2.1 Animals

Female triple-transgenic AD (3xTg-AD) mice (Stock No. 033930; Jackson Laboratory), co-expressing human MAPT, APP, and PSEN-1 transgenes [24], were obtained from the Shenzhen Center for Disease Control and Prevention and housed under standardized conditions (12-h light/dark cycle, ad libitum access to food and water). At 7 months of age—corresponding to the early pre-symptomatic phase of AD pathology [23]—mice were randomly assigned to either a normal diet (ND: 18% kcal from fat, 3.1 kcal/g; Teklad) or a high-fat diet (HFD: 60% kcal from fat [31% lard, 3.7% soybean oil], 20% kcal from protein, 20% from carbohydrates; Research Diets, D12492) for 16 weeks. This HFD duration was selected based on prior studies demonstrating its efficacy in inducing metabolic dysfunction and cognitive deficits in 3xTg-AD models [8, 25-29]. Female mice (n = 10–12 per group) were exclusively used to ensure phenotypic consistency and to account for their heightened amyloid pathology compared to males [30]. Age-matched female C57BL/6J mice served as the wild-type control (WT). All procedures complied with ARRIVE guidelines and were approved by the Animal Care and Use Committee of Guangzhou University of Traditional Chinese Medicine (Approval No. 20230526004).

2.2 Electroacupuncture (EA-ST36) Treatment

Throughout the 4-month HFD regimen, mice were treated with EA. Behavioral and metabolic assessments were conducted concurrently. At 7 months of age, mice were randomly allocated to either the Sham or EA group. EA stimulation was performed at the ST36 acupoints, following previously established protocols [23]. The parameters (2 Hz frequency, 1 mA intensity, 20 min/session, 5 sessions/week) were selected based on rodent studies demonstrating neuroprotection and anti-inflammatory effects [31]. The ST36 acupoints, located 5 mm below the fibular head and 2 mm lateral to the anterior tubercle of the tibia, were prepared by cleansing with alcohol, and needles were inserted to a 10 mm depth. Mice were anesthetized with 2% isoflurane (RWD Life Science, R510-22) and positioned in a stereotaxic apparatus (RWD Life Science, 71,000). An electrical stimulator (Huatuo SDZ, China) was connected to the needle handles, applying the aforementioned parameters over 16 weeks. Sham-EA animals received superficial needle insertion at ST36 without electrical stimulation. Mice were used for further analyses at 11.5 months of age, after treatment completion.

2.3 AAV-Mediated Tfeb and Tfe3 Knockdown in the Hippocampus

AAV-mediated hippocampal knockdown of TFEB and TFE3 was performed using shRNA constructs (BrainVTA Co. Ltd., China) to investigate their role in EA-ST36-mediated autophagy regulation. The Tfeb shRNA sequence (CCGGCGGCAGTACTATGA CTATGATCTCGAG-ATCATAGTCATAGTACTG CCGTTTTTG) and Tfe3 shRNA sequence (CCGGGCCTAACATCAAACGCGAGATCTCGAGATCTCGCGTTTGATGTTAGGCTTTTTG) were synthesized based on prior studies [30]. Viral constructs—rAAV-U6-shRNA(TFE3)-U6-shRNA(TFEB)-CMV-EGFP-SV40 pA, AAV9 (5.87E+12 vg/mL), and control rAAV-U6-shRNA(Scramble)-U6-shRNA(Scramble)-CMV-EGFP-SV40 pA, AAV9 (5.10E+12 vg/mL)—were produced by BrainVTA Co. Ltd. (Wuhan, China). Female 3xTg-AD mice (7–8 months, n = 36) received bilateral hippocampal CA1 injections (coordinates: −2.06 mm AP, ±1.50 mm ML, 1.4 mm DV from bregma) of 2 μL viral solution at 4 nL/s. Injections were administered two weeks prior to EA-ST36 treatment, followed by behavioral and biochemical analyses.

2.4 Metabolic Analysis

Body weight was monitored weekly throughout the study. After 4 months of treatment, insulin sensitivity and glucose tolerance were evaluated via intraperitoneal insulin tolerance tests (ITTs) and glucose tolerance tests (GTTs), respectively (n = 10–12 mice/group). To control for hormonal fluctuations, all tests were performed during the diestrus phase of the estrous cycle in female mice. Following a 6-h fast, mice received intraperitoneal injections of human insulin (1 unit/kg) or glucose (1 g/kg) [32]. Blood glucose levels were measured from saphenous vein samples using a glucometer (Model 710, Yuwell-Jiangsu Yuyue Medical Equipment & Supply Co. Ltd) at baseline (t = 0) and 15, 30, 60, and 120 min postinjection. The area under the curve (AUC) for glucose concentration was calculated as previously described [33].

2.5 Behavioral Assessment

The Morris Water Maze (MWM) was employed to assess hippocampus-dependent spatial memory, following established protocols [34]. Tests were conducted during the animals' active light phase (07:00–19:00), with each mouse's swim trajectory recorded via a video analysis system (Shanghai Jiliang Software Technology Co. Ltd., Shanghai, China). Pool temperature was maintained at 21°C ± 1°C, and an escape platform (4.5 cm diameter) was submerged 0.8–1 cm below the water surface. Mice underwent five days of training, with four 60-s trials per day (30-min intervals between trials). Mice unable to find the platform were placed on it for 10 s. On the sixth day, the platform was removed for a 60-s probe trial to assess retention.

The Y-maze test evaluated spatial working memory by measuring spontaneous alternation behavior (SAB), a hippocampal-dependent metric of memory and attention [35]. Mice were permitted to explore the maze for 5 min or until 15 arm entries. SAB was calculated as the number of alternations divided by total alternation opportunities, using the first 15 arm choices for analysis to control for variability.

2.6 Immunohistochemistry

Mice were anesthetized and perfused with 70 mL of ice-cold PBS, followed by 20 mL of 4% paraformaldehyde (PFA) in PBS. Tissues were fixed overnight at 4°C, and frozen brain sections (40 μm) were prepared using a microtome (HM 525, Leica Biosystems). Sections were permeabilized with 0.3% Triton X-100 (Beyotime Biotech, ST795) and blocked with 5% donkey serum (Jackson ImmunoResearch Inc., 102,974) in PBS, then incubated with primary antibodies against NLRP3, ASC, or SQSTM1/p62 at dilutions of 1:300 or 1:500 overnight at 4°C. After PBS washes, sections were incubated with Alexa Fluor-conjugated secondary antibodies (1:1500, Thermo Fisher Scientific, A11008, A11012, A11001, A11005) for 1.5 h at room temperature. Nuclear staining was performed with DAPI, and sections were mounted and imaged using a Nikon A1R confocal microscope (Nikon Instruments Inc.). Images were processed with Adobe Photoshop CS (San Jose, CA, USA) and quantified using ImageJ (NIH, USA).

2.7 Tissue Extraction and Western Blot Analysis

The hippocampus was dissected and homogenized in RIPA buffer (1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS) containing protease and phosphatase inhibitors (Thermo Fisher Scientific, A32957, A32955). Sarkosyl fractionation was conducted to separate soluble and insoluble protein fractions, following previously described protocols [36]. Protein concentrations were determined via BCA assay (Beyotime Biotechnology, P0009). Equal protein amounts were subjected to 8%–15% SDS-PAGE and transferred onto PVDF membranes (GE Amersham, 10600023). Membranes were blocked with 5% nonfat milk and incubated with primary antibodies overnight, followed by incubation with secondary antibodies for 2 h at room temperature. Signals were detected using ECL kits (Thermo Fisher Scientific, 32106, 34580) and analyzed with ImageJ software.

Antibodies used: Anti-LAMP1 (ab24170), anti-CTSD (ab75852), anti-p-RELA/ NF-κB-p65 (S536) (ab86299), anti-CASP1 (ab179515), anti-IL1B (ab9722) and anti-SQSTM1 (ab109012, used for IHC) were purchased from Abcam.Anti-NLRP3 (NBP2-12446), Anti-LC3B (NB100-2220) was purchased from Novus Biologicals. Anti-TFEB (A303-673A) was purchased from Bethyl Laboratories. Anti-phospho-TFEB (Ser142; ABE1971), Anti-TFE3 (HPA023881) and Anti-SQSTM1/p62 (P0067, used for WB) were purchased from Sigma-Aldrich. Anti-H3F3A/histone H3 (D1H2; 4499), anti-PSEN1 (5643), anti-phospho-MTOR (Ser2448; 2971), anti-MTOR (2983), anti-phospho-AMPKα (Thr172; 2535), anti-AMPKα (5831), anti-ULK1(#8054), anti-phospho-ULK1 (Ser757) (#6888) were purchased from Cell Signaling Technology. Anti-AT8 (Ser202/Thr205) (MN1020), anti-AT270 (Thr181) (MN1050), anti-AT100(Thr212, Ser214)(MN1060), Anti-Tau Monoclonal Antibody (HT7)(MN1000), Anti-APP (51–2700)were purchased from Thermo Fisher Scientific.Anti-β-amyloid 1–42 (AB5078P) was purchased from Millipore. Anti-β-amyloid (6E10; 803,017) was purchased from Biolegend. Anti-TUBB/β-tubulin (H-235; sc-9104), anti-GAPDH (G-9; sc- 365,062), anti-ASC (sc-271054), and anti-ACTB/β-actin (sc-47,778) were purchased from Santa Cruz Biotechnology.HRP-conjugated goat anti-mouse (115-035-003) and goat anti-rabbit (111–035-003) secondary antibodies were purchased from Jackson ImmunoResearch.

2.8 Statistical Analysis

Data are presented as mean ± SEM. Analyses were conducted using GraphPad Prism 8.0.1. Behavioral data were evaluated using two-way ANOVA with Bonferroni post hoc tests. Comparisons between two groups were conducted using unpaired t-tests, while one-way ANOVA with Dunnett's multiple comparison test was applied for three or more groups. Outliers were identified via the ROUT method (Q = 1%). Significance was set at p < 0.05, 0.01, or 0.001.

3.1 EA-ST36 Attenuates HFD-Induced Metabolic Dysfunction in 3xTg Mice

Mice were fed either a ND or an HFD for 16 weeks. HFD-fed 3xTg-AD mice exhibited significant weight gain compared to the normal diet (ND) group (Figure S1A), consistent with HFD-induced obesity. EA-ST36 treatment partially attenuated this weight gain, though the effect was not statistically significant. Metabolic analyses (Figure S1B,C) demonstrated that EA-ST36 improved glucose tolerance in HFD-fed mice, suggesting a beneficial effect on metabolic dysfunction. However, insulin sensitivity (Figure S1E) was not significantly altered by HFD or EA-ST36. These results indicate that EA-ST36 primarily targets neuropathology rather than systemic metabolism, though metabolic improvements may contribute indirectly to cognitive rescue.

3.2 EA-ST36 Attenuates HFD-Induced Learning and Memory Deficits in 3xTg-AD Mice

To enhance memory deficits, 3xTg-AD mice were administered a high-fat diet (HFD) starting at 7 months, a regimen known to exacerbate AD-like pathology [2]. As anticipated, HFD worsened behavioral disorders and cognitive deficits in these mice. To evaluate the protective effects of EA-ST36, Morris Water Maze (MWM) and Y-maze tests were conducted. In the MWM test, HFD-fed 3xTg-AD mice exhibited significantly longer escape latencies than mice on a standard diet, particularly on training days 4 and 5 (Figure 1D). Notably, treatment with EA-ST36 significantly reduced escape latencies on these days, indicating a protective effect (Figure 1D). However, Sham-EA treatment did not produce a notable reduction in escape latency (Figure 1D). In the probe trial, mice treated with EA-ST36 spent more time in the target quadrant and crossed the platform area more frequently than those treated with Sham-EA (Figure 1C,E,F). Spatial working memory was assessed through the Y-maze task. HFD-fed 3xTg-AD mice showed a lower spontaneous alternation percentage compared to mice on a normal diet, suggesting HFD-induced memory impairment. Importantly, EA-ST36 treatment significantly restored spatial working memory in HFD-fed 3xTg-AD mice (Figure 1G), while Sham-EA had no effect. No significant differences were observed in the number of arm choices between mice on standard and high-fat diets (Figure 1H). These findings suggest that EA-ST36 can ameliorate HFD-induced deficits in spatial learning and working memory in 3xTg-AD mice.

3.3 HFD Exacerbates Tau Accumulation and NLRP3 Inflammasome Activation in 3xTg-AD Mice

Given that 3xTg-AD mice exhibit both tau and Aβ accumulation in the brain, we explored the mechanisms underlying the exacerbation of the AD phenotype due to HFD by quantifying total tau, phosphorylated tau at Thr181 (AT270), T212/S214, full-length APP (Fl-APP), and C-terminal fragments (CTF-α/β). As expected, 3xTg-AD mice showed elevated levels of total and phosphorylated tau in the hippocampus compared to Non-Tg mice (Figure 2A,B). Similarly, Fl-APP and CTF levels were higher in 3xTg-AD mice than Non-Tg controls (Figure 2A,B). Interestingly, 4 months of HFD exposure did not significantly affect APP/CTF or tau protein levels in 11-month-old mice, consistent with previous studies [6, 27, 28]. To further examine tau pathology, we quantified Aβ and tau phosphorylation levels in the soluble and insoluble fractions of hippocampal tissue. As previously reported [5, 29, 37, 38], HFD exposure significantly increased total insoluble tau and its phosphorylation at AT270 and AT8 in the hippocampus of 3xTg-AD mice (Figure 2C,D). However, HFD exposure did not alter soluble or insoluble Aβ levels (6E10) (Figure 2C–E).

Prior studies have shown NLRP3 inflammasome activation and elevated IL-1β levels in AD patients and AD transgenic mouse models, such as Tau P301S (PS19) and 3xTg-AD mice [39, 40]. While HFD is widely known to worsen brain Aβ and tau pathology, the role of the NLRP3 inflammasome has been less explored. We investigated the effects of HFD on NLRP3 inflammasome activation, observing that HFD exposure activated NF-κB transcription, resulting in increased levels of NLRP3, pro-IL-1β, and ASC. This activation was confirmed by elevated phosphorylated NF-κB-p65 (RELA), NLRP3, pro-/cleaved-CASP1, and cleaved IL-1β levels in the hippocampus of HFD-fed 3xTg-AD mice (Figure 3A,B). Additionally, ASC oligomers and monomers were elevated, indicating NLRP3–ASC complex formation in the hippocampus of HFD-fed 3xTg-AD mice. Together, these findings suggest that HFD-induced cognitive impairments in 3xTg-AD mice are closely associated with increased insoluble tau aggregates and NLRP3 inflammasome activation in the hippocampus.

3.4 EA-ST36 Inhibits Tau Aggregation and NLRP3 Inflammasome Formation in HFD-Exacerbated 3xTg-AD Mice

To investigate the therapeutic effects of EA-ST36 on HFD-aggravated Alzheimer's pathology, we first assessed its impact on tau aggregation. While EA-ST36 had no significant effect on soluble phosphorylated tau levels in HFD-fed 3xTg-AD mice (Figure 4A,B), it markedly reduced sarkosyl-insoluble phosphorylated tau (AT8) aggregates, indicating enhanced clearance of pathological tau species (Figure 4B).

Next, we evaluated EA-ST36's role in modulating NLRP3 inflammasome activation. Hippocampal levels of NLRP3, pro-caspase-1 (pro-CASP1), and pro-/cleaved IL-1β were significantly reduced in EA-ST36-treated mice compared to HFD controls (Figure 5A,B). Furthermore, EA-ST36 diminished ASC dimer/oligomer formation, reflecting suppression of NLRP3–ASC inflammasome assembly (Figure 5B). Sham-EA failed to reduce FL-APP, insoluble tau (AT8), or NLRP3–ASC complexes, confirming the specificity of EA-ST36's effects (Figure S2A–D).

Immunohistochemical analysis further validated these findings, demonstrating reduced NLRP3–ASC colocalization in the hippocampal CA3 region following EA-ST36 treatment (Figure 5C,D). Additionally, EA-ST36 decreased phosphorylated NF-κB-p65 (RELA) levels, suggesting its inhibitory effects on NLRP3 inflammasome activation may be mediated through the NF-κB signaling pathway.

3.5 EA-ST36 Promotes Autophagy-Lysosomal Pathway via TFEB/TFE3 Activation

We then investigated the mechanisms by which HFD exacerbates AD pathology in 3xTg-AD mice. Impaired autophagic flux, characterized by sustained autophagy induction but defective lysosomal clearance, has been observed in HFD-induced nonalcoholic fatty liver disease [41] and brain tissues from 3xTg-AD mice [42]. Additionally, tau aggregates and NLRP3 inflammasome components are known to degrade through TFEB/TFE3-mediated autophagy-lysosome pathways (ALP) [39]. In this study, we explored lysosomal dysfunction and impaired autophagic flux in HFD-fed 3xTg-AD mice by assessing total and phosphorylated TFEB/TFE3 levels and ALP targets, including MAP1LC3B/LC3B, SQSTM1/p62, CTSD (cathepsin D), and LAMP1 in hippocampal lysates. After 4 months of HFD exposure, 3xTg-AD mice displayed significantly increased SQSTM1/p62 and phosphorylated TFEB (p-TFEB S142), with a trend towards elevated LC3B-II expression that did not reach statistical significance (Figure 6A,B). Furthermore, these mice exhibited a marked reduction in mature CTSD (m-CTSD), suggesting inhibition of late-stage autophagy, resulting in autophagosome accumulation and increased LC3-II expression on autophagosome membranes. These findings indicate that the worsening of AD pathology in HFD-fed 3xTg-AD mice may be attributed to deficient autophagic degradation.

Given that tau aggregates and inflammasome components (NLRP3, ASC, and pro-IL-1β) can be degraded through the ALP, we evaluated whether EA-ST36 reduces NLRP3 inflammasome components via autophagic degradation. To assess the impact of EA-ST36 on autophagy markers, we measured LC3B and SQSTM1 levels in Triton X-100-soluble and SDS-soluble hippocampal lysates, following established guidelines [30]. Our findings showed no significant change in LC3B-II levels in the TX-100-soluble fractions between the 3xTg + HFD and 3xTg + HFD + EA-ST36 groups (Figure 7A,B). However, both LC3B-II and SQSTM1 were significantly reduced in the TX-100-insoluble fractions in the 3xTg + HFD + EA-ST36 group (Figure 7A,B). Regarding lysosomal markers, EA-ST36 treatment significantly increased both pro- and mature CTSD levels, though no significant changes were observed in LAMP1 levels (Figure 7C,D). These findings suggest that EA-ST36 promotes lysosomal biogenesis, facilitating degradation of insoluble SQSTM1 and tau aggregates. Further analysis demonstrated that EA-ST36 increased total TFEB levels and significantly elevated total TFE3 levels in the hippocampus of HFD-fed 3xTg-AD mice. Additionally, EA-ST36 treatment significantly reduced phosphorylation of TFEB at S142, a key site regulated by mTORC1 and AMP [43](Figure 7C,D). Collectively, these data suggest that EA-ST36 activates TFEB/TFE3 to promote lysosomal biogenesis in HFD-fed 3xTg-AD mice.

3.6 EA-ST36 Activates TFEB/TFE3 by Inhibiting mTOR/ULK1 Signaling

To further understand how EA-ST36 activates TFEB/TFE3 in HFD-fed 3xTg-AD mice, we examined upstream kinases, including mTORC1 and AKT. Interestingly, EA-ST36 significantly reduced phosphorylation of mTOR and its substrate, ribosomal protein S6 (RPS6), along with ULK1 (S757) in the hippocampus of HFD-fed 3xTg-AD mice (Figure 7E,F). However, EA-ST36 had no effect on AKT phosphorylation or AMPK phosphorylation in the hippocampus of HFD-fed 3xTg-AD mice (Figure 7E,F). These results suggest that EA-ST36 activates TFEB/TFE3 by inhibiting the mTOR/ULK1 signaling pathway.

3.7 EA-ST36 Facilitates Autophagic Degradation of ASC Oligomers

For autophagic degradation, SQSTM1 binds to LC3 to facilitate degradation of ubiquitinated ASC protein aggregates via autophagy [44]. To determine whether HFD-induced ASC degradation is associated with ALP, we evaluated ASC and SQSTM1 colocalization in hippocampal regions of 3xTg mice fed with or without HFD. We observed large ASC specks colocalized with SQSTM1 puncta in the hippocampi of HFD-fed 3xTg mice (Figure 8A,B), indicating accumulation of undigested ASC–SQSTM1 aggregates due to impaired autophagic degradation. Conversely, EA-ST36 treatment significantly reduced ASC area and ASC-associated SQSTM1 (Figure 8A,B), indicating that EA-ST36 promotes autophagic recognition and degradation of ASC oligomers via SQSTM1. These findings suggest that EA-ST36 facilitates degradation of NLRP3 inflammasome components through the TFEB/TFE3-mediated autophagy-lysosomal pathway.

3.8 EA-ST36-Induced Cognitive Improvements Depend on TFEB/TFE3

Lastly, we determined the specific roles of TFEB/TFE3 in EA-ST36-induced NLRP3 inflammasome degradation and memory improvement in HFD-exacerbated 3xTg-AD mice. To achieve this, AAV-sh-Scramble and AAV-sh-Tfeb&Tfe3 were bilaterally injected into hippocampal CA1 regions. Strong GFP signals in the hippocampus confirmed efficient delivery of shRNAs (Figure S1A). However, hippocampal injection of AAV-sh-Tfeb&Tfe3, but not AAV-sh-Scramble, resulted in high mortality rates (~20%) (Figure S2B). Western blots confirmed TFEB&TFE3 knockdown in surviving mice (Figure 9A,B). In TFEB/TFE3 double-knockdown (TFEB&TFE3-KD) mice, EA-ST36 treatment did not reduce NLRP3, cleaved-IL1B, or pro-/cleaved-CASP1 levels (Figure 9A,B). Notably, the reduction of pro-IL1B by EA-ST36 was unaffected by TFEB&TFE3-KD, indicating that TFEB/TFE3-mediated ALP is not essential for pro-IL1B turnover (Figure 9A,B). Importantly, TFEB&TFE3-KD also abolished EA-ST36's ability to reduce phosphorylated tau (AT8) in the hippocampus, as shown in Figure S2, confirming that TFEB/TFE3 activation is critical for both NLRP3 inflammasome degradation and tau clearance. Furthermore, the EA-ST36-induced improvement in spatial learning memory was blocked in TFEB&TFE3-KD mice (Figure 9C–F). These results indicate that TFEB/TFE3 are required for EA-ST36-induced degradation of NLRP3 inflammasome components and memory improvement. This dual mechanism—reducing neuroinflammation and tauopathy—likely synergizes to mitigate HFD-induced cognitive deficits (Figure 10).