and DAPI (blue). (d-f) TIGAR fluorescence intensity fold change in the nucleus were calculated from the images of Figure. 1c, d and e, respectively. The mean value for each mouse or independent experiment were statistically analyzed and presented in the graph. The analysis was based on the green, red and blue channels of single-cell images. N = 40-60 cells per group from three independent experiments or three mice. Scale bar = 20 μm. n = 3-6 mice or 3 independent neuronal cell cultures with 2-3 replicates each. Bar represents mean ± S.D. (d-f) Unpaired Student's t test. Download Figure 1-2, TIF file.
Figure 1-3
The original uncropped blots of Figure 1. Download Figure 1-3, TIF file.
Figure 1-4
The original uncropped blots of Figure 1-1. Download Figure 1-4, TIF file.
TIGAR alleviates intraneuronal ER and mitochondrial damage following cerebral ischemia/reperfusion injury
To investigate the role of TIGAR in cerebral ischemia and its associated ER stress, we utilized Tigar transgenic mice and TigarNestin-CKO mice (Extended Data Fig. 2-1). PCR and Western blot results showed that TIGAR was significantly upregulated in Tigar transgenic (TG) mice (Extended Data Fig. 2-1a,b). TigarNestin-CKO mice were also successfully generated by crossing Tigarfl/fl mice with Nestin-Cre (Extended Data Fig. 2-1c–f). Following tMCAO/R surgery, TTC staining revealed a significant reduction in cerebral infarct areas in TG mice compared with control WT mice. Conversely, TigarNestin-CKO mice, with neuron-specific knock-out of Tigar, exhibited larger cerebral infarct areas than control Tigarfl/fl mice (Fig. 2a,c). Neurobehavioral assessments further demonstrated that Tigar overexpression ameliorated neurological deficits, whereas conditional knock-out of Tigar in neurons exacerbated neurological deficits (Fig. 2b,d). Next, we applied electron microscopy to examine the neuronal damage and organelle morphology in TG mice or TigarNestin-CKO mice after cerebral ischemia/reperfusion. In the sham group, the morphology of ER, mitochondria, and other organelles in cortical neurons of WT and TG mice were basically normal. Following I/R in WT mice, neurons exhibited markedly dilated ER with medium densities in ER lumen (representing protein aggregates), as well as swollen mitochondria with ruptured or absent cristae, indicating that I/R causes ER stress and mitochondrial damage in neurons. However, post-I/R in TG mice showed a significant improvement in the morphology of the ER and mitochondria within neurons (Fig. 2e). Similarly, Tigarfl/fl and TigarNestin-CKO mice in the sham operation group displayed no obvious changes in organelle morphology, similar to WT or TG mice. However, post-I/R in TigarNestin-CKO mice resulted in more pronounced mitochondrial damage, ER dilation, and protein aggregates within ER lumen compared with the Tigarfl/fl group (Fig. 2f). These results suggest that TIGAR may play a protective role against cerebral ischemia/reperfusion injury by alleviating ER and mitochondrial damage within neurons.
TIGAR alleviates intraneuronal ER and mitochondrial damage following cerebral ischemia/reperfusion injury. Tigar transgenic mice (TG) or their control wild-type (WT) and TIGAR neuronal conditional knock-out mice (TigarNestin-CKO) or their control mice (Tigarfl/fl) were subjected to tMCAO/R. The successful establishment of TG and TigarNestin-CKO mice is presented in Extended Data Figure 2-1. a, c, The infract regions appeared white in TTC-stained brain sections after tMCAO/R. b, d, Neurological deficits were assessed by Longa Score Scale after tMCAO/R. The middle line in the graph is the mean. e, f, The ultrastructure of neurons in the mouse cerebral cortex was examined with electron microscope (EM). The bottom row shows the enlarged endoplasmic reticulum (ER, labeled by red arrow) or mitochondria (M, labeled by red star) taken from the boxed areas. N, nucleus. Scale bar, 1 μm or 500 nm. n = 3–6 mice. Bar represents mean ± SD. Unpaired Student's t test.
Figure 2-1
The expression of TIGAR in TIGAR transgenic mice and TIGAR neuron-conditional knockout mice. (a) Brains from WT and TG mice were processed for Western blot analysis. (b) PCR analysis was performed for the identification of Wild-type (WT) and TIGAR transgenic (TG) mice. (c) The schematic diagram of CRISPR/Cas9 technology for editing the Tigar gene. (d) Strategy of mice mating. (e) Brain lysates from Tigarfl/fl and TigarNestin-CKO mice were analyzed by Western blot. Uncropped blots are shown in Extended Data Figure 2-2. (f) PCR analysis was performed for the identification of Tigarfl/fl and Nestin-Cre+/+ mice. The flox mice were knocked out after mating with Nestin-Cre+/+ mice expressing Nestin-Cre recombinase, resulting in the loss of function of Tigar in neuron. n = 6 mice. Bar represents mean ± S.D. Unpaired Student's t test. Download Figure 2-1, TIF file.
Figure 2-2
The original uncropped blots of Figure 2-1. Download Figure 2-2, TIF file.
TIGAR inhibits ER stress-induced neuronal apoptosis during cerebral ischemia/reperfusion (I/R)
To elucidate the impact of TIGAR on neuronal ER stress induced by ischemia/reperfusion, Tigar transgenic (TG) and WT mice were utilized as previously mentioned. TIGAR-overexpressed HT22 cells (LV-TIGAR-HT22) were generated and validated using immunofluorescence and Western blotting (Fig. 3a,b). Immunoblotting analysis revealed that TIGAR overexpression in mice significantly attenuated I/R-induced upregulation of ER stress-related proteins as GRP78, CHOP, and c-caspase12 in the cortex compared with WT mice (Fig. 3c). In primary cortical neurons obtained from embryos of WT and Tigar transgenic (TG) mice, and in LV-vector-HT22 (NC) and LV-TIGAR-HT22 (OV) cells, the levels of GRP78, c-caspase12, ATF4, and CHOP were significantly higher in response to OGD/R or tunicamycin (TM) compared with naive treatment, suggesting OGD/R-induced ER stress activates the ATF4-CHOP signaling pathway. The results also showed low and almost undetectable expression of CHOP in the naive treated control or TIGAR overexpression cells, consistent with previous studies indicating minimal CHOP mRNA and protein levels under nonstress conditions in neuronal cells (Poone et al., 2015; Southwood et al., 2016; Sun et al., 2018). However, OGD/R or TM treatment led to a significant increase in CHOP protein expression. Importantly, TIGAR overexpression reversed OGD/R- or TM-induced upregulation of GRP78, c-caspase12, ATF4, and CHOP proteins in TIGAR overexpression cells when compared with NC or WT neuronal cells (Fig. 3d,e). Moreover, TIGAR overexpression (OV) mitigated ER stress-induced apoptosis, as evidenced by improved cell viability and reduced apoptosis rates in LV-TIGAR-HT22 (OV) following TM treatment compared with LV-vector-HT22 (NC) cells (Fig. 3f–h). The above results indicate that TIGAR overexpression can alleviate ER stress-dependent apoptosis in neurons during cerebral ischemia/reperfusion.
TIGAR inhibits ER stress-induced neuronal apoptosis during cerebral ischemia/reperfusion. a, b, HT22 cells were infected with LV-TIGAR (OV) or LV-negative control (NC) for 72 h. The efficiency of TIGAR overexpression was evaluated with GFP fluorescence (a) and Western blotting (b). Scale bar, 100 μm. c, Wild-type (WT) and Tigar transgenic mice (TG) were subjected to tMCAO/R (I/R) and ipsilateral cortices were dissected 6 h after reperfusion. d, LV-vector-HT22 (NC) or LV-TIGAR-HT22 (OV) cells were treated with tunicamycin (TM, 12 μM) to induce ER stress and harvested at 12 h after treatment. The identification of TIGAR overexpression in HT22 cells is presented in Extended Data Figure 3-1. e, The primary neurons from WT or Tigar transgenic mouse embryos were subjected to OGD/R. c–e, Protein levels were measured with Western blotting. f, Cell viability was examined with Cell Counting Kit-8 assay. g, h, Apoptosis was analyzed using flow cytometry. Uncropped blots are shown in Extended Data Figure 3-1. n = 6 mice or three independent neuronal cell cultures with 2–3 replicates each. Error bar represents mean ± SD. Two-way analysis of ANOVA with Sidak's post hoc test.
Figure 3-1
The original uncropped blots of Figure 3. Download Figure 3-1, TIF file.
TIGAR interacts with ATF4 to inhibit ER stress during cerebral ischemia/reperfusion
During ER stress, the transcription factor ATF4 translocates to the nucleus to regulate downstream signals of ER stress (Gupta et al., 2016). We measured the protein levels of ATF4 in whole cell lysate (WCL), cytoplasm, and the nucleus using Western blotting. ATF4 (Fig. 4a,b) demonstrated a parallel increase with TIGAR (Fig. 4c,d) in both the nucleus and WCL under conditions of in vivo I/R and in vitro ER stress models. Immunofluorescence assay also showed elevated immunoreactivities of TIGAR and ATF4 in neuronal nucleus in both in vivo and in vitro settings (Fig. 4e–g; Extended Data Fig. 4-1). Notably, the colocalization of TIGAR and ATF4 within the neuronal nucleus was evident in response to I/R (Fig. 4e; Extended Data Fig. 4-1a,d). Similarly, there was a significant increase in the colocalization of TIGAR and ATF4 within the nucleus of TM-treated HT22 cells or OGD/R-treated primary neurons (Fig. 4f,g; Extended Data Fig. 4-1b,c,e,f). These data suggest that TIGAR and ATF4 proteins translocated to the nucleus and exhibit colocalization during ER stress induced by cerebral ischemia/reperfusion.
ATF4 nuclear translocation and colocalization with TIGAR are increased after cerebral ischemia/reperfusion or ER stress. a, c, Mice were subjected to tMCAO/R (I/R) and cortex were dissected 6 h after reperfusion. b, d, HT22 cells were treated with tunicamycin (TM, 12 μM) to induce ER stress and handled at 6 h after treatment. a–d, Nucleus, cytoplasm and whole cell lysate (WCL) were extracted and subjected to Western blot analysis. Uncropped blots are shown in Extended Data Figure 4-2. e, The brain sections were labeled with the anti-TIGAR antibody (red), anti-ATF4 antibody (green), DAPI (blue), and NeuN (cyan). The immunofluorescence was examined in the peri-infarct region of cortex. Scale bar, 20 μm. f, g, HT22 cells or primary cortical neurons were labeled with the anti-TIGAR antibody (red), anti-ATF4 antibody (green), and DAPI (blue). Scale bar, 10 μm. e–g, The original images of all channels are shown in Extended Data Figure 4-1. Insets were the enlarged cells displayed obvious nuclear translocation of ATF4 and colocalization with TIGAR. Pearson’s correlation coefficients were calculated from these images. The analysis was based on the green, red, and blue channels of single-cell images. N = 40 cells from three independent experiments. n = 3–6 mice or three independent neuronal cell cultures with 2–3 replicates each. The mean value for each mouse or independent experiment were statistically analyzed and presented the data in Extended Data Figure 4-1. Error bar represents mean ± SD. a–d, Two-way analysis of ANOVA with Sidak's post hoc test. e–g, Unpaired Student’s t test.
Figure 4-1
Neuronal ATF4 nuclear translocation and colocalization with TIGAR are increased after cerebral ischemia/reperfusion or ER stress. Panel a-c are the original magnified images of all channels in Figure. 4c, d and e, respectively. (a) Mice were subjected to tMCAO/R (I/R) and cortex were dissected 6 h after reperfusion. The brain sections were labeled with the anti-TIGAR antibody (red), anti-ATF4 antibody (green), DAPI (blue) and NeuN (cyan). Scale bar = 20 μm. (b) HT22 cells were treated with tunicamycin (TM, 12 μM) to induce ER stress and handled at 6 h after treatment. Cells were labeled with the anti-TIGAR antibody (red), anti-ATF4 antibody (green), and DAPI (blue). Scale bar = 10 μm. (c) Primary cortical neurons were subjected to oxygen glucose deprivation and reperfusion (OGD/R). The cells were labeled with the anti-TIGAR antibody (red), anti-ATF4 antibody (green) and DAPI (blue). Scale bar = 20 μm. (d-f) Pearson correlation coefficients were calculated from the images of Figure. 4c, d and e, respectively. The mean value for each mouse or independent experiment were statistically analyzed and presented in the graph. The analysis was based on the green, red and blue channels of single-cell images. N = 40 cells from three independent experiments. n = 3-6 mice or 3 independent neuronal cell cultures with 2-3 replicates each. Bar represents mean ± S.D. (d-f) Unpaired Student's t test. Download Figure 4-1, TIF file.
Figure 4-2
The original uncropped blots of Figure 4. Download Figure 4-2, TIF file.
The observed colocalization of TIGAR and ATF4 in the nucleus promoted us to hypothesize that TIGAR might interact with ATF4. To investigate the potential interaction between TIGAR and ATF4, co-immunoprecipitation (CoIP) and GST pull-down assays were employed. Notably, TIGAR and ATF4 co-immunoprecipitated in lysates from the ischemic cortex (Fig. 5a,b), as well as in HEK293T cells transfected with FLAG-TIGAR (WT) and ATF4 (Fig. 5c,d). Moreover, the GST pull-down assay using GST-TIGAR confirmed a specific interaction between ATF4 and GST-TIGAR in brain lysates (Fig. 5e).
TIGAR interacts with ATF4 during cerebral ischemia/reperfusion. a, b, Mice were subjected to tMCAO/R (I/R) and cortices were dissected 6 h after reperfusion. The brain lysates were immunoprecipitated with anti-ATF4 (a) or anti-TIGAR (b) antibodies. c, d, HEK293T cells were harvested at 24 h after transfected with p3×FLAG-TIGAR and ATF4. The cellular lysates were immunoprecipitated with anti-ATF4 (c) or anti-FLAG (d) antibodies. The immunoprecipitates, along with a portion of the lysates containing 60 µg of protein used as input, were analyzed by Western blotting. e, GST-TIGAR fusion protein or GST bound to glutathione-agarose were incubated with lysates of brain, and analyzed by Western blotting with anti-ATF4. a–e, Uncropped blots are shown in Extended Data Figure 5-1. f, g, Wild-type (WT) or TIGAR transgenic (TG) mice were subjected to tMCAO/R (I/R) and the cortices were dissected 6 h after reperfusion. h, i, LV-vector-HT22 (NC) or LV-TIGAR-HT22 (OV) cells were treated with TM (12 μM) to induce ER stress and harvested at 12 h after treatment. The mRNA levels of Atf4 and Chop were measured with RT-qPCR. j, ChIP assay was performed by using anti-FLAG antibody in LV-TIGAR-HEK293T cells. IgG served as a negative control. DNA enrichment of ChIP sample was measured with RT-qPCR using CHOP and ATF3 primers. n = 3–6 mice or three independent cell cultures with 2–3 replicates each. Two-way analysis of ANOVA with Tukey's post hoc test.
Figure 5-1
The original uncropped blots of Figure 5. Download Figure 5-1, TIF file.
Given ATF4's role as an active transcription factor modulating Chop and Atf3 expression during ER stress (Fusakio et al., 2016), we investigated the impact of TIGAR on Atf4 transcriptional activity by RT-qPCR. In WT mice subjected to cerebral ischemia/reperfusion, Atf4 and Chop mRNA expression in the cortex significantly increased. Remarkably, TG mice exhibited a decrease in Atf4 and Chop transcription compared with WT mice following I/R (Fig. 5f,g). Similarly, in NC cells, TM treatment significantly increased Atf4 and Chop mRNA levels, but TIGAR overexpression mitigated this increase (Fig. 5h,i). We also performed ChIP-RT-qPCR in LV-TIGAR-HEK293T cells using anti-FLAG-TIGAR antibody. The data revealed an interaction of FLAG-TIGAR with the DNA of Chop and Atf3 (Fig. 5j), suggesting that TIGAR may bind to ATF4's downstream genes Chop and Atf3 to regulate the transcription of these genes. These results collectively demonstrate that TIGAR exerts an inhibitory effect on ER stress-dependent apoptosis by binding to ATF4 to influence its transcriptional activity and downstream signaling pathways.
ATF4 overexpression reverses TIGAR-mediated attenuation of cerebral ischemia/reperfusion injury and ER stress-related apoptosis
To further determine whether TIGAR regulates ER stress through ATF4, lentivirus-mediated overexpression of ATF4 was applied in HT22 neuronal cells (Extended Data Fig. 6-1a). Consistent with the previous data, TIGAR overexpression decreased TM-induced upregulation of CHOP and c-caspase12, while ATF4 overexpression reversed TIGAR's effect on these proteins (Fig. 6a). Moreover, TIGAR overexpression significantly ameliorated TM-induced cell damage and increased cell viability, whereas ATF4 overexpression abolished TIGAR's protective effect against TM-induced cell injury (Fig. 6b). Parallel observations were made in primary neurons isolated from WT or TG fetal mice (Extended Data Fig. 6-1b). Primary neurons from TG mice showed reduced OGD/R-induced expression of CHOP and c-caspase12, but this reduction was reversed by ATF4 overexpression (Fig. 6c). Additionally, TG neurons exhibited significantly higher cell viability than the WT group after undergoing OGD/R. However, the protective effect was reversed by ATF4 overexpression (Fig. 6d).
Exogenous ATF4 negated the protective effect of neuronal TIGAR against cerebral ischemia/reperfusion-induced ER stress-dependent apoptosis. a, b, Control or LV-TIGAR stabled HT22 cells were infected with LV-ATF4 or LV-vector. Then the cells were treated with TM (12 μM). c, d, The primary neurons from WT or Tigar transgenic mouse embryos were transfected with LV-ATF4 or LV-vector. Then the cells were treated with OGD/R. a, c, Protein levels of CHOP and caspase12 were measured with Western blotting. b, d, The cell viability was examined with Cell Counting Kit-8 assay. e, The protocols of the in vivo experiment. f, The infract regions appeared white in TTC-stained sections after tMCAO/R. g, Neurological deficits were assessed by Longa Score Scale after tMCAO/R. h, Caspase12 and CHOP protein levels were measured with Western blotting after tMCAO/R. The expression of ATF4 in HT22 cells, primary neurons, and the brains of TIGAR transgenic mice after infection is presented in Extended Data Figure 6-1. Uncropped blots are shown in Extended Data Figure 6-2. n = 6–12 mice or three independent cell cultures with 2–3 replicates each. Error bar represents mean ± SD. One-way analysis of ANOVA with Tukey's post hoc test.
Figure 6-1
The expression of ATF4 in HT22 cells, primary neurons and the brain of TIGAR transgenic mice after infection. (a) HT22 cells or (b) primary neurons were infected with LV-ATF4 or LV-vector. The efficiency of ATF4 overexpression was evaluated with GFP fluorescence and Western blotting. Scale bar = 100μm (a) or 50 μm (b). (c) TIGAR transgenic mice (TG) were injected with AAV-VEC or AAV-ATF4 and sacrificed after 30 days. Representative images of GFP protein expression in the brain. H-C: hippocampus and cortex. S-C: striatum and cortex. BF: bright field. Scale bar = 100 µm. Brain lysates were assayed by western blotting. Uncropped blots are shown in Extended Data Figure 6-3. n = 6 mice or 3 independent neuronal cultures with 2-3 replicates each. Bar represents mean ± S.D. (a) Unpaired Student's t test. (b, c) Two-way analysis of ANOVA with Tukey's post hoc test. Download Figure 6-1, TIF file.
Figure 6-2
The original uncropped blots of Figure 6. Download Figure 6-2, TIF file.
Figure 6-3
The original uncropped blots of Figure 6-1. Download Figure 6-3, TIF file.
To confirm TIGAR mitigates cerebral ischemia/reperfusion-induced ER stress in vivo by targeting ATF4, we employed AAV9 to overexpress ATF4 in the brains of Tigar transgenic mice. The tMCAO model was performed 30 d after brain stereotaxic injection of AAV virus (Fig. 6e), following validation of AAV9-ATF4 infection efficiency (Extended Data Fig. 6-1c). After 24 h of tMCAO/reperfusion, TG mice exhibited significantly improved neurological deficits compared with WT mice injected with negative control virus. However, this amelioration was notably compromised in TG mice overexpressing ATF4, and ATF4 overexpression in WT mice exacerbated neurological deficits even beyond those observed in WT mice (Fig. 6g). TTC staining showed a significant reduction in cerebral infarct size in TG mice compared with WT mice, consistent with previous results. Importantly, the cerebral infarct area of TG mice overexpressing ATF4 did not exhibit the reduction observed in TG mice, resembling the infarct size in WT mice. Additionally, WT mice overexpressing ATF4 exhibited an even large infarct area (Fig. 6f). These results suggest that ATF4 overexpression exacerbated cerebral ischemia/reperfusion injury and counteracts TIGAR's protective effects. Consistent with earlier observations, immunoblotting showed that TIGAR overexpression inhibited protein expression of c-caspase12 and CHOP after cerebral ischemia–reperfusion. Notably, ATF4 overexpression partially reversed TIGAR's inhibitory effect on these proteins (Fig. 6h). These findings suggest that ATF4 overexpression reverses the protective effect of TIGAR on ER stress-induced apoptosis, further implying that TIGAR inhibits ER stress-induced neuronal injury by targeting ATF4 during cerebral ischemia/reperfusion.
TIGAR translocates to the nucleus to interact with ATF4 via Q141/K145 residues rather than its known phosphatase activity and mitochondrial localization domains
Previous studies have shown that TIGAR functions as a fructose-2,6-bisphosphatase, promoting the PPP. However, its mitochondrial localization is independent of phosphatase activity and relies on its C-terminal amino acids 258–261 (Cheung et al., 2012). To investigate the effect of TIGAR on inhibiting ATF4 in relation to its known functions, we applied a phosphatase functional region mutant (Δtm) and a mitochondrial localization mutant (Δ258-261) of FLAG-TIGAR, along with FLAG-WT-TIGAR. Firstly, we explored the effect of mutated TIGAR on the Atf4 transcriptional activity by luciferase reporter assay. As expected, the expression of WT-TIGAR reduced the luciferase activity of ATF4 in HEK293T cells treated with TM. Intriguingly, the luciferase activity of ATF4 remained significantly reduced in the cells transfected with the phosphatase functional region mutant (Δtm) or mitochondrial localization mutant (Δ258-261) of TIGAR (Extended Data Fig. 7-1a). These results imply that the inhibitory effect of TIGAR on ATF4 transcription may not depend on its phosphatase activity and mitochondrial localization domains.
Next, to identify the key domains in the TIGAR protein responsible for its translocation to the nucleus and interaction with ATF4, we conducted molecular docking and simulation study using the crystal structures of TIGAR (PDB:3dyc) and ATF4 (PDB:1ci6). The molecular prediction indicated that Gln141, Lys145, or TIGAR may form hydrogen bonds with residues of ATF4 (Extended Data Fig. 7-1b). In addition, PredictProtein (https://predictprotein.org) predicted that the 59-63 of TIGAR might be involved in RNA binding (Bernhofer et al., 2021). Accordingly, we thus constructed truncated and mutated TIGAR plasmids, including FLAG-TIGAR ΔN, FLAG-TIGAR ΔC, FLAG-TIGAR Δ59-63AAAAA, and FLAG-TIGAR Q141A/K145A (Fig. 7a). Consistent with previous data, WT- TIGAR translocated to the nucleus after OGD/R. Similarly, the Δtm, Δ258-261, ΔN, ΔC, and Δ59-63AAAAA mutants of FLAG-TIGAR also displayed significantly increased nuclear localization of FLAG fluorescence after OGD/R. However, the nuclear localization of Q141A/K145A mutant was markedly reduced compared with WT after OGD/R (Fig. 7b; Extended Data Fig. 7-1c).
TIGAR translocates to the nucleus and interact with ATF4 depending on its Q141/K145. a, Schematic diagram of TIGAR and its mutants. The predicted binding sites for TIGAR and ATF4 are presented in Extended Data Figure 7-1. b, HT22 cells were transfected with Flag-WT-TIGAR, Flag-TIGAR Δtm, Flag-TIGAR Δ258-261, FLAG-TIGAR ΔN, FLAG-TIGAR ΔC, FLAG-TIGAR Δ59-63AAAAA, or FLAG-TIGAR Q141A/K145A plasmids before being subjected to OGD/R. Scale bar, 5 μm. This image shows the nuclear translocation of Δtm, Δ258-261, and Q141A/K145A mutants. The complete images and associated statistical plots are presented in Extended Data Figure 7-1. c, HEK293T cells were transfected with Flag-WT-TIGAR, Flag-TIGAR Δtm, Flag-TIGAR Δ258-261, or FLAG-TIGAR Q141A/K145A plasmids before being treatment with TM. The cellular lysates were immunoprecipitated with anti-IgG or anti-Flag antibodies. Then total lysates and immunoprecipitates were analyzed by Western blotting with anti-Flag or anti-ATF4. The graph on the right shows the corresponding statistics. d, e, HT22 cells were subjected to OGD/R after transfection with relevant plasmids. d, The protein levels of CHOP and c-caspase 12 were measured by Western blotting. Uncropped blots are shown in Extended Data Figure 7-2. e, The cell viability was examined with Cell Counting Kit-8 assay. n = 3 independent neuronal cultures with three replicates each. Error bar represents mean ± SD. One-way analysis of ANOVA with Tukey's post hoc test.
Figure 7-1
TIGAR translocates to the nucleus depending on its Q141/K145 rather than on the phosphatase and mitochondrial localization structural domains. (a) Analysis of ATF4 luciferase activity in HEK293 T cells. HEK293 T cells were co-transfected with ATF4 5'-UTR and FLAG-WT-TIGAR, FLAG -TIGAR Δtm or FLAG -TIGAR Δ258-261 plasmids for 24 h. Then the cells were treated with TM to induce ER stress. (b) The predicted binding sites for TIGAR and ATF4. Gln141 of TIGAR interacts with Leu285 of ATF4, and Lys145 of TIGAR interacts with Leu281 of ATF4. The favorable residues are labeled in black. (c) The original images of all channels in Figure 7b. HT22 cells were transfected with Flag-WT-TIGAR, Flag-TIGAR Δtm, Flag-TIGAR Δ258-261 FLAG-TIGAR ΔC, FLAG-TIGAR ΔN, FLAG-TIGAR Δ59-63AAAAA or FLAG-TIGAR Q141A/K145A plasmids before subjected to OGD/R. Scale bar = 5 μm. The FLAG (green)-nucleus (blue) colocalization were quantified as Pearson’s correlation coefficients of images. The analysis was based on green and blue channels of single-cell images. N = 40 cells from three independent experiments. n = 3 independent cell cultures with 2-3 replicates each. Bar represents mean ± S.D. Two-way analysis of ANOVA with Tukey's post hoc test. Download Figure 7-1, TIF file.
Figure 7-2
The original uncropped blots of Figure 7. Download Figure 7-2, TIF file.
We further investigated the interaction between TIGAR mutants and ATF4. Consistent with previous results, WT-TIGAR co-immunoprecipitated with ATF4 and suppressed its expression. However, the CoIP of TIGAR and ATF4 was obviously decreased after TM-induced ER stress in the HEK293T cells transfected with Q141A/K145A mutant but not Δtm and Δ258-261 mutants (Fig. 7c). These results suggest that TIGAR's interaction with ATF4 depends on residues Q141/K145, rather than on the phosphatase activity and mitochondrial localization domains. Next, we examined the effects of TIGAR mutants on ER stress-induced apoptosis. WT-TIGAR reduced the upregulation of CHOP and c-caspase 12 induced by OGD/R, as well as increase cell viability, suggesting that TIGAR attenuated ER stress-dependent apoptosis (Fig. 7d,e). However, the Q141A/K145A mutant, but not the ΔN, ΔC, and Δ59-63AAAAA mutants, failed to reduce the expression of c-caspase 12 and CHOP and improve cell viability, underscoring the essential role of residues Q141/K145 for TIGAR in rescuing ER stress. Taken together, these findings indicate that TIGAR translocates to the nucleus and interacts with ATF4 to reduce ER stress, with residues Q141/K145 playing a critical role in this regulatory mechanism.
Discussion
The known function of TIGAR is to promote PPP via its Fru-2, 6-BPase activity, resulting in the production of NADPH. NADPH maintains the reduced GSH, thioredoxin [Trx (SH)2], and glutaredoxin, which contribute to scavenging ROS and preventing protein thiol oxidative damage (Tang et al., 2021; TeSlaa et al., 2023; P. Zhang et al., 2024). Previous studies, including our own, have demonstrated that the neuroprotection of TIGAR in cerebral ischemia and preconditioning is partly related to its antioxidative and anti-inflammatory effects (Li et al., 2014; Zhou et al., 2016; Chen et al., 2018; Zhang et al., 2019; Liu et al., 2022; Wang et al., 2024). This study builds on these findings and explores the role of TIGAR nonenzymatic functions in regulating ER stress and subsequent neuronal damage during ischemia–reperfusion. The results showed that TIGAR translocates to the nucleus and interacts with ATF4 to reduce the ER stress-dependent apoptosis, thus alleviating neuronal injury. This study revealed a new mechanism of TIGAR's neuroprotective action via regulating ATF4-mediated ER stress signaling.
Cerebral ischemia/reperfusion injury is a complex cascade involving oxidative stress, calcium overload, mitochondrial dysfunction, inflammatory reaction, and neuronal apoptosis (Wu et al., 2021; Y. Han et al., 2022; Sun et al., 2023; Zhou et al., 2023; H. Zhang et al., 2024). Recent studies have shown that ER stress is closely related with cerebral ischemia/reperfusion injury (Pan et al., 2021; Kumar et al., 2023). In the context of ER stress, UPR is activated due to the aggregation of unfolded and misfolded proteins in the ER, which is triggered through signal proteins such as PERK, IRE1α, and ATF6, to restore ER homeostasis (Wiseman et al., 2022). In the early stage of ER stress, the function of ER can be recovered, while enduring and severe ER stress may induce apoptosis (You et al., 2021; Mandula et al., 2022). The expression of CHOP began increasing 3 h after cerebral ischemia/reperfusion and reached its peak at 24 h, which then mediated neuronal apoptosis. In addition, caspase12 is also involved in ER stress-dependent apoptosis during cerebral ischemia (Zuo et al., 2018; Xu et al., 2021). Consistent with these reports, we found that ER stress-related proteins increased after I/R. Importantly, TIGAR and ER stress responses were simultaneously upregulated during I/R. Moreover, TigarNestin-CKO mice exhibit more pronounced neuronal ER swelling and protein aggregation in the ER lumen after I/R, in association with more severe cerebral ischemic injury, while this phenotype was significantly improved in Tigar TG mice. In addition, overexpression of TIGAR can reverse the upregulated levels of ER stress-related proteins and rescue ER stress-dependent apoptosis both in vivo and in vitro. All these data demonstrate that TIGAR can attenuate cerebral ischemic injury and inhibit ER stress-induced neuronal apoptosis during cerebral ischemia/reperfusion.
As previously reported, TIGAR is widely distributed in cells including mitochondria, ER, cytoplasm, and nucleus (Li et al., 2014). Intriguingly, this study found that neuronal TIGAR protein could translocate to the nucleus during I/R or ER stress. We extracted cellular components and found that the expression of neuronal TIGAR protein in the nucleus was significantly upregulated after I/R. Moreover, there is a distinct nuclear localization of TIGAR in neurons subjected to I/R or ER stress. We thus hypothesized that this translocation may underlie TIGAR's inhibitory effects on ER stress-dependent apoptosis.
ATF4, also known as cAMP response element binding protein 2 (CREB-2), is a member of ATF/CREB transcription factor family of basic leucine zipper domain proteins. ATF4 functions as an activated transcription factor that operates downstream of the eIF2α pathway and serves as a crucial regulator within the ER stress and integrated stress response (Kumar et al., 2023; Lu et al., 2024). During the UPR, eIF2α reduces global protein synthesis but selectively promotes ATF4 expression, which then translocates to the nucleus through its nuclear targeting KKLKK signal located within its basic region to regulate the expression of downstream genes such as Sestrin2, Asparagine synthetase (ASNS), LC3, Gadd34, Chop, Atf3, Tribbles Pseudokinase 3 (Trib3), etc. (Rzymski et al., 2010; Lee et al., 2018; Tian et al., 2020; D. Han et al., 2022; Xu et al., 2022; Ye et al., 2023; Zou et al., 2024). Under conditions of mild or transient stress, ATF4 initially regulates the expression of stress-adaptive genes and engages in antioxidant defense, autophagy, and amino acid metabolism, thereby playing a pivotal role in enhancing cellular survival under stress (Almanza et al., 2019; Shi et al., 2021). Specifically, ATF4 activates the antioxidant defense mechanisms of Sestrin2 to reduce reactive oxygen species (ROS) accumulation (Li et al., 2021). ATF4 also regulates genes involved in amino acid metabolism, such as ASNS, supporting protein synthesis and cellular survival during nutrient deprivation (Zou et al., 2024). Furthermore, ATF4 plays a role in autophagy regulation by promoting the expression of autophagy-related genes like LC3, facilitating the clearance of damaged proteins and organelles while maintaining cellular homeostasis (B'Chir et al., 2013; Vanhoutte et al., 2021). However, under extreme or prolonged stress, ATF4 can contribute to neuronal apoptosis by upregulating proapoptotic genes such as Chop (Chen et al., 2016; Demmings et al., 2021; Liu et al., 2021). In this context, ATF3 also amplifies apoptotic signals, thereby enhancing the detrimental effects of ATF4 under sustained stress (Tian et al., 2020; Kang et al., 2024). Additionally, TRIB3 has been shown to promote apoptosis via ATF4-CHOP pathways, further complicating the role of ATF4 in the balance between cell survival and death. Collectively, these findings highlight the dual nature of ATF4 as both a protector and a promoter of apoptosis, depending on the cellular context (Ohoka et al., 2005; Aime et al., 2015; Demmings et al., 2021).
In this study, we observed low expression of ATF4 protein in nonstress state but a significant increase after I/R or ER stress. The transcriptome analysis and Western blotting revealed activation of the ATF4-CHOP pathway following cerebral ischemia/reperfusion. Therefore, the present study focused on exploring the role of ATF4 in mediating neuronal apoptosis during cerebral ischemia–reperfusion. Importantly, nuclear extraction analysis demonstrated simultaneous nuclear translocation of TIGAR and ATF4. There was an obvious colocalization of TIGAR and ATF4 in the neuronal nucleus after I/R or ER stress. CoIP and GST pull-down assays further verified that TIGAR can interact with ATF4. ChIP assay also demonstrated that TIGAR may interact with the downstream genes of ATF4. Furthermore, by overexpressing lentivirus carrying ATF4 in neuronal cells and primary neurons cultured in vitro, as well as AAV9-mediated overexpression of ATF4 in Tigar transgenic mice in vivo, we demonstrated that ATF4 overexpression reversed the protective effects of TIGAR against cerebral ischemia/reperfusion injury. This highlights the central role of the functional relevance between TIGAR and ATF4 interaction in modulating ER stress and neuronal apoptosis. These results indicate that TIGAR may interact with ATF4 to inhibit its downstream gene expression, thereby alleviating ER stress-dependent apoptosis during cerebral ischemia.
In addition to maintaining metabolic homeostasis of cells by Fru-2, 6-BPase activity, TIGAR can also regulate some signaling proteins through nonenzymatic action. Specifically, TIGAR can translocate to mitochondria to interact with HK2, thus reducing mitochondrial ROS and promoting cell survival during hypoxia (Cheung et al., 2012). And the mitochondrial localization of TIGAR is independent of the phosphatase activity but is dependent on its C-terminal amino acids 258–261 (Cheung et al., 2012). Interestingly, we found that TIGAR translocated to the nucleus independently of its phosphatase activity and mitochondrial localization domains.
We next predicted through molecular docking simulation that TIGAR protein may bind to ATF4 via its Q141/K145 residues. These amino acids do not belong to phosphatase activity domain and two conserved pockets that can mediate dephosphorylation by binding to phosphate molecules on the substrate (Tang et al., 2021), suggesting that TIGAR may inhibit ER stress and promote cell survival through a nonenzymatic function. Importantly, the TIGAR Q141A/K145A mutation, instead of ΔN, ΔC, and Δ59-63AAAAA mutations, abolished TIGAR's translocation to the ER and nucleus and the interaction with ATF4. These results indicate that the Q141/K145 residues of TIGAR protein is required for its translocation to ER and nucleus and interaction with ATF4. Previous studies have shown that ATF4 protein enters the nucleus to bind and regulate the expression of downstream target genes through its DNA binding region (280–299 aa), while malfunction of its DNA-binding region blocks the downstream expression of ATF4 (Wortel et al., 2017). Intriguingly, molecular docking data also showed that the binding to Q14/K145 of TIGAR was L281/L285 belonging to the DNA-binding domain of ATF4. This may further suggest that TIGAR binds to ATF4 to inhibit its downstream target gene expression.
This study reveals that TIGAR can effectively inhibit the transcriptional activity of ATF4 following I/R or ER stress, independent of its phosphatase activity or mitochondrial localization; however, the underlying mechanism remains elusive. As a transcription factor, ATF4 mRNA levels are elevated under conditions such as integrated stress response, ER stress, and oxidative stress (Liu, 2020; Lu et al., 2024). TIGAR exerts its antioxidative effects through multiple mechanisms such as promoting NADPH wproduction via the PPP and inhibiting succinate dehydrogenase (SDH) activity to mitigate mitochondrial ROS generation (Li et al., 2014; Wang et al., 2024). Notably, TIGAR activates Nrf2 in a PPP-independent manner to maintain redox homeostasis (H. Wang et al., 2022; Liu et al., 2022). Recent studies have identified Nrf2 as a potential regulator involved in ferroptosis and ER stress under the context of chronic intermittent hypoxia (CIH). Interestingly, inhibition of ferroptosis by Nrf2 leads to downregulation of ATF4 and CHOP expression, thereby attenuating ER stress (Zhong et al., 2024). These findings prompt us to speculate that TIGAR may regulate ATF4 transcription through the activation of Nrf2. Further investigations are warranted to elucidate the mechanisms underlying the regulatory role of TIGAR in reducing ATF4 transcription.
In conclusion, this study found that neuronal TIGAR translocates to the nucleus and interact with ATF4 following ER stress induced by cerebral ischemia–reperfusion, mediated by its Q141/K145 residues. Subsequently, TIGAR suppresses the expression proapoptotic genes downstream of ATF4, thereby attenuating ER stress-dependent apoptosis and alleviating neuronal damage (Fig. 8). This study expands on the mechanism of TIGAR's neuroprotective effect in cerebral ischemia/reperfusion injury. The Q141/K145 residues are critical for the interaction between TIGAR and ATF4, as well as for the inhibition of ATF4 target genes.
Proposed pathway underlying TIGAR involves translocation to the nucleus and interaction with ATF4, inhibiting ER stress-dependent neuronal apoptosis during cerebral ischemia/reperfusion. Neuronal overexpression of TIAGR ameliorates ischemic injury. Neuronal TIGAR translocates to the nucleus and interacts with ATF4 following ER stress induced by cerebral ischemia–reperfusion, mediated by its Q141/K145 residues. TIGAR suppresses the expression of proapoptotic genes downstream of ATF4, thereby attenuating ER stress-dependent apoptosis and alleviating neuronal damage.
Data Availability
The data that support the findings of this study are available from the corresponding author upon reasonable request. Some data may not be made available because of privacy or ethical restrictions.
Footnotes
This work was supported by grants from the National Natural Science Foundation of China (No. 82473917, 82173811), Jiangsu Key Laboratory of Neuropsychiatric Diseases (BM2013003) and the Priority Academic Program Development of the Jiangsu Higher Education Institutes (PAPD).
↵*L.C. and J.T. contributed equally to this work.
The authors declare no competing financial interests.
- Correspondence should be addressed to Rui Sheng at sheng_rui{at}163.com.
