TXNIP/NLRP3 aggravates global cerebral ischemia-reperfusion injury-induced cognitive decline in mice

Abstract

Global cerebral ischemia/reperfusion (GCI/R) injury poses a risk for cognitive decline, with neuroinflammation considered pivotal in this process. This study aimed to unravel the molecular mechanisms underlying GCI/R injury and propose a potential therapeutic strategy for associated cognitive deficits. Utilizing bioinformatics analysis of a public microarray profile (GSE30655 and GSE80681) in cerebral ischemic mice, it was observed that neuroinflammation emerged as a significant gene ontology item, with an increase in the expression of thioredoxin-interacting protein (TXNIP) and NLRP3 genes. Experimental models involving bilateral occlusion of the common carotid arteries in mice revealed that GCI/R induced cognitive impairment, along with a time-dependent increase in TXNIP and NLRP3 levels. Notably, TXNIP knockdown alleviated cognitive dysfunction in mice. Furthermore, the introduction of adeno-associated virus injection with TXNIP knockdown reduced the number of activated microglia, apoptosis neurons, and levels of oxidative stress and inflammatory cytokines in the hippocampus. Collectively, these findings underscore the significance of TXNIP/NLRP3 in the hippocampus in exacerbating cognitive decline due to GCI/R injury, suggesting that TXNIP knockdown holds promise as a therapeutic strategy.

Graphical abstract

In brief: Using bioinformatics analysis based on GSE30655 and GSE80681 datasets revealed an increase in the expression of genes TXNIP and NLRP3 in mice experiencing ischemia/reperfusion injury. The introduction of TXNIP knockdown via adeno-associated virus injection demonstrated a mitigation of neuroinflammation and cognitive decline induced by global cerebral ischemia/reperfusion injury. This amelioration was associated with microglial activation, mediated by the inhibition of the TXNIP/NLRP3 signaling pathway. Activations were denoted by “+”, inhibitions by “–”, increases by “↑”, and decreases by “↓”.

1. Introduction

Global cerebral ischemia-reperfusion (GCI/R) injury is a severe clinical condition frequently observed during events such as cardiac arrest, hemorrhagic shock, severe trauma, cardiopulmonary bypass, carotid endarterectomy, and neurosurgery [1,2]. This type of injury can result in learning and memory loss and may even lead to patient deaths in the days and months following reperfusion [3,4]. Furthermore, early cognitive impairment induced by GCI/R injury could potentially elevate the morbidity of vascular dementia and Alzheimer's disease [5,6]. Neuroinflammation was considered to play a crucial role in the GCI/R-induced cognitive deficits [7,8]. However, the underlying molecular mechanism and potential therapeutic strategy need further elucidation.

Microglia, the natural immune cells in the brain and spinal cord, establish an inflammatory microenvironment [9,10]. Following ischemic injury, microglia can be activated and accumulate in the hippocampus. While moderately activated microglia can assist in clearing cellular debris and repairing brain tissue, prolonged activation leads them to release pro-inflammatory factors, thereby causing neuronal cell death [11,12]. Therefore, regulating the excessive activation of microglia might offer therapeutic benefits against neuroinflammation caused by GCI/R injury.

Thioredoxin-interacting protein (TXNIP), also known as vitamin D3 upregulated protein-1, is a pivotal mediator in the thioredoxin (TRX) antioxidant system, crucial for the redox signaling complex and relevant to the pathophysiology of certain diseases [13]. Recent studies indicate that elevated TXNIP levels contribute to the pathogenesis of complex diseases such as metabolic disorders, neurological disorders, and inflammatory illnesses [14,15]. The assembly of the nucleotide-binding oligomerization domain-like receptor protein 3 (NLRP3) inflammasome, comprising NLRP3, apoptosis-associated speckle-like protein, and caspase-1, could increase the risk of cognitive impairment, involving immune cell activation, cell apoptosis, oxidative stress, mitochondrial autophagy, and neuroinflammation [16]. Under pathological conditions, activation of the NLRP3 inflammasome can trigger the transformation of procaspase-1 to caspase-1, resulting in the production of mature interleukin 1β (IL-1β) and IL-18. Furthermore, increased inflammatory cytokines can promote immune cell activation and infiltration, resulting in an inflammatory cascade, oxidative stress, and cell apoptosis [17]. Meanwhile, the cellular powerhouse mitochondria are sensitive to hypoxia, and mitochondria are inclined to be impaired and produce reactive oxygen species (ROS) during I/R injury. When mitochondrial autophagy is insufficient to remove damaged mitochondria, the increased ROS could induce TXNIP dissociation from TRX, binding to the NLRP3 inflammasome—an indispensable step for activation. Additionally, research indicates that the NLRP3 inflammasome and downstream IL-1β activate microglial responses, establishing a feedback loop of neuroinflammation [18,19]. Consequently, the TXNIP/NLRP3 signaling pathway likely plays a pivotal role in regulating microglia activation.

In this study, a pilot bioinformatics analysis based on a public microarray profile (GSE30655 and GSE80681) of cerebral ischemic mice was conducted. The potential targets were then validated in a mice model of bilateral common carotid artery occlusion (BCCAO). Adeno-associated virus (AAV)-carrying shRNA was used to investigate the underlying mechanism.

2. Materials and methods

2.1. Gene expression omnibus (GEO) data sets analysis

Gene expression profiles related to cerebral I/R injury were retrieved and downloaded from the GEO database (http://www.ncbi.nlm.nih.gov/geo). The criteria for selecting datasets included focal I/R injury from middle cerebral artery occlusion (MCAO) mice or global I/R injury from BCCAO mice. Two expression profile microarray datasets (GSE30655 and GSE80681) were chosen for analysis [20,21]. Cerebral cortices were collected at 24 h reperfusion after 60-min middle cerebral artery occlusion in the GSE30655, while the hippocampal CA1 subfield samples were collected at 3 h reperfusion after 10-min bilateral common carotid artery occlusion In the GSE80681.In the analysis, samples from each dataset were categorized into two groups: the sham surgery group and the ischemia group. Differentially expressed genes (DEGs) were identified using the R software "Limma" package. The screening criteria for DEGs were |log2 FC| > 1 and an adjusted P-value (adj.P) < 0.05. Subsequently, the Venn diagram tool (http://bioinformatics.psb.ugent.be/webtools/Venn/) was used to compare and analyze the intersection of DEGs from different datasets. The intersection DEGs were utilized for Gene Ontology (GO) enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses via the Database for Annotation, Visualization, and Integrated Discovery database [22]. Finally, gene set enrichment analysis (GSEA) was conducted using the online platform (https://www.bioinformatics.com.cn) [23] to assess the statistical significance of the inflammasomes and NLR signaling pathway gene sets (https://www.gsea-msigdb.org/) in the GSE30655 and GSE80681 datasets.

2.2. Animals and GCI/R model

The study received approval from the Animal Care Committee at Southwest Medical University in China. Adult male C57BL/6 mice, weighing between 18 and 22 g, were obtained from Chengdu Dossy Experimental Animals in China. These mice were housed in a room maintained at a consistent temperature (22 ± 2 °C) under a 12-h light/dark regime, with a one-week acclimation before surgery. All experimental procedures strictly adhered to the guidelines set by the National Institutes of Health for the use of experimental animals.

Subsequently, mice were randomly assigned to either the sham or the I/R groups. Using a previously described modeling method with slight modifications [24], mice were intraperitoneally anesthetized with 2% pentobarbital sodium (45 mg/kg). The bilateral common carotid arteries (CCAs) were exposed through a midline neck incision. In the I/R group, the CCAs were temporarily obstructed for 30 min using microvascular clamps. In the sham group, the CCAs were dissected without any occlusion. Laser Speckle Contrast Imaging (LSCI) was employed during modeling to monitor changes in cerebral blood flow (CBF) [25,26], confirming the effective establishment of the GCI/R model. A heating pad was utilized to maintain the mice's body temperature throughout the process.

2.3. Animal experimental design

The animal experimental flow diagram is illustrated in Fig. 1.

Fig. 1.

Animal experimental design and animal groups. AAV, adeno-associated virus; ICV, intracerebroventricular; WB, Western blotting; qRT-PCR, quantitative real-time polymerase chain reaction; ELISA, enzyme-linked immunosorbent assay; BWC, brain water content; HE, hematoxylin-eosin staining; IF, immunofluorescence; TUNEL, terminal deoxynucleotidyl transferase dUTP nick end labeling; MWM, Morris water maze.

2.3.1. Experiment 1

To evaluate the time course of endogenous TXNIP, NLRP3, and cleaved caspase-1 temporal levels in the hippocampus of mice following GCI/R, the mice were randomly divided into four groups: sham, and 12, 24, and 72 h after I/R (n = 6/group). Western blotting analysis was conducted to assess the protein levels at the various time points after GCI/R.

2.3.2. Experiment 2

To screen the most effective interference fragment of TXNIP, the mice were randomly divided into five groups: negative control (NC), control shRNA, TXNIP shRNA1, TXNIP shRNA2, and TXNIP shRNA3 (n = 6/group). A 5 μL AAV vector carrying control shRNA, TXNIP shRNA1, TXNIP shRNA2, or TXNIP shRNA3 was injected into the lateral cerebral ventricle of the mice in each respective group. The shRNA sequences are listed as follows: control shRNA (5′-CCTAAGGTTA AGTCGCCCTC GCTCGAGCGA GGGCGACTTA ACCTTAGG-3′), TXNIP shRNA1 (5′-GCTGGATAGA CCTAAACATC TCGAGATGTT TAGGTCTATC CAGC-3′), TXNIP shRNA2 (5′-GCAAACAGAC TTTGGACTAC TCGAGTAGTC CAAAGTCTGT TTGC-3′), and TXNIP shRNA3 (5′-GCCTCAGAGT GCAGAAGATC TCGAGATCTT CTGCACTCTG AGGC-3′). After 28 days, each group of mice was sacrificed, and the hippocampus was collected for Western blotting and quantitative real-time polymerase chain reaction (qRT-PCR).

2.3.3. Experiment 3

To investigate the effect of TXNIP knockdown on hippocampal damage and cognitive decline through the NLRP3/caspase-1 pathway in mice after GCI/R, the mice were randomly divided into five groups: sham, I/R, sham + control shRNA, I/R + control shRNA, and I/R + TXNIP shRNA (n = 30/group). Before establishing the sham or I/R groups, a volume of 5 μL AAV-containing control shRNA or TXNIP shRNA was intracerebroventricularly administered 28 days prior. early. In total, 18 mice in each group were sacrificed 3 days after modeling for Western blotting, enzyme-linked immunosorbent assay (ELISA), brain water content (BWC) measurement, hematoxylin-eosin (HE) staining, ionized calcium-binding adapter molecule 1 (Iba1) immunofluorescence staining, and terminal deoxynucleotidyl transferase dUTP nick end-labeling (TUNEL) staining. The remaining 12 mice in each group underwent the Morris water maze (MWM) test on days 7–12 after modeling.

2.4. Intracerebroventricular administration

The intracerebroventricular administration was conducted 28 days prior to GCI/R, following a previously described protocol [27]. The AAV vectors were produced by Yunzhou Biosciences (Guangzhou, China) and included AAV-control shRNA, AAV-TXNIP shRNA1, AAV-TXNIP shRNA2, and AAV-TXNIP shRNA3. Mice, anesthetized with pentobarbital sodium, were securely positioned within a stereotaxic apparatus. A volume of 5 μL AAV vectors was precisely administered into both ventricles (AP –0.3 mm; ML ±1.0 mm; DV 3.0 mm) at a rate of 0.5 μL/min. Following the injection, the needle remained in place for an additional 5 min before withdrawal. The incision was then sutured, and each mouse, upon regaining consciousness, was housed individually.

2.5. MWM test

To assess hippocampus-dependent spatial learning and memory in mice, the MWM test was conducted between days 7–12 post-GCI/R, following the methodology of previous studies [28,29]. The apparatus consisted of a circular pool, 50 cm in height and 120 cm in width, filled up to 30 cm with a milk-water mixture maintained at 22 ± 1 °C. The pool was divided into four equidistant quadrants. During the training phase, spanning 5 days (from day 1 to day 5), mice were placed into different starting quadrants and tasked with locating a submerged platform with a 10 cm diameter, positioned 1 cm below the water's surface. If a mouse failed to find the platform within 60 s, it was manually guided to the platform and allowed to stay there for 15 s. The time taken to discover the concealed platform, referred to as escape latency, was recorded. On the sixth day, a probe trial was conducted after removing the platform. The time spent in the target quadrant, the number of platform crossings, and swimming velocity were recorded using a digital tracking system (XINRUAN Information Technologies, Shanghai, China).

2.6. BWC measurement

The wet-to-dry brain weight ratio was measured to determine BWC, following the methodology of previous studies [30]. Upon euthanasia, the entire brain was promptly excised to measure the wet weight. Subsequently, the brain samples underwent desiccation at 65 °C for 48 h to determine the dry weight. The BWC was calculated using the formula: BWC = (wet weight − dry weight)/wet weight × 100%.

2.7. HE staining

The euthanized mice underwent transcardial perfusion with 4% paraformaldehyde, after which the brains were extracted, fixed, and paraffin-embedded. Subsequently, the brain samples were sectioned into 3 μm coronal slices for HE staining [31]. In each mouse, the hippocampal CA1 region of each section was examined under a light microscope (Olympus, Tokyo, Japan). The number of neuron cells/mm² in the CA1 subregion at 400 × magnification for each slide was quantified using Image-Pro Plus v6.0. The resulting average was then used in subsequent statistical analyses.

2.8. TUNEL staining

Brain specimens underwent sucrose dehydration before being sectioned into 8-μm-thick coronal frozen slices. To evaluate hippocampal cell apoptosis, TUNEL staining was employed following the manufacturer's guidelines (Roche, Basel, Switzerland) [32]. These sections were then incubated with a mixture of reagents (terminal deoxynucleotidyl transferase enzyme reaction solution blended with tetramethylrhodamine labeling solution in a 1:9 ratio) in a dark setting for 1 h. Subsequently, nuclei were stained using 4′,6-diamidino-2-phenylindole (DAPI) for 5 min. The stained sections were examined using a fluorescence microscope (Olympus, Tokyo, Japan). The average count of TUNEL-positive cells within the hippocampal CA1 subregion was calculated from four arbitrary fields at 400 × magnification. The data were presented as the ratio of TUNEL-positive cells (%).

2.9. Immunofluorescence staining

Immunofluorescence staining was performed following established protocols as described previously [33]. Briefly, 8-μm-thick coronal sections were fixed in paraformaldehyde for 30 min, permeabilized using 0.2% Triton X-100 for 15 min, blocked with 1% bovine serum albumin for 30 min, and then incubated with the primary antibody rabbit anti-Iba1 (1:100; Cell Signaling Technology, Danvers, MA, USA) overnight at 4 °C. After rinsing, the sections were incubated with the FITC-labeled goat anti-rabbit secondary antibody (1:200; Affinity, Cincinnati, OH, USA) in a dark setting at 37 °C for 1 h. Finally, the sections were stained with DAPI for 5 min. Three sections from each mouse were observed under a fluorescence microscope (Olympus, Tokyo, Japan). The Iba-1 positive cells/mm² in the CA1 area of the hippocampus were quantified using Image-Pro Plus v6.0.

2.10. Inflammatory cytokines and oxidative stress

Hippocampal homogenates were centrifuged at 12,000 rpm at 4 °C for 15 min or at 3000 rpm at 4 °C for 10 min, respectively. The concentrations of TNF-α, IL-6, IL-1β, and IL-18 in the hippocampus were then determined using commercial ELISA kits (Meimian Technology, Jiangsu, China) according to the manufacturer's guidelines [34]. The malondialdehyde (MDA) content and superoxide dismutase (SOD) activity were assessed using commercial biochemical kits (Nanjing Jiancheng, China), following previously described methodologies.

2.11. qRT-PCR

Following a standard protocol, total RNA was isolated from hippocampal samples using the RNA simple total RNA kit (Tiangen, Beijing, China). The extracted RNA was then converted into circular DNA (cDNA) with the ReverTra Ace qPCR RT Master Mix (Toyobo, Osaka, Japan). Subsequently, this cDNA was amplified using the SuperReal PreMix Plus (SYBR Green; Tiangen, Beijing, China) in a Roche qRT-PCR apparatus. The amplification settings included an initial denaturation phase at 95 °C for 15 min, followed by five cycles of 95 °C for 10 s and then 60 °C for 32 s. β-actin was used as an internal reference for normalization. For the naive group, the mRNA level, determined by the ratio of target mRNA to β-actin, served as the baseline to evaluate the fold variations in mRNA levels of the other test groups. The primer sequences were defined as follows: TXNIP (forward: 5′- ATACTCCTTGCTGATCTACG-3′, reverse: 5′-TGGGGTATCTGGG-ATGTTTA-3′) and β-actin (forward: 5′- TTTGCAGCTCCTTCGTTGC-3′, reverse: 5′-TCGTCA TCCATGGCGAACT-3′).

2.12. Western blot analysis

Protein concentration was determined using the enhanced BCA protein assay kit (Beyotime, Shanghai, China). Subsequently, 30 μg protein underwent separation through sodium dodecyl sulfate-polyacrylamide gel electrophoresis. It was then transferred to either a polyvinylidene fluoride (Amersham Biosciences, Piscataway, NJ, USA) or a nitrocellulose sheet (Beyotime). The membrane was treated with 5% skim milk for 1 h at 25 °C, followed by overnight incubation at 4 °C using primary antibodies: anti-TXNIP (1:1000; Cell Signaling Technology), anti-NLRP3 (1:1000; Cell Signaling Technology), anti-cleaved caspase-1 (1:200; Santa Cruz Biotechnology, Santa Cruz, CA, USA), and anti-β-actin (1:5000; Proteintech, Wuhan, China). Afterwards, the membrane was exposed to the appropriate horseradish peroxidase-linked secondary antibody (1:5000; Solarbio, Beijing, China) for 60 min at 25 °C. The bands corresponding to the antibodies were detected with an ECL solution (Amersham, Buckinghamshire, UK). For measurement purposes, bands that were not saturated were selected, and their intensities were assessed using Image-Pro Plus v6.0.

2.13. Statistical analysis

Statistical analyses were conducted using R v4.2.1 and GraphPad Prism v8.3. Differences between multiple groups were analyzed by the Kruskal-Wallis rank sum test or one-way analysis of variance followed by Tukey's post-hoc test. A P-value of less than 0.05 was considered statistically significant.

3. Results

3.1. TXNIP/NLRP3 overexpression is associated with an inflammatory response in ischemic mice

A total of 10 samples were included from dataset GSE30655, comprising seven ischemic group mice and three sham group mice. This analysis identified 511 DEGs, with 385 upregulated genes and 126 downregulated genes (Fig. 2A). Simultaneously, dataset GSE80681 included six samples, with three mice in the ischemic group and three in the sham group, revealing 415 DEGs, including 328 upregulated genes and 87 downregulated genes (Fig. 2B). The Venn diagram tool illustrated 98 intersection DEGs between these two datasets, among them the gene TXNIP (Fig. 2C). Subsequently, we obtained the TXNIP/NLRP3 pathway-related gene sets, including inflammasome and NLR signaling pathway gene sets, from the MSigDB database.

Fig. 2.

Bioinformatics analysis for the gene expression profile in GSE30655 and GSE80681 datasets. (A–B) Volcano plot of the DEGs of GSE30655 and GSE80681 datasets between ischemia and sham groups. (C) Venn plot of intersection DEGs between GSE30655 and GSE80681 datasets. (D–E) GSEA analyses of gene sets for inflammasomes and NLR signaling pathways in GSE30655 and GSE80681 datasets. NES, normalized enrichment score. (F) Histogram of GO and KEGG enrichment analyses.

GSEA analysis conducted on dataset GSE30655 demonstrated an increase in inflammasomes (normalized enrichment score [NES] = 1.66, P = 0.02) and NLR signaling pathway-related genes (NES = 1.65, P = 0.01) in the ischemia group mice compared with the sham group mice (Fig. 2D). Furthermore, the results indicated that the TXNIP/NLRP3 genes belong to the leading-edge subset. However, GSEA analyses in dataset GSE80681 (Fig. 2E) did not reveal a significant difference (P > 0.05). Additionally, GO and KEGG pathway analyses suggested that an inflammatory response may play an indispensable role in cerebral I/R injury (Fig. 2F).

Taken together, the TXNIP/NLRP3 genes, connecting inflammatory and oxidative stress, may play key roles in cerebral I/R injury, particularly in focal ischemia. This study further investigates whether they can regulate inflammation in GCI/R injury.

3.2. TXNIP/NLRP3 expression increase in the hippocampus of GCI/R mice

The LSCI analysis revealed a significant decrease in CBF in the I/R group compared with the sham group, indicating the successful establishment of the GCI/R model (Fig. 3A). Subsequent Western blot analysis demonstrated elevated protein levels of TXNIP, NLRP3, and cleaved caspase-1 at 12, 24, and 72 h post-GCI/R, with the expression peak observed at 72 h post-GCI/R (Fig. 3B–E). These results suggest that GCI/R can induce the activation of the TXNIP/NLRP3/cleaved caspase-1 pathway.

Fig. 3.

Time course of the endogenous protein levels of TXNIP, NLRP3, and cleaved caspase-1 in the hippocampus after GCI/R. (A) Laser speckle contrast imaging of cerebral blood flow, the yellow arrows indicate common carotid arteries, the green arrows indicate vagus nerve and the black arrows indicate microvascular clamps. (B) Representative bands of Western blot data. (C–E) Quantitative analysis of the Western blot bands. The values are shown as the mean ± SD. *P < 0.05, **P < 0.01 vs. Sham group (n = 6).

3.3. TXNIP knockdown ameliorates pathological damage and brain edema caused by GCI/R

At 28 days after intracerebroventricular administration of AAV vectors containing either control shRNA or TXNIP shRNA, the efficacy of the TXNIP interference fragments was assessed through qRT-PCR and Western blot analyses. The results revealed a significant reduction in hippocampal TXNIP mRNA and TXNIP protein levels in the TXNIP shRNA1, TXNIP shRNA2, and TXNIP shRNA3 groups compared with the NC and control shRNA groups. Notably, the TXNIP shRNA2 group exhibited the most pronounced suppression (Fig. 4A–C). Consequently, the TXNIP shRNA2 sequence was chosen for subsequent studies.

Fig. 4.

TXNIP knockdown ameliorated pathological damage and brain edema caused by GCI/R. (A–C) TXNIP knockdown efficacy evaluation. The values are shown as the mean ± SD. *P < 0.05, **P < 0.01 vs. NC group; ^#P < 0.05, ^##P < 0.01 vs. Control shRNA group (n = 6). (D) HE staining of hippocampal CA1 area ( × 400), scale bar = 20 μm. (E) Quantitative analysis of normal neurons in the hippocampal CA1 area. (F) BWC. The values are shown as the mean ± SD. *P < 0.05, **P < 0.01 vs. Sham group; ^#P < 0.05, ^##P < 0.01 vs. I/R group (n = 6).

To evaluate the protective effects of TXNIP knockdown against brain injury 72 h post-GCI/R, HE staining and BWC assessments were conducted (Fig. 4D–F). The results depicted the typical structural characteristics of the hippocampal CA1 zone in the sham and sham + control shRNA mice. Conversely, the I/R and I/R + control shRNA groups exhibited significantly shrunken and darkly stained damaged neurons post-GCI/R. Furthermore, fewer normal neurons in the hippocampal CA1 area and increased BWC were observed in the I/R and I/R + control shRNA groups compared with the sham group at 72 h after GCI/R. However, in the I/R + TXNIP shRNA group, the hippocampal CA1 normal neuron count increased, and BWC decreased compared with the I/R group. These results indicate that TXNIP knockdown mitigated pathological damage and brain edema caused by GCI/R.

3.4. TXNIP knockdown mitigates cognitive decline caused by GCI/R

The potential neuroprotective role of TXNIP knockdown in alleviating spatial reference learning and memory deficits induced by GCI/R was investigated using the MWM test. In the training phase (Fig. 5A), a noticeable increase in escape latency was observed in both the I/R and I/R + control shRNA groups compared to the sham group. This increase suggests that GCI/R may lead to spatial reference learning dysfunction, but this dysfunction appeared to be mitigated with TXNIP knockdown. During the probe trial (Fig. 5B–D), both the time spent in the target quadrant and the number of platform crossings significantly decreased in the I/R and I/R + control shRNA groups compared to the sham group, indicating potential memory impairments caused by GCI/R. However, these impairments were seemingly lessened with TXNIP knockdown. Moreover, no noticeable variation in swimming velocities was observed among all the groups, indicating that GCI/R did not adversely affect the mice's motor skills during this study.

Fig. 5.

TXNIP knockdown mitigated cognitive decline caused by GCI/R. (A) Escape latency of mice in the training trials. (B) The time in the target quadrant in the probe trial. (C) The number of platform crossings in the probe trial. (D) Swimming speed in the probe trial. The values are shown as the mean ± SD. *P < 0.05, **P < 0.01 vs. Sham group; ^#P < 0.05, ^##P < 0.01 vs. I/R group (n = 12). (E–F) Quantitative analysis of the inflammatory factors including TNF-α, IL-6, IL-1β, and IL-18. (G–H) Quantitative analysis of the oxidative stress markers (MDA and SOD). The values are shown as the mean ± SD. *P < 0.05, **P < 0.01 vs. Sham group; ^#P < 0.05, ^##P < 0.01 vs. I/R group (n = 6).

3.5. TXNIP knockdown reduces inflammatory response and oxidative stress in the hippocampus caused by GCI/R

Our investigation extended to assess the levels of inflammatory cytokines, specifically TNF-α, IL-6, IL-1β, and IL-18. Additionally, we examined oxidative stress through MDA content and SOD activity in the hippocampal region of the mice 72 h post-GCI/R (Fig. 5E–J). A marked increase in both inflammatory cytokines and indicators of oxidative stress was observed in the I/R and I/R + control shRNA groups compared to the sham group. However, this increase was notably diminished in the I/R + TXNIP shRNA group relative to the I/R group.

3.6. TXNIP knockdown inhibits hippocampal neuronal apoptosis and microglial activation caused by GCI/R

Apoptosis in the CA1 region of the mouse hippocampus was assessed using the TUNEL assay 72 h post-GCI/R (Fig. 6A–C). Negligible apoptotic cells were observed in both the sham and sham + control shRNA groups. In contrast, there was a marked increase in the number of apoptotic cells within the hippocampal CA1 region in the I/R and I/R + control shRNA groups compared to the sham group. However, the intervention with TXNIP knockdown notably mitigated these apoptotic changes, as evidenced by a significant reduction in TUNEL-positive cells in the I/R + TXNIP shRNA group relative to the I/R group.

Fig. 6.

TXNIP knockdown inhibits hippocampal neuronal apoptosis and microglial activation caused by GCI/R. (A) Representative images of TUNEL staining in the hippocampal CA1 area ( × 400); nuclei were labeled with blue fluorescence (DAPI), and TUNEL-positive cells were labeled with red fluorescence, scale bars = 20 μm. (B) Representative images of Iba-1 immunofluorescence staining in hippocampal CA1 area ( × 400); Iba-1 was labeled with green fluorescence, scale bars = 20 μm. (C) Quantitative analyses of TUNEL-positive cells in the hippocampal area. (D) Quantitative analyses of Iba-1-positive cells in the hippocampal area. The values are shown as the mean ± SD. *P < 0.05, **P < 0.01 vs. Sham group; ^#P < 0.05, ^##P < 0.01 vs. I/R group (n = 6).

Microglial activation is a prominent factor in neuroinflammation. To investigate this, we employed immunofluorescence staining with the Iba-1 antibody 72 h post-GCI/R (Fig. 6B–D). The results showed that the sham and sham + control shRNA groups exhibited minimal Iba-1-positive cells characterized by slender, ramified processes. In contrast, there was a notable increase in the number of Iba-1-positive cells in the I/R and I/R + control shRNA groups compared to the sham group. These cells displayed marked morphological changes, including branch retraction, thickening, and a transition to an amoeba-like form. Interestingly, the I/R + TXNIP shRNA group presented diminished microglial activation relative to the I/R group.

3.7. TXNIP knockdown obstructs the NLRP3/cleaved caspase-1 signaling pathway in the hippocampus caused by GCI/R

Western blotting was employed to evaluate the protein levels of TXNIP, NLRP3, and cleaved caspase-1 in the mouse hippocampus 72 h post-GCI/R (Fig. 7A–D). Both the I/R and I/R + control shRNA groups exhibited elevated levels of these proteins compared to the sham group. Notably, the I/R + TXNIP shRNA group demonstrated a significant reduction in the levels of TXNIP, NLRP3, and cleaved caspase-1 compared to the I/R group. This indicates that the inhibition of TXNIP effectively mitigates the activation of the TXNIP/NLRP3/cleaved caspase-1 signaling pathway in the hippocampus following GCI/R.

Fig. 7.

TXNIP knockdown obstructed the NLRP3/cleaved caspase-1 signaling pathway in the hippocampus caused by GCI/R. (A) Representative bands of Western blot data. (B–D) Quantitative analysis of the Western blot bands. The values are shown as the mean ± SD. *P < 0.05, **P < 0.01 vs. Sham group; ^#P < 0.05, ^##P < 0.01 vs. I/R group (n = 6).

4. Discussion

Bioinformatics analysis based on public microarray profiles (GSE30655 and GSE80681) of cerebral ischemic mice revealed an increased expression of the gene TXNIP in both datasets. GO and KEGG pathway analyses indicated that the inflammatory response may play a crucial role in cerebral I/R injury. Additionally, TXNIP/NLRP3-related genes, including inflammasomes and NLR signaling pathway gene sets, were obtained from the MSigDB database. GSEA analyses demonstrated the enrichment of TXNIP and NLRP3 genes in MCAO mice (GSE30655), belonging to the leading-edge subset. However, no significant difference was found in BCCAO mice (GSE80681), likely due to the small sample size (only six samples). Consequently, the impact of the TXNIP/NLRP3 pathway in GCI/R injury mice was further investigated in this study. Experiments in BCCAO mice revealed that GCI/R injury induced cognitive impairment and time-dependent increases in the levels of TXNIP and NLRP3. Notably, TXNIP knockdown successfully restored cognitive function in mice.

Hippocampal neurons, crucial for spatial learning and memory functions, are particularly vulnerable to ischemic/reperfusion injury, especially the CA1 neurons [8,35]. The C57/BL6 mouse strain is commonly employed to establish the GCI/R model due to its incomplete circle of Willis [36,37]. Previous research using the BCCAO model indicated that 5–20 min of ligation of bilateral CCA ischemia results in compromised integrity of the blood-brain barrier, damage to the white matter, and subsequent harm to hippocampal neurons. Furthermore, spatial memory and learning functions deteriorate during the initial two postoperative weeks [[38], [39], [40]]. In a previous study, we determined that occluding the bilateral carotid arteries for up to 30 min produces a more stable and reproducible GCI/R model [41]. Consequently, in this study, we employed a 30-min ligation BCCAO model. Post-ligation, CBF observed by LSCI analysis significantly decreased compared to the sham group, confirming the reproducibility of the model in this study.

Microglia, recognized as innate immune cells within the central nervous system, exert a pivotal influence on neuroinflammation and oxidative stress [[42], [43], [44]]. Under normal conditions, microglia engage in immune surveillance, characterized by their ramified thin processes. However, when exposed to injury-induced stress, they undergo activation, transforming to display branch retraction, thickening, and an amoeba-like morphology. This activated state prompts the secretion of cytokines such as IL-6, TNF-α, and ROS, amplifying neuronal damage. Our investigations revealed that GCI/R significantly increased the proportion of activated microglia with these morphological changes. Iba1 immunofluorescence analysis demonstrated a notable increase in the count of Iba1-positive cells. Interestingly, TXNIP knockdown led to a reduction in microglial activation. Furthermore, TXNIP knockdown decreased hippocampal levels of inflammatory cytokines (TNF-α, IL-6, IL-1β, and IL-18) and oxidative stress induced by GCI/R. From these observations, we infer that the protective effects of TXNIP knockdown against GCI/R-induced cognitive decline may involve the suppression of microglial-induced proinflammatory cytokine and ROS release.

Previous research has emphasized TXNIP's dual role in disease progression and therapeutic interventions [45]. TXNIP regulates critical pathways influencing apoptosis, inflammation, and cell proliferation across diverse cell types. During episodes of oxidative stress, TXNIP relocates from the nucleus, associating with TRX1 in the cytosol and TRX2 within mitochondria. This interaction leads to an increase in apoptosis signal-regulating kinase 1, triggering apoptosis [46]. Additionally, TXNIP can directly engage with the NLRP3 inflammasome, inducing pyroptosis by stimulating IL-1β secretion. This action occurs through the NLRP3/Caspase-1 pathway, intensifying inflammation, magnifying oxidative stress, and establishing a deleterious feedback loop [47]. Concurrently, TXNIP gene expression maintains a strong connection with the cell cycle, where it acts to suppress cancer cell proliferation and migration [48,49].

Evidence indicates elevated TXNIP levels in various neurodegenerative and cerebrovascular conditions, including Alzheimer's disease (AD), stroke, and subarachnoid hemorrhage (SAH) [50]. In AD, TXNIP's primary role revolves around inflammation [51]. Specifically, TXNIP/NLRP3 inflammasomes become activated by endoplasmic stress, a consequence of irregular accumulations of phosphorylated Tau within the hippocampus [52]. Apoptosis serves as the principal pathological mechanism driving early brain injury post-SAH [53]. Interestingly, past investigations have identified treatments that indirectly reduce TXNIP levels, thereby alleviating inflammation and apoptosis in neurological conditions. For instance, Dl-3-n-butylphthalide can suppress the TXNIP/NLRP3 pathway through the amplification of nuclear factor erythrocyte 2-related factor 2 [54], while compounds like GW0742 or estradiol attenuate TXNIP by elevating miR-17-5p or miR-106b-5p, safeguarding brain cells from apoptosis and inflammatory harm [55,56].

NLRP3 inflammasomes contribute to microglial activation and the release of inflammatory factors. Research indicates a frequent association between cognitive impairment and inflammation, coupled with microglial activation. Excessive NLRP3 activation exacerbates the pathological progression of cognitive decline. Although the intricate signaling cascade of the NLRP3 inflammasome offers various targets for inhibition, current evidence for NLRP3 inflammasome as a therapeutic target is limited and insufficient [57]. Meanwhile, few animal and clinical investigations have explored NLRP3 as a therapeutic target for cognitive impairment, and safety and efficacy should be tested. TXNIP, as the upstream signal of the NLRP3 inflammasome, is a promising target for NLRP3 inflammasome inhibition. Therefore, in our current research, we reduced TXNIP expression by administering an intracerebroventricular injection of AAV-carrying TXNIP shRNA. Our findings revealed that this targeted TXNIP knockdown substantially improved cognitive deficits, decreased the release of TNF-α, IL-6, IL-1β, and IL-18, along with a reduction in apoptosis-positive hippocampal cells, actions mediated by the inhibition of microglia via the NLRP3/cleaved caspase-1 pathway in GCI/R injury mice.

While this research provides valuable insights, it is essential to acknowledge certain limitations. Firstly, the availability of only a few GCI/R injury microarray profile datasets for bioinformatics analysis may result in insufficient representativeness. This constraint could impact the generalizability of the findings. Additionally, our discussion highlights the dual role of TXNIP in disease progression and therapeutic interventions. Focusing exclusively on TXNIP knockdown might lead to an oversight of its potential effects on other biological pathways. The multifaceted role of TXNIP could have unforeseen impacts on cell proliferation and migration. Lastly, considering that microRNA (miRNAs) or long non-coding RNAs (lncRNAs) can regulate the levels of the NLRP3 inflammasome by specifically targeting the binding of TXNIP mRNA, identifying potential miRNAs or lncRNA associated with NLRP3-related cognitive impairment warrants further research.

5. Conclusion

The significance of TXNIP/NLRP3 signaling in the hippocampus is crucial in exacerbating cognitive decline resulting from GCI/R injury. The knockdown of TXNIP emerges as a promising therapeutic strategy for mitigating these detrimental effects.

Founding

This work was supported by grants from the National Natural Science Foundation of China (No. 81873930), partly supported by the Sichuan Science and Technology Program (No.2022YFS0632 and 2022YFS0615), the Youth Program of Southwest Medical University (No.2021ZKQN057), and He Jiang People's Hospital-Southwest Medical University Science and Technology Strategic Cooperation Project (No.2021HJXNYD03).

Institutional review board statement

The animal study protocol was approved by the Animal Ethics Committee of Southwest Medical University (China) (Animal license No. 20211122-012).

Data availability statement

Data will be made available on request.

CRediT authorship contribution statement

Chengjie Yang: Writing – original draft, Investigation, Funding acquisition, Formal analysis. Jing Mo: Investigation. Qingmei Liu: Formal analysis. Wei Li: Formal analysis. Ye Chen: Formal analysis. Jianguo Feng: Methodology. Jing Jia: Methodology. Li Liu: Methodology. Yiping Bai: Funding acquisition, Conceptualization. Jun Zhou: Writing – review & editing, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors have no relevant financial or non-financial interests to disclose.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

figs1.

figs2

figs3.

figs4.

figs5.

figs6.

figs7.

figs8.

figs9.

figs10.