ApoE4 increases GSK-3β levels to worsen Alzheimer-like brain damage and memory problems in mice with type 2 diabetes

All mice were housed at 23°C–25°C under a 12‐h light/dark cycle. Food and water were provided ad libitum throughout the experiment. All animal experiments were approved by the Institutional Animal Care and Use Committee of Tongji Medical College, Huazhong University of Science and Technology. The humanized ApoEε3/ε3 knock‐in mice (ApoE3 KI mice, RRID: 029018) and ApoEε4/ε4 knock‐in mice (ApoE4 KI mice, RRID: 027894) were generated by the Jackson Laboratory. In the ApoE4 KI mice, exons 2, 3, and the majority of exon 4 of the endogenous ApoE gene were replaced by the human ApoE gene sequence. To generate ApoE3 KI mice, CRISPR/Cas9 genome editing was employed to introduce a C‐to‐T nucleotide substitution into the human APOE sequence of the ApoE4 KI allele, resulting in the R130C mutation. The animals were distributed into various experimental groups with balanced littermates and gender, ensuring each group contained an equivalent number of male and female siblings.

Antibodies and reagents used in this study are listed as Table. S1

To establish a type 2 diabetes mellitus (T2DM) model in 10‐week‐old mice, a 60% high‐fat diet (HFD, D12492; Research Diets) was fed for 8 weeks. Following this dietary intervention, streptozotocin (STZ, 30 mg/kg, 572201; Sigma‐Aldrich) was dissolved in citrate buffer (pH 4.35) and administered via intraperitoneal injection over five consecutive days. Control mice received an equivalent volume of citrate–phosphate buffer. Prior to each injection, all mice were fasted for 6–8 h, with free access to water maintained. Individual body weights were recorded to accurately determine the dose of STZ. Given the instability of STZ, the solution was prepared immediately before use, kept away from light, and administered within 30 min to ensure efficacy. The successful induction of the T2DM model was confirmed by fasting blood glucose levels exceeding 11.1 mmol/L.

To prepare a solution at a concentration of 4 mg/mL, mix 10% DMSO with 20% PEG400 to dissolve the GSK‐3β inhibitor 9‐ING‐41 (Elraglusib, HY‐113914; MCE). Due to the limited solubility of 9‐ING‐41, ultrasonic assistance may be utilized to enhance the dissolution process. The solution should be freshly prepared for each use. Administer the drug via intraperitoneal injection at a dosage of 20 mg/kg in 19‐week‐old ApoE4‐T2DM mice, ensuring accurate calculation of the injection volume for each mouse. Injections should be conducted three times a week at the same time, continuing for a total duration of 4 weeks.

Following the collection of blood from the mice via orbital puncture, transfer the samples into appropriately labeled Eppendorf tubes. Allow the samples to stand overnight at 4°C. Subsequently, centrifuge the samples at 3000 rpm for 15 min at 4°C. Carefully transfer the supernatant into new, labeled Eppendorf tubes. Transport the samples on dry ice to Servicebio Technology Co. Ltd. (Wuhan) for analysis. An automatic biochemical analyzer (Chemray 240) from Leadbio Life Science Technology Co. Ltd. (Shenzhen) will be employed for the measurements.

The skulls of the mice were carefully removed, and the hippocampal tissue was dissected on ice. The samples were homogenized in RIPA lysis buffer containing PMSF (Millipore) and a protease inhibitor cocktail (Sigma) at a ratio of 10 μL per mg of tissue. Ultrasonic disruption was performed on ice for 10–15 s, followed by centrifugation at 12,000 rpm for 15 min at 4°C. The supernatant was collected, and the protein concentration was assessed using the BCA assay kit (Thermo Fisher). Proteins were then separated by SDS‐PAGE (10%) for approximately 1.5 h and subsequently transferred to nitrocellulose membranes (Merck Millipore; 0.45 μm) for 1 h. The membranes were blocked with 5% bovine serum albumin (BSA) for 1 h. Following this, the membranes were incubated sequentially with primary and secondary antibodies. Protein bands were visualized using an ECL Imaging System (610007‐8Q; Clinx Science Instruments Co. Ltd.) and quantified with ImageJ software.

Mice were anesthetized with 2% isoflurane 1 day post‐behavioral testing and perfused intracardially with 0.9% NaCl followed by 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer (pH 7.4). Brains were removed and cryoprotected in 25% and 30% sucrose for 2 days, then sectioned into 30 μm slices using a cryostat (Leica CM1900). For immunohistochemistry, free‐floating sections were treated with 3% H₂O₂ in methanol for 30 min to block endogenous peroxidase activity, followed by blocking with 5% BSA for 30 min at room temperature. Sections were incubated with primary antibodies overnight at 4°C, then dehydrated through a graded ethanol series (75%, 95%, 100%) and cleared in xylene (two 10‐min rounds). Slices were mounted with neutral balsam, dried in the dark, and immunoreactivity visualized using a DAB staining kit (ZSGB‐BIO). Images were captured with an automatic slide scanner (SV120; Olympus). For immunofluorescence, sections were blocked in 5% BSA with 0.1% Triton X‐100 for 30 min at room temperature, incubated with primary antibodies overnight at 4°C, washed in PBS (three 5‐min washes), and incubated with secondary antibodies at 37°C for 1 h. Nuclei were counterstained with DAPI. Images were obtained using an automatic slide scanner (SV120; Olympus) and a confocal microscope (Zeiss Carl LSM 800). Quantitative analysis was conducted using ImageJ to assess the staining area and the number of GFAP‐ and Iba1‐positive cells. One slice per mouse was analyzed, with 6–7 mice per group.

Brain sections were rinsed with phosphate‐buffered saline (PBS) for 15 min and subsequently stained with 0.3% Thioflavin S (#1326‐12‐1; Sigma) in 50% ethanol at room temperature for 10 min. The staining duration was adjusted according to the observed intensity under the microscope. After staining, the sections were decolorized in 50% ethanol for 3–5 min, washed with PBS, and mounted using a coverslipping medium containing DAPI.

After mounting the brain sections on adhesive slides, they were stained using Nissl stain for 10 min. The sections were subsequently decolorized in 75% ethanol for approximately 5 min, followed by sequential treatments with 80%, 90%, and 100% ethanol for 2 min each. The staining and decolorization durations were adjusted based on the observed staining intensity under the microscope. Following dehydration, the sections were cleared in xylene for 1 h and mounted using neutral balsam.

For mouse serum, orbital blood was collected and incubated at 4°C overnight. The sample was then centrifuged at 3000 rpm for 15 min, and the supernatant was analyzed immediately. For mouse cortical tissue, nine volumes of physiological saline were added to the tissue (1 mg of tissue to 9 μL of saline). The mixture was homogenized on ice to produce a 10% homogenate and then centrifuged at 3000 rpm for 10 min. The supernatant was analyzed according to the instructions of the reagent kit, and Aβ40 and Aβ42 levels were detected using commercial ELISA kits (E‐EL‐H0542c 90 and E‐EL‐H0543c; Elabscience).

After the mouse was anesthetized with 2% isoflurane, the brain tissue was carefully extracted by removing the skull. The tissue was then immersed in Golgi staining solution, which was prepared by mixing 5% potassium dichromate, 5% silver nitrate, and an equal volume of 5% potassium dichromate solution in a 1:1:2 ratio for 2 weeks in the dark. Following the staining process, the tissue was transferred to a preservation solution containing 300 g of sucrose, 500 mL of phosphate buffer (PB), 200 mL of ethanol, and 4 g of polyvinylpyrrolidone for 2 days, also in the dark. Sections of 100 μm thickness were cut while the brain was immersed in PB.

The sections underwent a series of washes: twice in distilled water (3 min each), followed by a rinse in 50% ethanol for 5 min, and then treatment with a 66% ammonia solution for approximately 10 min. They were rinsed again twice in distilled water (5 min each), incubated in 5% sodium thiosulfate for 10 min in the dark, and subsequently washed twice more in distilled water (5 min each). The sections were mounted on slides with 2% gelatin, dehydrated through a graded series of ethanol (75%, 85%, 90%, 100%) for 2 min each, cleared in xylene for 20 min, and finally sealed with neutral gum. Subsequently, these sections were imaged using a light microscope (Nikon, Japan) for further analysis. Spine densities were quantified as the number of spines per 10 mm of dendrite length, measured using ImageJ 2.0 software.

During the experiment, the mouse was placed in a square enclosure measuring 60 × 60 × 50 cm³. The floor of the enclosure was evenly divided into 16 squares, including a central zone (comprising the central four squares) and 12 peripheral zones. The mouse was introduced into the enclosure with its back facing the experimenter and was allowed to explore freely for 5 min. The environment was kept quiet throughout the test, and a video tracking system (Taimeng Software Co. Ltd., Chengdu, China) was used to record and analyze the mouse's behavior.

The experiment lasted 2 days. On the first day, two identical objects, A and B, were placed in opposite corners of the enclosure, and the mouse was allowed to explore the objects for 5 min. On the following day, object B was replaced with a novel object C, and the mouse was also given 5 min to explore the objects. The video tracking system (Anymaze Technology SA, Stoelting Co., IL) identified the mouse's exploration of the objects based on the proximity of the mouse's head to each object. The exploration times for objects A and C were recorded as TA and TC, respectively, and the discrimination index was calculated as (TC − TA)/(TC + TA). Throughout the experiment, the environment was kept quiet, and handling of the mouse was gentle. Before testing each new mouse, the enclosure was cleaned of feces and disinfected with 75% ethanol.

The experiment was conducted over 2 days. On Day 1, two identical objects, A and B, were placed in opposite corners of the enclosure, and the mouse was allowed to explore the objects for 5 min. On the following day, object B was relocated to a different corner of the enclosure while object A remained in the same corner. The mouse was also given 5 min to explore the objects. The video tracking system (Anymaze Technology SA, Stoelting Co., IL) recorded the mouse's exploration of the objects based on the proximity of the mouse's head to each object. The exploration times for objects A and B were recorded as TA and TB respectively. On Day 2, the discrimination index was calculated as (TB − TA)/(TB + TA). Throughout the experiment, the environment was kept quiet, and handling of the mouse was gentle. Before testing each new mouse, the enclosure was cleaned of feces and disinfected with 75% ethanol.

The water maze (a circular pool, 1.2 m in diameter and 0.6 m in height) was divided into four quadrants, with a platform placed 2 cm below the water surface in one quadrant. The pool's walls were marked with brightly colored signs, and a camera system (Taimeng, China) was used to track the mice's movements and speed. Before the experiment, mice were acclimated in the experimental room for at least 6 h, and heaters and dry bedding were prepared. The learning phase lasted 5 days, with three trials per day and at least a 30‐min interval between trials in each quadrant. Mice were gently placed in the water to find the platform, with a maximum swim time of 60 s per trial. If a mouse found the platform, its time was recorded, and it was allowed to remain on the platform for 15 s. If the platform was not located within 60 s, the mouse was guided to it and allowed to stay for 15 s. On Day 6, spatial memory was assessed by removing the platform and placing the mice in the quadrant opposite to the platform's original location. Mice were given 60 s to search for the platform, and the number of crossings over the former platform location was recorded, along with the percentage of distance moved in the target quadrant.

Data analysis and plotting were performed using GraphPad Prism 8 (La Jolla, CA). The Shapiro–Wilk test was used to evaluate normality, and non‐parametric tests were employed when required. Unpaired t test with Welch's correction was used for comparisons between two groups. For comparisons involving three or more groups, either one‐way ANOVA and Bonferroni's multiple comparison test or two‐way ANOVA and Bonferroni's post hoc test were applied for multiple comparisons. All experiments and analysis were performed in a blind manner, and experiments were replicated at least three times to ensure sufficient statistical power. Data are expressed as mean ± SEM or min‐median‐max unless otherwise specified. The probability (p) value < 0.05 was considered statistically significant.

Rodents possess a single ApoE genotype that closely resembles human ApoE4 [31, 32]. To investigate the potential underlying mechanisms and the related preventive strategies for ApoE4‐exacerbated cognitive impairment in T2DM, we developed T2DM mouse models in humanized ApoE3 and ApoE4 knock‐in mice (Figure 1a). The mice carrying either the ApoE3 or ApoE4 genotype were randomly assigned to two groups: a normal diet group and a T2DM model group with an equal ratio of male to female mice maintained in each group. The T2DM model group was fed a high‐fat diet for 8 weeks, followed by intraperitoneal injections of streptozotocin (30 mg/kg/day for five consecutive days) (Figure 1b). During the model induction, body weight and blood glucose levels were dynamically monitored (Figure 1c,d). A fasting blood glucose level exceeding 11.1 mmol/L or a random blood glucose level greater than 16.7 mmol/L was considered indicative of successful establishment of the T2DM model [33, 34]. Furthermore, we performed weekly random blood glucose measurements for 4 weeks after STZ injection in ApoE3‐T2DM and ApoE4‐T2DM mice that continued to be fed a high‐fat diet to rule out non‐specific toxicity to brain tissue (Figure S1a).

T2DM mice exhibited symptoms resembling those of clinical diabetes patients, including hyperglycemia and body weight loss (Figure 1c,d). Additionally, T2DM mice displayed liver enlargement, a gray‐white appearance, a firm texture, and an oily surface film (Figure S1b). Compared with E3‐T2 or E4 mice, E4‐T2 mice exhibited a significantly increased liver index (Figure S1c), while serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels remained unchanged (Figure S1d,e). Furthermore, T2DM mice showed elevated serum levels of triglycerides (TG), cholesterol (CHO), and low‐density lipoprotein (LDL) compared with the normal diet group (Figure S1f–i). We measured the serum glucose (GLU) and glycated serum protein (GSP) levels in the mice, noting that while GLU reflects immediate glycemic status, GSP indicates average blood glucose levels over the past few weeks. Both GLU and GSP levels were significantly higher in T2DM mice compared to non‐T2DM mice (Figure S1j,k). These results confirm that both ApoE3‐T2DM and ApoE4‐T2DM mice showed elevated blood glucose, and an exacerbated hepatic lipid accumulation was seen in ApoE4‐T2DM mice.

GSK‐3β is a critical enzyme involved in the phosphorylation of tau protein [35], and the aggregation of hyperphosphorylated tau is a hallmark of AD [36]. To determine whether ApoE4 affects the phosphorylation of tau protein in T2DM mice, we assessed the levels of total tau (detected by tau5 antibody) and tau phosphorylation at multiple AD‐related epitopes in the hippocampus of mice. Immunohistochemical staining revealed that the levels of phosphorylated tau at pS202/pT205 sites (detected by AT8 antibody) were significantly increased in 6‐month‐old E4‐T2 mice compared with E3‐T2 and E4 mice (Figure 2a,b). Additionally, Western blotting analysis indicated that both total tau and phosphorylated tau levels (including the pS404, pS396, and pS199 epitopes) were significantly elevated in the hippocampus of E4‐T2 mice compared with E3‐T2 mice. Furthermore, ApoE4 enhanced tau phosphorylation at the pS404, pS396, and pS199 epitopes in non‐diabetic mice (Figure 2c,d). These results suggest that both ApoE4 and T2DM contribute to increased tau phosphorylation, with the most pronounced AD‐like tau pathologies observed in ApoE4‐T2DM mice.

Abnormal aggregation of β‐amyloid (Aβ), forming senile plaques, is a hallmark of AD [37]. Studies have indicated that ApoE plays a critical role in regulating the metabolism and clearance of Aβ [38]. Specifically, the ApoE4 allele may increase the risk of AD by accelerating the aggregation and deposition of Aβ in the brain [37, 39]. To investigate the potential influence of ApoE4 on Aβ pathology in T2DM mice, we performed Thioflavin S staining on 6‐month‐old mice. No positively stained plaques were observed in these mice, unlike the age‐matched 5xFAD mice, which express human APP and PSEN1 transgenes harboring five AD‐associated mutations and display significant Aβ plaque deposition in the hippocampus, especially in the DG region (Figure S2a). Then, we quantified the Aβ42/Aβ40 ratio in both brain and serum using ELISA. We found that E4‐T2 mice exhibited an increased Aβ42/Aβ40 ratio in the cortex compared with E3‐T2 mice (Figure S2b). Additionally, no significant difference but an increasing trend of Aβ42/Aβ40 ratio was observed in the serum of E4‐T2 mice compared with E3‐T2 mice (Figure S2c). These results suggest that the influence of ApoE4 and T2DM on the brain Aβ deposition is limited.

Accumulating evidence has suggested that gliosis and neuroinflammation play pivotal roles in the pathology of AD [40]. Immunofluorescence staining showed that ApoE4 significantly increased the number of GFAP‐labeled astrocytes and IBA1‐labeled microglia in hippocampal CA1 of T2DM mice (Figure 3a–c). Furthermore, the number of astrocytes and microglia was increased in T2DM mice with different humanized ApoE genotypes (Figure 3a–c). Activated microglia play a central role in the progression of neurodegenerative diseases, including AD, by modulating inflammatory responses through the release of various mediators, which subsequently affect the overall function and structure of the nervous system [41]. These inflammatory mediators include tumor necrosis factor‐alpha (TNF‐α), interleukin‐6 (IL‐6), and nitric oxide (NO) [42]. Western blotting analysis showed that the IL‐6 expression level was increased in the hippocampus of T2DM mice with different humanized ApoE genotypes (Figure 3d,e). Additionally, TNF‐α expression was increased in E4 mice compared with E3 mice (Figure 3d,e). These observations suggest that T2DM promotes hippocampal glia‐associated expression of inflammatory factors and ApoE4 exacerbates the effects.

The aggregation of phosphorylated tau and activation of glial cells are closely linked to synaptic impairment. Severe synaptic deficits may lead to neuronal death [43]. To evaluate the impact of ApoE4 and T2DM on synaptic integrity, we measured the levels of synapse‐related proteins and spine density in the hippocampus of mice. Western blotting analysis showed that both ApoE4 and T2DM significantly reduced the levels of presynaptic synapsin‐1 (Syn‐1) and synaptophysin (SYP). Additionally, ApoE4 decreased the levels of postsynaptic NMDA receptor subunit 1 (GluN1) and glutamate receptor 2 (GluR2), although these changes were not statistically significant (Figure 3f,g). Golgi staining showed that spine density in the hippocampal DG was significantly decreased in the E4‐T2 and E4 groups compared with the E3‐T2 and E3 groups, respectively (Figure 3h,i). These findings suggest that ApoE4 exacerbates synaptic deficits in both T2DM and non‐T2DM mice, and T2DM also contributes to synaptic damage. Additionally, neither ApoE4 nor T2DM led to decreased neuron numbers in the hippocampal CA1 subset during the observed period, as assessed by Nissl staining (Figure S3a,b). Collectively, these findings demonstrate that T2DM and ApoE4 exacerbate synaptic deficits without inducing neuronal loss in mice.

To investigate the effects of ApoE4 and T2DM on learning and memory, we conducted a series of behavioral tests to assess cognitive function in mice. In the novel object recognition (NOR) test, E4‐T2 mice showed a decreased discrimination index for the novel object compared with E3‐T2 mice. Similarly, E4 mice exhibited a decline in the discrimination index compared with E3 mice (Figure 4a,b). In the object place recognition (OPR) test, both E4‐T2 and E4 mice displayed poorer performance in discriminating objects moved to a new location than E3‐T2 and E3 mice (Figure 4a,c). During the learning trial of the Morris water maze (MWM) test, E3‐T2 mice showed longer latency on day 4 than E3 mice, while E4‐T2 mice exhibited prolonged latency on day 5 compared with E3‐T2 or E4 mice (Figure 4d). During the probe test on day 6, E4‐T2 mice showed fewer platform crossings (Figure 4e,g) and a lower percentage of distance traveled in the target quadrant (Figure 4h) compared with E3‐T2 mice. No significant differences were observed among the four groups in swimming speed (Figure 4f). Importantly, correlation analysis revealed a significant negative relationship between tau phosphorylation at multiple epitopes (pT231, pS404, pS396, and pS199) in the hippocampus and the number of platform crossings, demonstrating that increased tau phosphorylation is associated with worse spatial memory performance (Figure S4a–d). These data suggest that both ApoE4 and T2DM impair spatial learning and memory in mice. In the open field test, no significant differences were observed in total distance traveled or center area entries among the four groups of mice (Figure 4i–k), suggesting that neither ApoE4 nor T2DM significantly affects anxiety‐like behavior. Collectively, these results demonstrate a synergistic effect of ApoE4 and T2DM on cognitive impairment in mice, and ApoE4‐T2DM mice show the most severe cognitive deficits.

Our previous studies have shown that T2DM patients carrying the ApoE4 allele exhibit increased activity of GSK‐3β in platelets and more severe cognitive impairment [27, 28]. To investigate whether ApoE4 influences cerebral GSK‐3β expression and the insulin signaling pathway in T2DM mice, we first assessed the levels of total GSK‐3β (tGSK‐3β) and phosphorylated GSK‐3β at Tyr216 (pGSK‐3β‐Tyr216, active form) in 6‐month‐old mice (Figure 5a,b). Immunohistochemical analysis showed significantly increased expression of tGSK‐3β and the phosphorylated pGSK‐3β‐Tyr216 in the hippocampal dentate gyrus (DG) of E4‐T2 mice compared with E3‐T2 mice (Figure 5a–d). This finding was further validated by Western blotting analysis, which showed consistent upregulation of tGSK‐3β and pGSK‐3β‐Tyr216 in the hippocampus of E4‐T2 mice (Figure 5e,h). These results suggest that ApoE4 enhances GSK‐3β activation in T2DM mice.

GSK‐3β is an essential component of the insulin signaling pathway. Under physiological conditions, upstream PI3K‐AKT signaling pathways phosphorylate GSK‐3β at Ser9, leading to its inactivation [21, 22, 23]. This modification promotes glycogen synthesis and reduces blood glucose levels [22]. Studies have indicated that ApoE4 interacts with insulin receptors, thereby impairing cerebral insulin signaling [44]. Therefore, we examined the expression of insulin‐related signaling molecules in the hippocampus of 6‐month‐old mice A significantly reduced phosphorylated AKT at Ser473 (pAKT‐Ser473) with phosphorylated GSK‐3β at Ser9 (pGSK‐3β‐Ser9) was detected in the hippocampus of E4‐T2 mice compared with E3‐T2 and E4 mice (Figure 5e–h). Additionally, a significantly increased tGSK‐3β was also observed in the hippocampus of E4‐T2 mice (Figure 5e,h), which was further confirmed by immunohistochemistry (Figure 5c,d). These findings indicate that ApoE4 disrupts insulin signaling and enhances GSK‐3β activity in T2DM mice, which can contribute to the exacerbated AD‐related pathology in ApoE4‐T2DM populations.

To further evaluate the potential of GSK‐3β as a therapeutic target for AD‐like pathologies and cognitive impairments in ApoE4‐T2DM conditions, we administered intraperitoneally 9‐ING‐41, a specific small‐molecule inhibitor of GSK‐3β [45], into the ApoE4‐T2DM mice (Figure 6a). We observed that 9‐ING‐41 significantly inhibited GSK‐3β expression level with modulated phosphorylation status in the hippocampus of E4‐T2 mice (Figure 6d,e). Specifically, 9‐ING‐41 increased inhibitory phosphorylation of GSK‐3β at Ser9 and decreased the level of Tyr216 phosphorylation, the active form of GSK‐3β (Figure 6b–e). However, 9‐ING‐41 did not affect the expression or phosphorylation of AKT (Figure 6d,e), which serves as an upstream regulator in the insulin signaling pathway. Serum biochemical analysis indicated that the intraperitoneal injection of 9‐ING‐41 significantly reduced random blood glucose levels in E4‐T2 mice without causing changes in body weight (Figure S5a,b). Additionally, 9‐ING‐41 treatment induced a decreasing trend in TG and LDL in E4‐T2 mice (Figure S5g–j). Although 9‐ING‐41 did not affect liver index or liver transaminases (ALT and AST) in E4‐T2 mice (Figure S5d–f), the mouse livers treated with 9‐ING‐41 appeared redder, and the sensation of a surface oil film was notably reduced (Figure S5c). These data suggest that 9‐ING‐41 effectively downregulates GSK‐3β expression and activity in the hippocampus with a limited effect on glucose and lipid metabolism in the hippocampus of ApoE4‐T2DM mice.

To determine whether AD‐like pathologies could be alleviated by intraperitoneal injection of 9‐ING‐41, we assessed its effects on tau phosphorylation, gliosis, and synaptic dysfunction in the hippocampus of T2DM mice. The phosphorylation levels of tau at Ser199, AT8 (Ser202/Thr205), Thr231, and Ser404 in the hippocampus of E4‐T2‐9‐ING‐41 mice were significantly reduced compared with those of PBS‐injected E4‐T2 (E4‐T2‐PBS) mice measured by Western blotting (Figure 7a,b). Additionally, 9‐ING‐41 mitigated the activation of microglia and astrocytes (Figure 7c–e), with significantly decreased levels of inflammatory mediators in the hippocampus measured by Western blotting and immunofluorescence (Figure 7a,b,f,g). These results suggest that GSK‐3β inhibitor 9‐ING‐41 alleviates AD‐like tau hyperphosphorylation and gliosis in ApoE4‐T2DM mice.

To evaluate the effects of intraperitoneal injection of 9‐ING‐41 on cognitive function and the molecular mechanisms in ApoE4‐T2DM mice, we observed that 9‐ING‐41 significantly increased Syn‐1 level without affecting GluR2 and synaptophysin (SYP) levels in the hippocampus of E4‐T2 mice (Figure 8a,b). Spine density was significantly increased in E4‐T2‐9‐ING‐41 mice compared with E4‐T2‐PBS mice (Figure 8c,d). Furthermore, E4‐T2 mice treated with 9‐ING‐41 showed an increased discrimination index compared with the PBS‐treated group in the NOR and OPR tests (Figure 8e,f). During the MWM training trials, E4‐T2‐9‐ING‐41 mice exhibited shorter latencies to find the platform on days 2 and 4 compared with the E4‐T2‐PBS group (Figure 8g). Consistently, 9‐ING‐41‐treated mice showed an increased number of platform crossings (Figure 8i,j) and a higher percentage of distance traveled in the target quadrant during the test phase (Figure 8k). Additionally, there were no significant differences in swimming speed (Figure 8h). Overall, these data suggest that inhibiting GSK‐3β by 9‐ING‐41 effectively mitigates synaptic deficits and cognitive impairment in ApoE4‐T2DM mice.

Data available on request from the authors. The data that support the findings of this study are available from the corresponding author upon reasonable request.