An Engineered Multifunctional Fusion Protein Targeting Aβ Oligomers, Microglia and Autophagy Ameliorates Cognitive Deficits and Amyloid Pathology in Alzheimer’s Disease Mice Authors: Jie Zhu, Ya-ru Huang, Fei Chen, Mei-qi Wang, Xiao-lin Yu, Gui-feng Zhang, Rui-tian Liu Published: 2026 ABSTRACT In Alzheimer’s disease (AD), Amyloid-β (Aβ) oligomers function as key neurotoxic agents that underpin the disease’s progression. A diverse array of therapeutic entities, including peptides, single-chain variable fragments (scFvs), and small molecules, have demonstrated the ability to interact with Aβ oligomers, thereby suppressing their aggregation and associated neurotoxicity. Despite these advances, such agents frequently struggle to promote the phagocytosis and subsequent breakdown of aggregated Aβ by microglia. Moreover, the dense accumulation of Aβ oligomers may resist enzymatic hydrolysis within the acidic lysosomal lumen, contributing to lysosomal stress and dysfunction. To overcome these problems, we engineered a multifunctional fusion protein, p62-LIR–W20–Tuftsin (W20-LT), consisting of an oligomer-specific scFv, a microglia-targeting Tuftsin peptide, and a p62-LIR peptide to activate autophagy. In vitro assays demonstrated that W20-LT significantly outperformed the parental W20 by promoting the rapid microglial uptake of Aβ oligomers and enhancing their intracellular clearance through an autophagy-associated pathway. In APPswe/PS1dE9 (APP/PS1) mice, a low-dose regimen (0.5 µg, every 3 days) of W20-LT, but not W20, significantly ameliorated cognitive deficits and reduced amyloid pathology. Mechanistically, W20-LT was associated with enhanced autophagy-lysosomal pathway activity, as indicated by increased LC3B-II and reduced p62 levels, together with downregulated CatD and LAMP1 levels, thereby mitigating neuroinflammation. In summary, our findings suggest that W20-LT represents a promising proof-of-concept therapeutic strategy that combines scFv-based Aβ oligomer recognition with enhanced autophagy-associated clearance, thereby mitigating AD pathology.Graphical Abstract Schematic of the W20-LT-mediated aggrephagy strategy. The fusion protein W20-LT drives microglial uptake of Aβ oligomers via Neuropilin-1. Utilizing a p62-LIR motif to recruit LC3-II, it directs cargo into autophagosomes. This aggrephagy alleviates lysosomal stress, reduces inflammatory signaling, and promotes Aβ clearance and neuroprotection in AD mice. Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1007/s10571-026-01753-3. Introduction Significant pressure is exerted on global medical systems by Alzheimer’s disease (AD), which remains the most pervasive neurodegenerative pathology and the foremost cause of dementia globally (Selkoe and Hardy 2016). Neuropathologically, the disease is distinguished by two cardinal hallmarks: the external accumulation of Aβ-derived plaques and the cytoplasmic clumping of tau-based neurofibrillary tangles. Among the various amyloid species, soluble amyloid-β (Aβ) oligomers are increasingly recognized as the primary neurotoxins and pivotal drivers of synaptic failure and neuronal loss, which exhibits a robust correlation with the extent of cognitive impairment (Selkoe and Hardy 2016; Haass and Selkoe 2007). Consequently, strategies targeting the elimination of these toxic oligomers remain at the forefront of therapeutic development. Acting as the principal immune sentinels of the central nervous system (CNS), microglia play a fundamental role in the recognition and elimination of Aβ deposits. Under physiological conditions, microglia internalize Aβ primarily through receptor-mediated phagocytosis, directing the cargo to lysosomes for degradation (Whyte et al. 2017). However, in the AD brain, this clearance mechanism is frequently overwhelmed. Evidence suggests that the dense accumulation of Aβ oligomers may resist enzymatic hydrolysis within the acidic lysosomal lumen, contributing to lysosomal swelling, leakage of cathepsins, and ultimately lysosomal storage dysfunction (Nixon 2013; Whyte et al. 2017). Notably, although the autophagy-lysosomal pathway (ALP) serves as the principal cellular mechanism for catabolizing large protein aggregates (aggrephagy), extracellular Aβ ingested via phagocytosis is not intrinsically routed to this autophagic machinery (Holliger and Hudson 2005). Instead, it accumulates in the endocytic-lysosomal compartments, exacerbating cellular stress and neuroinflammation. Therefore, a therapeutic strategy that not only neutralizes Aβ toxicity but also promotes the routing of these aggregates toward the autophagy-lysosomal pathway could relieve lysosomal pressure and improve microglial degradation (Zhao et al. 2014). To date, various therapeutic agents, including peptides, single-chain variable fragments (scFv), and small molecule compounds, have been reported to bind Aβ oligomers effectively (Dong et al. 2018; Zha et al. 2016). scFvs, in particular, act as potent “neutralizers” that inhibit Aβ aggregation and block its neurotoxicity. Compared to full-length antibodies, scFvs offer a significant safety advantage by avoiding Fc receptor-mediated pro-inflammatory side effects (Bard et al. 2000). In our prior investigations, we characterized W20, a pan-amyloid oligomer-specific scFv capable of recognizing toxic conformations and mitigating neuropathological deficits in murine models of AD and Amyotrophic Lateral Sclerosis (ALS) (Dong et al. 2018; Zha et al. 2016; Zhao et al. 2014). However, a fundamental limitation shared by these “passive” binding agents—including peptides and scFvs—is that they lack intrinsic effector domains to actively drive the cellular degradation of the captured toxins (Bard et al. 2000; Najjar and Nishioka 1970). Without a mechanism to accelerate clearance, neutralized complexes may persist or be released, potentially failing to resolve the core issue of microglial lysosomal dysfunction. To overcome this, we hypothesized that coupling the specific recognition of scFvs with the degradation capacity of autophagy could confer these immunotherapies with “active clearance” capabilities. We engineered a tripartite fusion protein, p62-LIR–W20–Tuftsin (W20-LT), designed to act as a molecular bridge. W20-LT integrates three functional modules: the W20 scFv for Aβ oligomer recognition, a C-terminal Tuftsin peptide for enhanced microglial targeting via Neuropilin-1 (Nissen and Tsirka 2013; Teesalu et al. 2009), and an N-terminal p62-LIR motif designed to engage the autophagic machinery (Ichimura et al. 2013; Pankiv et al. 2007). By facilitating binding to LC3-II, W20-LT is designed to promote the autophagy-associated processing of internalized Aβ complexes, thereby potentially alleviating pressure on the phagocytic-lysosomal system. In this study, we evaluated the ability of W20-LT to enhance the uptake and intracellular clearance of Aβ oligomers, and examined whether it could improve cognitive function and alleviate amyloid-associated pathology in the APPswe/PS1dE9 (APP/PS1) mouse model. Materials and Methods Engineering and Production of the W20-LT Fusion Protein The recombinant fusion protein, W20-LT, was constructed using the Aβ oligomer-specific scFv (clone W20) as the core recognition unit (Zhao et al. 2014). An engineered acidic cluster derived from the p62/SQSTM1 LIR motif (DDDDDWTHL) (Ichimura et al. 2013), designed to mimic the electrostatic environment and binding affinity of the phosphorylated state, was conjugated to the N-terminus, whereas the C-terminus was functionalized with a microglia-targeting Tuftsin peptide (TKPR) (Nissen et al. 2013).To ensure adequate spatial separation and full exposure of the functional domains, flexible (G4S)3 linkers were inserted at both junctions. A termination codon was inserted directly downstream of the Tuftsin motif to ensure the presence of a free C-terminal arginine, a modification requisite for satisfying the CendR rule and facilitating NRP1 binding (Teesalu et al. 2009). To generate a matched negative control, W20-nsT was constructed using the same backbone and C-terminal Tuftsin motif as W20-LT, but the N-terminal p62-LIR sequence was replaced with a designed non-functional control peptide (DDDDDATHA), which preserves peptide length and the overall acidic character while disrupting the LIR-like binding determinant. The coding sequences of W20-LT and W20-nsT was subcloned into the pET-22b (+) expression plasmid, followed by protein production in the E. coli BL21 (DE3) host strain. Following IPTG induction at 22 °C, the protein was extracted from the periplasm and purified using a Protein L affinity chromatography column (Capto L, GE Healthcare), capable of selectively interacting with the kappa light chain region of the scFv. The purity and molecular weight were verified by SDS-PAGE. To facilitate in vivo tracking, the purified W20-LT protein was fluorescently labeled using the near-infrared dye Cy5-NHS ester (GE Healthcare), following the manufacturer’s recommended instructions (Mujumdar et al. 1993). Biophysical Characterization of Recombinant Proteins Following the methodology outlined in a previous study (Stine et al. 2011), we prepared the soluble Aβ1–42 oligomeric species. Lyophilized Aβ1–42 peptide (Zhongtai Biochemicals, Hangzhou, China) was dissolved in HFIP, evaporated to dryness, and subsequently dissolved in DMSO to achieve a final concentration of 5 mM. Oligomer formation was initiated by diluting the peptide stock to a final concentration of 100 µM in cold PBS, followed by incubation at 4 °C for 24 h. The binding affinity of W20 and W20-LT to these Aβ oligomers was determined by Enzyme-Linked Immunosorbent Assay (ELISA). Briefly, microplate wells were coated with Aβ oligomers (1 µg/well), and gradient dilutions of the W20-LT or W20 were added. Immunodetection was performed utilizing a horseradish peroxidase (HRP)-conjugated antibody directed against the Myc epitope. Additionally, the biophysical properties of the W20-LT were characterized. To ensure the quality of the fusion protein, the monomeric homogeneity and purity of W20-LT were first verified via Size Exclusion Chromatography (SEC) using a TSKgel G3000SWxl column (Tosoh Bioscience). Subsequently, its biophysical profiles, including hydrodynamic diameter and intrinsic fluorescence distribution, were comprehensively characterized on an Uncle all-in-one platform (Unchained Labs). Cell Culture and Models SH-SY5Y Cell Culture SH-SY5Y human neuroblastoma cells were supplied by Procell Life Science & Technology Co., Ltd. (Wuhan, China). The identity of the cell line was authenticated using Short Tandem Repeat (STR) profiling analysis, and the cultures were routinely confirmed to be negative for mycoplasma contamination. Cells were maintained in DMEM/F12 (Thermo Fisher Scientific, Cat# 11330-032) supplemented with 10% Fetal Bovine Serum (FBS; Gibco, Cat# A5256701) and 1% penicillin-streptomycin (Beyotime, Cat# C0222) and in a humidified incubator at 37 °C with 5% CO2. The culture medium was refreshed daily. Upon reaching approximately 90% confluence, cells were washed with sterile phosphate-buffered saline (PBS), dissociated using trypsin-EDTA (Beyotime, Cat# C0203), and centrifuged at 800 rpm for 5 min. The cell pellet was resuspended and subcultured at a split ratio of 1:2 All experiments were performed using cells within passages 5–20 to ensure phenotypic stability. Primary Microglial Culture Primary microglial cells were isolated from the cerebral cortices of postnatal day 0–2 (P0–P2) C57BL/6 mice following previously established protocols (Saura et al. 2003). Cortical tissues were dissociated and cultured in DMEM/F12 (Gibco, Cat# 11330-032) supplemented with 10% FBS (Gibco, Cat# A5256701) and 1% penicillin-streptomycin (Beyotime, Cat# C0222) and in a humidified incubator at 37 °C with 5% CO2 to establish mixed glial cultures. Every three days, a 50% volume medium exchange was performed. Following a maintenance period of 10–14 days, microglia were detached and collected by subjecting the culture flasks to mechanical agitation at 200 rpm for 2 h. The purity of the isolated microglia was confirmed to be > 95% by Iba1 immunofluorescent staining. In Vitro Cytotoxicity and Functional Assays of Recombinant Proteins For cytotoxicity assessment, SH-SY5Y cells were treated with W20 or W20-LT at a high concentration of 27 µM (1 mg/mL) for 24 h, followed by the evaluation of cell survival using the standard MTT protocol (n = 6 biological replicates per group) according to the established protocol (Mosmann 1983). For phagocytosis and degradation assays, primary microglia were incubated with Aβ oligomers (0.5 µM) and engineered proteins (0.5 µM) for 0–2 h (uptake phase). Intracellular Aβ levels were assessed by Dot blot and Immunocytochemistry as described in the respective sections (n = 6 biological replicates). Following the uptake phase, cells were washed to remove extracellular oligomers and incubated in fresh medium for an additional 0–24 h (degradation phase), after which residual intracellular Aβ was quantified by Dot blot. To evaluate autophagy pathway activation and inflammatory responses, in a separate experiment, microglia were treated with Aβ oligomers alone or pre-formed Aβ-antibody complexes (1:1 molar ratio, final concentration 0.5 µM) for 24 h. After treatment, cell lysates were prepared for immunoblotting of autophagy-related proteins, and culture supernatants were collected for ELISA measurement of TNF-α and IL-1β, as described in the corresponding sections below. Animals and Ethics Statement Male APP/PS1 transgenic mice, their wild-type (WT) littermates, and neonatal C57BL/6 pups (P0–P2, mixed sex) were supplied by Beijing HFK Bioscience Co., Ltd. (Beijing, China). All animals were received with appropriate quarantine documentation to ensure health status. Subsequently, they were maintained within an SPF-grade unit, where environmental factors including a 12-h light/dark cycle, temperature, and humidity were stringently regulated. Ex Vivo Imaging for Brain Distribution and Microglial Targeting Assessment A single intracerebroventricular (ICV) dose of Cy5-conjugated W20-LT (0.5 µg per mouse) was administered to 8-month-old APP/PS1 mice to investigate its cerebral retention characteristics. The surgical steps and stereotaxic parameters utilized are elaborated in the subsequent section.Mice were euthanized at designated intervals spanning from 2 h to 5 days post-administration, after which brain tissues were collected and preserved in 4% paraformaldehyde (PFA). To facilitate post-mortem fluorescence visualization, fixed whole brains were sectioned into four sagittal blocks. Imaging was conducted on an Amersham ImageQuant 800 apparatus, with the integrated fluorescence levels subsequently calculated via ImageJ (Schneider et al. 2012). After background subtraction, data were normalized to the peak intensity at 24 h (Day 1) and expressed as the percentage of maximum intensity (% of Max Intensity). To assess the specific targeting of W20-LT to microglia in vivo, a separate cohort of mice was administered the same dose of W20-LT. Brains were collected 24 h post-injection and processed for paraffin embedding. The co-localization of W20-LT with Iba1-positive microglia was subsequently analyzed via anti-Myc immunofluorescence, as detailed in the Immunohistochemistry section. Therapeutic Regimen and Administration For the evaluation of therapeutic efficacy, male APP/PS1 mice (8 months old) and their WT littermates were first subjected to stereotaxic surgery. Under isoflurane-induced anesthesia, a guide cannula was surgically positioned into the lateral ventricle using an RWD stereotaxic apparatus (Shenzhen, China) based on established coordinates (AP: −0.32 mm; ML: +1.2 mm; DV: −2.3 mm) (Paxinos and Franklin 2001). Following surgery, the APP/PS1 mice were allocated into three cohorts (n = 8 mice per group): AD-PBS (vehicle), AD-W20, and AD-W20-LT. Similarly, WT mice were divided into WT-PBS and WT-W20-LT control groups (n = 8). Treatment involved the administration of W20-LT (0.5 µg in 2 µL of PBS) or an equal volume of vehicle every three days over a 4-week period. To prevent backflow, the injection needle remained in situ for 5 min post-delivery. Behavioral assessments were initiated after a 5-day post-surgical recovery phase. Behavioral Assessments Behavioral tests commenced 5 days after the last drug administration. All tests were conducted in a blinded manner to ensure unbiased scoring (n = 8 mice per group). Morris Water Maze Spatial navigational abilities and long-term memory retention were evaluated utilizing the Morris Water Maze (MWM) paradigm (Morris 1984). The spatial acquisition training was conducted over 5 consecutive days, with two trials administered daily. In each trial, mice were allowed 60 s to locate the hidden submerged platform (n = 8 mice per group). If a mouse failed to find the platform within this time limit, it was manually guided to the platform and allowed to remain there for 2 min to reinforce spatial orientation. Following the training period, mice were allowed to rest on day 6. To evaluate memory retention on day 7, a probe trial was performed by withdrawing the platform. The animals were then allowed to explore the pool for a 60-second duration without the target. The time spent in the target quadrant and the number of platform crossings were recorded. Novel Object Recognition Recognition memory was evaluated using the Novel Object Recognition (NOR) test (Antunes and Biala 2012). During the initial familiarization phase on day one, animals were introduced into an open-field enclosure containing a pair of identical objects and permitted to explore the environment for a set duration. After a 24-h retention interval, one of the familiar objects was replaced with a novel object of a different shape and texture. In the test phase, mice were allowed to explore the objects, and the number of physical contacts (sniffing or touching) with each object was recorded. The discrimination index was calculated as the ratio of the number of touches on the novel object to the total number of touches on both objects (n = 8 mice per group). Y-Maze Test Short-term spatial memory and spontaneous locomotor activity were evaluated employing a modified version of the Y-Maze paradigm (Kraeuter et al. 2019). During the training phase, one arm (designated as the “Novel Arm”) was blocked with a removable partition. Mice were placed in the “Start Arm” and allowed to explore the remaining two open arms for 10 min. After a 2-h inter-trial interval, the partition was removed to open the Novel Arm. During the testing phase (5 min), mice were placed back in the maze with access to all three arms. The time spent in the Novel Arm and the number of entries into it were analyzed to evaluate spatial recognition memory. (n = 8 mice per group) Protein Extraction from Tissues and Cells Tissue Processing and Harvesting Following the conclusion of behavioral testing, following deep anesthesia with tribromoethanol, the mice were systemically perfused via the heart with chilled PBS containing heparin for 5 min to flush the circulatory system. The brain tissues were promptly dissected and divided into two hemispheres via a mid-sagittal incision. The left hemisphere was used for histological analysis and the right hemisphere was employed for biochemical testing. The hippocampus was dissected from the right hemisphere for biochemical analysis. Tissues were mechanically homogenized in RIPA lysis buffer (Beyotime, Cat# P0013B) supplemented with a protease inhibitor cocktail (Solarbio, Cat# IKM1130). The resulting homogenates were subjected to centrifugation at 13,000 × g for 30 min at 4 °C, after which the supernatant was harvested as the RIPA-soluble fraction, comprising detergent-soluble proteins. To extract insoluble aggregates, the resulting pellet was resuspended in 6 M guanidine hydrochloride, incubated, and centrifuged again at 13,000 × g for 30 min at 4 °C to obtain the guanidine-soluble fraction (containing insoluble proteins). Cell Protein Extraction Prior to downstream assays, the cultured primary microglia were rinsed thrice with chilled PBS to ensure the complete removal of residual culture medium. The cells were lysed on ice using the same RIPA lysis buffer supplemented with the protease inhibitor cocktail. The lysates were harvested and incubated on ice for 30 min with occasional vortexing. Subsequently, the samples were centrifuged at 13,000 × g for 30 min at 4 °C to remove cellular debris. The resulting supernatants were harvested, distributed into small aliquots, and maintained at −80 °C to preserve sample stability for subsequent analyses. Western Blot and Dot Blot Analysis Proteins extracted from cell lysates or brain homogenates were analyzed to evaluate protein expression and aggregation states (n = 6 mice per group). The specificity of the primary antibodies used in this study was validated based on manufacturer-provided data (including knockout/knockdown validation and specific peptide blocking assays) or referenced from established literature. For immunoblotting, equivalent protein quantities (15 µg) were resolved via SDS-PAGE (4–12%) and electroblotted onto nitrocellulose (NC) membranes. After blocking in PBST containing 5% non-fat milk for 1 h at ambient temperature, the membranes were probed overnight at 4 °C with the primary antibodies listed below: p62 (Abcam, Cat# ab109012, RRID: AB_10860015, 1:1,000), LC3B (Cell Signaling Technology, Cat# 2775, RRID: AB_915950, 1:1,000), LAMP1 (Abcam, Cat# ab24170, RRID: AB_448322, 1:1,000), Cathepsin D (Beyotime, Cat# AF1226, 1:1,000), PSD95 (Abcam, Cat# ab18258, RRID: AB_444362, 1:1,000), NeuN (Abcam, Cat# ab177487, RRID: AB_2532109, 1:1,000), GFAP (Cell Signaling Technology, Cat# 12389, RRID: AB_2631098, 1:1,000), Iba1 (Fujifilm Wako, Cat# 019–19741, RRID: AB_839504, 1:1,000), and β-actin (Abcam, Cat# ab8226, RRID: AB_306371, 1:2,000). After washing three times with PBST, the membranes were incubated with HRP-conjugated goat anti-mouse (Abcam, Cat# ab6789, AB_955439, 1:5,000) or anti-rabbit (Abcam, Cat# ab6721, AB_955447, 1:5,000) secondary antibodies for 1 h at room temperature. Protein bands were visualized using an ECL detection kit (Thermo Fisher Scientific, Cat# 34580) and imaged using an Amersham ImageQuant 800 system. For Dot blot analysis, 2 µg of each protein sample was spotted directly onto NC membranes. Once dry, membranes were pre-treated with 5% non-fat milk in PBST for blocking and subsequently exposed to the following primary antibodies overnight at 4 °C: antibodies against Aβ17–24 (4G8) (BioLegend, Cat# 800701, RRID: AB_2564633, 1:1,000), amyloid oligomers (A11) (Millipore, Cat# AB9234, RRID: AB_570698, 1:1,000), amyloid fibrils (OC) (Millipore, Cat# AB2286, RRID: AB_1977024, 1:1,000), and β-actin (Abcam, Cat# ab8226, RRID: AB_306371, 1:2,000). The membranes were then washed and incubated with the corresponding HRP-conjugated secondary antibodies. Signals were visualized using ECL reagents (Thermo Fisher Scientific) and quantified using ImageJ software. Enzyme-Linked Immunosorbent Assay (ELISA) To assess the inflammatory response and amyloid pathology, the levels of proinflammatory cytokines and amyloid-beta peptides were quantified using commercial ELISA kits (n = 6 mice per group). Specifically, Mouse TNF-α (Abclonal, Cat# RK04875), IL-1β (Solarbio, Cat# SEKM-0002), Aβ1–40 (IBL, Cat# 27718), and Aβ1–42 (IBL, Cat# 27719) were measured. For sample preparation, microglial culture supernatants were collected and centrifuged after 24 h of treatment for in vitro assessments. For in vivo studies, the RIPA-soluble fractions of hippocampal homogenates were utilized, and the total protein concentration was determined using the BCA Protein Assay Kit prior to the assay. All ELISA procedures were performed strictly according to the manufacturers’ instructions, with absorbance measured at 450 nm using a microplate reader. For in vivo samples, the concentrations of TNF-α, IL-1β, Aβ1–40, and the levels of Aβ were corrected for the total protein concentration, with results quantified as picograms per milligram (pg/mg) of total protein. Immunocytochemistry and Immunohistochemistry Immunocytochemistry Cultured cells grown on coverslips were fixed with 4% PFA for 15 min at 25 ℃. After washing 3 times with PBST, cells were blocked with 5% donkey serum in PBS containing 0.1% Triton X-100 for 1 h. The specificity of primary antibodies was validated as described in the Western blot section. The coverslips were then incubated overnight at 4 °C with primary antibodies: anti-Aβ (4G8) (BioLegend, Cat# 800701, RRID: AB_2564633, 1:1,000) and anti-Iba1 (Fujifilm Wako, Cat# 019–19741, RRID: AB_839504, 1:500). Subsequent to primary antibody binding, cells were rinsed thoroughly and then labeled with an Alexa Fluor 488-conjugated donkey anti-rabbit secondary antibody (1:500, Thermo Fisher Scientific, Cat# A-21206, RRID: AB_2535792). This incubation was carried out for 60 min at ambient temperature. To identify cell nuclei, a Hoechst 33,342 counterstain (1:5,000) was applied (n = 6 mice per group). Immunohistochemistry After being fixed by immersion in 4% PFA for 24 h at 4–8℃, brain tissues were processed through sequential ethanol concentrations and xylene for dehydration and clearing, respectively. Once embedded in paraffin blocks, the tissues were sliced into 10 μm serial sections. To prepare for analysis, sections were deparaffinized, rehydrated, and subjected to antigen unmasking by boiling in citrate buffer (pH = 6.5) for a duration of 20 min. The sections were subsequently subjected to blocking and permeabilization in 10% donkey serum supplemented with 0.1% Triton X-100 for 1 h at ambient temperature. Subsequently, sections were incubated overnight at 4 °C with the following primary antibodies: anti-Iba1 (Fujifilm Wako, Cat# 019–19741, RRID: AB_839504, 1:500), anti-Myc tag (Abcam, Cat# ab32, RRID: AB_303599, 1:1,000), anti-NeuN (Abcam, Cat# ab177487, RRID: AB_2532109, 1:500), anti-PSD95 (Abcam, Cat# ab18258, RRID: AB_444362, 1:500), and anti-GFAP (Cell Signaling Technology, Cat# 12389, RRID: AB_2631098, 1:500). For the detection of amyloid fibrils, sections were stained with 0.01 mg/mL Thioflavin S (ThS) (Sigma-Aldrich, Cat# T1892). Subsequent to rinsing with 1% PBST, the tissue sections were treated with appropriate fluorophore-tagged secondary antibodies (Alexa Fluor series, Thermo Fisher Scientific) for immunofluorescent detection or processed for DAB colorimetric detection using a DAB detection kit (Zhongshan Golden Bridge Biotechnology) (n = 6 mice per group). Image Acquisition and Analysis Fluorescence visualization was performed on a Leica TCS SP8 confocal laser scanning platform, which was configured with a 63× oil-immersion lens (NA = 1.4) or a 20× dry objective. All confocal imaging was performed at room temperature. Excitation and emission wavelengths were set according to the specific fluorophores used (e.g., 405 nm for Hoechst, 488 nm for Alexa Fluor 488/ThS, and 561 nm for Alexa Fluor 594). Both the capture and subsequent analysis of images were managed through the LAS X software environment (Leica Microsystems).To visualize the DAB-labeled sections, brightfield micrographs were acquired on an Olympus IX73 inverted microscope system coupled with a DP80 digital camera. Images were acquired using 4× and 40× objectives at 25 °C. Quantitative analysis for all images was performed using ImageJ software. 3D morphological reconstruction of microglia was performed using Imaris 9.9 software (Bitplane). Nissl Staining To evaluate neuronal loss and structural integrity, brain sections were stained using a Nissl Staining Kit (Solarbio, Cat# G1430) according to the manufacturer’s instructions (n = 6 mice per group). Briefly, tissue sections underwent Nissl staining using Cresyl Violet solution, followed by differentiation and dehydration steps. Micrographs were acquired at 25℃on an Olympus IX73 inverted microscope fitted with a DP80 camera and a 40× objective. To assess neuronal viability, surviving neurons in the hippocampal CA1 subfield were counted across three non-overlapping, randomly selected regions per section using ImageJ (Schneider et al. 2012). To eliminate subjective bias, the entire analysis was conducted in a blinded fashion, with the investigator unaware of the experimental groupings. Statistical Analysis To provide a clear view of the data spread, individual observations were plotted over bar graphs representing the mean ± SD. The number of technical replicates for molecular and histological assays was maintained at ≥ 3. Sample sizes for statistical analyses were maintained at n ≥ 6 to ensure adequate power; specifically, behavioral experiments included n = 8 mice per group, and biochemical or histological assays included n = 6 independent biological replicates. Sample size sufficiency was confirmed using pilot power analyses (power > 0.8, α = 0.05). To minimize subjective bias, animals were randomly assigned to experimental groups, and investigators remained blinded to group allocation during both experimental procedures and statistical analysis. The Gaussian distribution of all data was examined through the Shapiro–Wilk test. Additionally, Levene’s procedure was employed to substantiate the homogeneity of variance across the experimental groups. Statistical tests were selected based on the following criteria: Non-Parametric Tests: To ensure the robustness of our findings, non-parametric methodologies were implemented whenever datasets deviated from a normal distribution or displayed non-uniform variances. For comparisons between two independent cohorts, the Mann–Whitney U test was utilized. In contrast, multi-group evaluations were conducted via the Kruskal–Wallis test, with subsequent Dunn’s post hoc analysis to pinpoint specific inter-group differences.Parametric Tests: Parametric tests were employed only when data satisfied both normal distribution and homogeneity of variance assumptions. Two-group comparisons: Unpaired Student’s t-test.Multiple-group comparisons: One-way analysis of variance (ANOVA) was used for single-factor comparisons. Two-way ANOVA was utilized for comparisons involving two independent variables (e.g., Treatment × Time or Treatment × Genotype). Post hoc tests (Tukey’s for one-way, Bonferroni’s for two-way) were performed to determine specific group differences. Non-Parametric Tests: To ensure the robustness of our findings, non-parametric methodologies were implemented whenever datasets deviated from a normal distribution or displayed non-uniform variances. For comparisons between two independent cohorts, the Mann–Whitney U test was utilized. In contrast, multi-group evaluations were conducted via the Kruskal–Wallis test, with subsequent Dunn’s post hoc analysis to pinpoint specific inter-group differences. Non-Parametric Tests: To ensure the robustness of our findings, non-parametric methodologies were implemented whenever datasets deviated from a normal distribution or displayed non-uniform variances. For comparisons between two independent cohorts, the Mann–Whitney U test was utilized. In contrast, multi-group evaluations were conducted via the Kruskal–Wallis test, with subsequent Dunn’s post hoc analysis to pinpoint specific inter-group differences. Parametric Tests: Parametric tests were employed only when data satisfied both normal distribution and homogeneity of variance assumptions. Two-group comparisons: Unpaired Student’s t-test.Multiple-group comparisons: One-way analysis of variance (ANOVA) was used for single-factor comparisons. Two-way ANOVA was utilized for comparisons involving two independent variables (e.g., Treatment × Time or Treatment × Genotype). Post hoc tests (Tukey’s for one-way, Bonferroni’s for two-way) were performed to determine specific group differences. Parametric Tests: Parametric tests were employed only when data satisfied both normal distribution and homogeneity of variance assumptions. Two-group comparisons: Unpaired Student’s t-test.Multiple-group comparisons: One-way analysis of variance (ANOVA) was used for single-factor comparisons. Two-way ANOVA was utilized for comparisons involving two independent variables (e.g., Treatment × Time or Treatment × Genotype). Post hoc tests (Tukey’s for one-way, Bonferroni’s for two-way) were performed to determine specific group differences. Two-group comparisons: Unpaired Student’s t-test. Two-group comparisons: Unpaired Student’s t-test. Multiple-group comparisons: One-way analysis of variance (ANOVA) was used for single-factor comparisons. Two-way ANOVA was utilized for comparisons involving two independent variables (e.g., Treatment × Time or Treatment × Genotype). Post hoc tests (Tukey’s for one-way, Bonferroni’s for two-way) were performed to determine specific group differences. Multiple-group comparisons: One-way analysis of variance (ANOVA) was used for single-factor comparisons. Two-way ANOVA was utilized for comparisons involving two independent variables (e.g., Treatment × Time or Treatment × Genotype). Post hoc tests (Tukey’s for one-way, Bonferroni’s for two-way) were performed to determine specific group differences. Statistical processing was executed using GraphPad Prism 9.0 or Python-based computational tools. We have provided precise P values for significant findings within the figures; notably, for any results where P > 0.001, the exact values are listed, whereas those below this threshold are denoted as P < 0.001. A comprehensive breakdown of statistical metrics, including test statistics, degrees of freedom (df), and APA-compliant P values, can be found in Supplementary Table S1. Results Rational Design, Biophysical Characterization, and In Vivo Targeting of W20-LT To engineer a tripartite fusion protein capable of simultaneously facilitating microglial clearance and neutralizing Aβ toxicity, we constructed W20-LT by fusing a phosphomimetic p62-LIR motif and a microglia-targeting Tuftsin peptide to the N- and C-termini of the Aβ oligomer-specific scFv (W20), respectively (Fig. 1A). SDS-PAGE analysis confirmed a molecular weight shift from approximately 30 kDa (parental W20) to 35 kDa (W20-LT), indicating the successful integration of functional peptides and high purity of the recombinant protein (Fig. 1B). SEC-HPLC displayed a single symmetrical peak with a retention time of 10.2 min, indicating that W20-LT exists primarily as a monomer (Fig. 1C). DLS further confirmed a uniform size distribution with an average hydrodynamic diameter of approximately 5 nm (Fig. 1D). Importantly, ELISA results demonstrated that W20-LT retained dose-dependent binding affinity to Aβ oligomers comparable to that of the parental W20, suggesting that the terminal modifications did not compromise antigen recognition (Fig. 1E). Furthermore, no significant cytotoxicity was observed in SH-SY5Y cells treated with W20-LT at concentrations up to 1 mg/mL (Fig. 1F). Following biophysical characterization, we evaluated the brain retention and cell-specific targeting of W20-LT in APP/PS1 mice. Ex vivo fluorescence imaging of sagittal brain sections following ICV injection of Cy5-labeled W20-LT revealed robust signals distributed throughout the brain parenchyma (Fig. 1G). Quantitative analysis showed that fluorescence signals rapidly distributed within the brain parenchyma, reaching a peak intensity at 24 h post-injection. Unlike conventional scFvs that are often rapidly cleared, W20-LT exhibited an extended retention profile, maintaining approximately 56% of its maximal intensity at day 3, before declining to roughly 22% by day 5 (Fig. 1H). Based on this decay kinetics, the effective half-life in the brain is estimated to be approximately 2 days (48 h). This pharmacokinetic profile provides a strong rationale for the 3-day dosing interval employed in subsequent efficacy studies (Fig. 3A), as this regimen ensures the maintenance of steady-state therapeutic levels while preventing excessive accumulation. To experimentally verify the cellular basis for this retention, immunofluorescence staining was performed 24 h post-injection. The results demonstrated extensive colocalization between W20-LT and the microglial marker Iba1 (Fig. 1I), suggesting that the fusion protein is preferentially associated with microglia within the brain tissue. Fig. 1The design, characterization and in vivo distribution of the multifunctional recombinant protein W20-LT. A Schematic representation of the gene construct and prokaryotic expression vector for W20-LT. The p62-LIR peptide was fused to the N-terminus, and the Tuftsin peptide was fused to the C-terminus of the Aβ oligomer-specific scFv (W20), both of which were separated by flexible (G4S)3 linkers. The construct was cloned into the pET-22b (+) vector containing a T7 promoter, a lac operator, and a pelB signal peptide for periplasmic expression. B SDS-PAGE analysis followed by Coomassie Brilliant Blue staining characterizing the molecular weight and purity of W20 and W20-LT proteins (n = 3 biological replicates). C Purity and aggregation analysis of W20-LT via Size Exclusion Chromatography (SEC-HPLC) using a TSKgel G3000SWxl column (n = 3 biological replicates). The x-axis represents retention time (min). D Hydrodynamic size distribution of W20-LT. The size profile was determined by Dynamic Light Scattering (DLS) using an Uncle stability platform (Unchained Labs). The x-axis indicates the hydrodynamic diameter (nm) (n = 3 biological replicates). E Binding affinity of W20 and W20-LT to Aβ oligomers determined by ELISA. Aβ oligomers were coated on the plate, and gradient dilutions of W20 or W20-LT were added to evaluate binding capabilities (n = 3 biological replicates). The x-axis represents the log10 concentration of the proteins. F Cytotoxicity assessment of W20 and W20-LT on SH-SY5Y cells using the MTT assay. Cells were treated with a high concentration (1 mg/mL) of recombinant proteins for 24 h. G Ex vivo fluorescence imaging of Cy5-labeled W20-LT retention in the brain. 8-month-old APP/PS1 mice received an intracerebroventricular (ICV) injection of Cy5-W20-LT (0.5 µg). Brains were harvested at indicated time points (2 h to 5 days), fixed with paraformaldehyde, cut into four sagittal sections, and imaged using an Amersham ImageQuant 800 system (n = 3 biological replicates). H Quantitative analysis of Cy5 fluorescence intensity from panel G (n = 3 mice per group). Data are expressed as fold change relative to day 1. I Immunofluorescence staining showing the colocalization of W20-LT with microglia in the brains of 8-month-old APP/PS1 mice. Brain sections were prepared 24 h after ICV injection of W20-LT (0.5 µg). Green: W20-LT; Red: Iba1 (microglia marker). Scale bar = 25 μm (n = 3 biological replicates). All data are presented as mean ± SD. Statistical analysis was performed using one-way ANOVA followed by Tukey’s post hoc test (F). Exact P values for key comparisons are displayed in the figure; for complete statistical details, please refer to Table S1 The design, characterization and in vivo distribution of the multifunctional recombinant protein W20-LT. A Schematic representation of the gene construct and prokaryotic expression vector for W20-LT. The p62-LIR peptide was fused to the N-terminus, and the Tuftsin peptide was fused to the C-terminus of the Aβ oligomer-specific scFv (W20), both of which were separated by flexible (G4S)3 linkers. The construct was cloned into the pET-22b (+) vector containing a T7 promoter, a lac operator, and a pelB signal peptide for periplasmic expression. B SDS-PAGE analysis followed by Coomassie Brilliant Blue staining characterizing the molecular weight and purity of W20 and W20-LT proteins (n = 3 biological replicates). C Purity and aggregation analysis of W20-LT via Size Exclusion Chromatography (SEC-HPLC) using a TSKgel G3000SWxl column (n = 3 biological replicates). The x-axis represents retention time (min). D Hydrodynamic size distribution of W20-LT. The size profile was determined by Dynamic Light Scattering (DLS) using an Uncle stability platform (Unchained Labs). The x-axis indicates the hydrodynamic diameter (nm) (n = 3 biological replicates). E Binding affinity of W20 and W20-LT to Aβ oligomers determined by ELISA. Aβ oligomers were coated on the plate, and gradient dilutions of W20 or W20-LT were added to evaluate binding capabilities (n = 3 biological replicates). The x-axis represents the log10 concentration of the proteins. F Cytotoxicity assessment of W20 and W20-LT on SH-SY5Y cells using the MTT assay. Cells were treated with a high concentration (1 mg/mL) of recombinant proteins for 24 h. G Ex vivo fluorescence imaging of Cy5-labeled W20-LT retention in the brain. 8-month-old APP/PS1 mice received an intracerebroventricular (ICV) injection of Cy5-W20-LT (0.5 µg). Brains were harvested at indicated time points (2 h to 5 days), fixed with paraformaldehyde, cut into four sagittal sections, and imaged using an Amersham ImageQuant 800 system (n = 3 biological replicates). H Quantitative analysis of Cy5 fluorescence intensity from panel G (n = 3 mice per group). Data are expressed as fold change relative to day 1. I Immunofluorescence staining showing the colocalization of W20-LT with microglia in the brains of 8-month-old APP/PS1 mice. Brain sections were prepared 24 h after ICV injection of W20-LT (0.5 µg). Green: W20-LT; Red: Iba1 (microglia marker). Scale bar = 25 μm (n = 3 biological replicates). All data are presented as mean ± SD. Statistical analysis was performed using one-way ANOVA followed by Tukey’s post hoc test (F). Exact P values for key comparisons are displayed in the figure; for complete statistical details, please refer to Table S1 W20-LT Enhances Microglial Uptake and Promotes Autophagic Degradation of Aβ Oligomers In Vitro To dissect the distinct functional contributions of the microglial-targeting domain (Tuftsin) and the autophagy-directing motif (p62-LIR) in W20-LT, we engineered a control fusion protein, W20-nsT, which retains the C-terminal Tuftsin but contains a non-functional N-terminal peptide instead of the p62-LIR motif. We systematically evaluated the cellular uptake efficiency of these engineered proteins in primary microglia (Fig. 2A). Dot blot analysis of cell lysates revealed that unmodified W20 mediated limited Aβ internalization over the 2-h incubation period, while both W20-nsT and W20-LT significantly increased intracellular Aβ levels (Fig. 2B, C). Notably, no significant difference in uptake efficiency was observed between W20-nsT and W20-LT, which was further corroborated by immunocytochemistry assay performed at the 2-h time point showing extensive colocalization of internalized Aβ with Iba1 in both the W20-nsT and W20-LT groups, whereas the W20 group displayed significantly lower fluorescence intensity (Fig. 2D, E). In addition, we investigated the metabolic fate of the internalized complexes using a pulse-chase degradation assay (Fig. 2F). Microglia were incubated with W20-nsT-Aβ complex or W20-LT-Aβ complex for 2 h to allow uptake, followed by a washout with fresh medium to remove extracellular species, and intracellular Aβ levels were then monitored over time. The W20-LT group exhibited a significantly faster decay of intracellular Aβ signals than the W20-nsT group over the 24-h chase period (Fig. 2G, H), suggesting that the N-terminal p62-LIR motif contributes to the intracellular clearance of the cargo. To verify whether the effect of W20-LT was associated with autophagy-related processing, we performed a separate 24-h treatment experiment in which microglia were exposed to Aβ oligomers alone or pre-formed Aβ-antibody complexes. Under these conditions, W20-LT but not W20 or W20-nsT treatment significantly increased the LC3B-II/I ratio and modulated p62 levels compared to Aβ treatment alone (Fig. 2I–K), consistent with altered autophagy-related processing (Klionsky et al. 2021). To characterize the neuroinflammatory response instigated by Aβ oligomers—which are established stimulators of the NLRP3 pathway (Halle et al. 2008)—we quantified the expression levels of TNF-α and IL-1β. Under the same treatment conditions, ELISA analysis of culture supernatants collected after 24 h showed that W20-LT treatment significantly reduced the levels of TNF-α and IL-1β induced by Aβ oligomers (Fig. 2L, M). In contrast, the levels of TNF-α and IL-1β in the W20-nsT group did not decrease and even exceeded those in the Aβ-only control.. This exacerbation suggests that enhancing microglial uptake without a corresponding degradation mechanism in the W20-nsT group may lead to intracellular accumulation and increased lysosomal stress. Consequently, given the potentially unfavorable “uptake-without-degradation” phenotype, W20-nsT was not included in subsequent in vivo efficacy studies. Collectively, these data suggest that the C-terminal Tuftsin motif enhanced microglial uptake, while the N-terminal p62-LIR motif contributed to the autophagy-associated intracellular clearance of Aβ and was associated with reduced inflammatory responses. Fig. 2Microglial uptake, intracellular degradation, and autophagic clearance of Aβ complexes. A Schematic illustration of the phagocytosis assay. Primary microglia were incubated with Aβ complexes (generated by pre-incubating Aβ oligomers with different engineered proteins: W20, W20-nsT [W20 fused with an N-terminal irrelevant peptide and C-terminal Tuftsin], or W20-LT [W20 fused with the N-terminal p62-LIR and C-terminal Tuftsin]) to evaluate uptake capacity. B Representative dot blot images showing the kinetics of intracellular Aβ uptake. Aβ oligomers (0.5 µM final concentration) were pre-incubated with the proteins (1:1 molar ratio) for 1 h at room temperature to form complexes, then added to microglia for 0–2 h. β-Actin served as the loading control. “0 h” represents samples collected immediately after the addition of Aβ complexes. C Quantitative analysis of intracellular Aβ levels from panel B, normalized to β-actin using ImageJ software. D Immunocytochemistry showing the intracellular distribution of Aβ (green, 4G8) within microglia (red, Iba1) after 2 h of treatment. Scale bar = 25 μm. E Quantification of intracellular Aβ fluorescence intensity from panel D using ImageJ. F Schematic illustration of the degradation assay. G Representative dot blot images showing the degradation of internalized Aβ complexes (W20-nsT vs. W20-LT) in microglial lysates over 0–24 h post-washout. H Quantitative analysis of residual intracellular Aβ levels from panel G, calibrated to β-actin using ImageJ. I Representative Western blot images showing autophagy activation markers (p62 and LC3B) in microglia treated with Aβ oligomers alone or different Aβ complexes for 24 h. β-Actin served as the loading control. J, K Quantitative analysis of p62 protein levels (J) and the LC3B-II/I ratio (K) from panel I using ImageJ. L, M ELISA analysis of proinflammatory cytokines TNF-α (L) and IL-1β (M) in the culture supernatant of microglia treated with Aβ oligomers or different Aβ complexes for 24 h. All data are presented as mean ± SD (n = 6 biologically independent samples). Statistical comparisons were performed using the Kruskal-–Wallis test followed by the Mann-Whitney U test (E, J–M; C at 0 and 1 h; H at 0 and 6 h), one-way ANOVA followed by Tukey’s post hoc test (C at 0.5 and 2 h), or independent sample t-test (H at 12 and 24 h). P values in C indicate comparisons of W20-nsT or W20-LT versus the W20 group at corresponding time points; P values in H indicate comparisons between W20-nsT and W20-LT. Exact P values for key comparisons are displayed in the figure; for complete statistical details, please refer to Table S1 Microglial uptake, intracellular degradation, and autophagic clearance of Aβ complexes. A Schematic illustration of the phagocytosis assay. Primary microglia were incubated with Aβ complexes (generated by pre-incubating Aβ oligomers with different engineered proteins: W20, W20-nsT [W20 fused with an N-terminal irrelevant peptide and C-terminal Tuftsin], or W20-LT [W20 fused with the N-terminal p62-LIR and C-terminal Tuftsin]) to evaluate uptake capacity. B Representative dot blot images showing the kinetics of intracellular Aβ uptake. Aβ oligomers (0.5 µM final concentration) were pre-incubated with the proteins (1:1 molar ratio) for 1 h at room temperature to form complexes, then added to microglia for 0–2 h. β-Actin served as the loading control. “0 h” represents samples collected immediately after the addition of Aβ complexes. C Quantitative analysis of intracellular Aβ levels from panel B, normalized to β-actin using ImageJ software. D Immunocytochemistry showing the intracellular distribution of Aβ (green, 4G8) within microglia (red, Iba1) after 2 h of treatment. Scale bar = 25 μm. E Quantification of intracellular Aβ fluorescence intensity from panel D using ImageJ. F Schematic illustration of the degradation assay. G Representative dot blot images showing the degradation of internalized Aβ complexes (W20-nsT vs. W20-LT) in microglial lysates over 0–24 h post-washout. H Quantitative analysis of residual intracellular Aβ levels from panel G, calibrated to β-actin using ImageJ. I Representative Western blot images showing autophagy activation markers (p62 and LC3B) in microglia treated with Aβ oligomers alone or different Aβ complexes for 24 h. β-Actin served as the loading control. J, K Quantitative analysis of p62 protein levels (J) and the LC3B-II/I ratio (K) from panel I using ImageJ. L, M ELISA analysis of proinflammatory cytokines TNF-α (L) and IL-1β (M) in the culture supernatant of microglia treated with Aβ oligomers or different Aβ complexes for 24 h. All data are presented as mean ± SD (n = 6 biologically independent samples). Statistical comparisons were performed using the Kruskal-–Wallis test followed by the Mann-Whitney U test (E, J–M; C at 0 and 1 h; H at 0 and 6 h), one-way ANOVA followed by Tukey’s post hoc test (C at 0.5 and 2 h), or independent sample t-test (H at 12 and 24 h). P values in C indicate comparisons of W20-nsT or W20-LT versus the W20 group at corresponding time points; P values in H indicate comparisons between W20-nsT and W20-LT. Exact P values for key comparisons are displayed in the figure; for complete statistical details, please refer to Table S1 W20-LT Treatment Ameliorates Cognitive Deficits in APP/PS1 Mice To investigate the impact of W20-LT on cognitive performance, APP/PS1 mice were administered ICV injections of PBS, W20, or W20-LT (0.5 µg per mouse) every three days. After 1 month of therapeutic intervention (Fig. 3A), behavioral testing was conducted, beginning with the MWM to assess spatial acquisition and memory retention. During the training sessions, the AD-PBS group exhibited pronounced learning impairments, evidenced by persistently high escape latencies. In contrast, W20-LT treatment markedly attenuated these deficits, as shown by a significant reduction in escape latency compared to AD-PBS controls (Fig. 3B, C). In the probe trial, W20-LT significantly increased the time spent within the target quadrant (Fig. 3D) and the frequency of platform site crossings (Fig. 3E). Notably, these cognitive gains were not attributed to differences in locomotor activity, as swimming speeds remained uniform across all groups (Fig. 3F). To broaden our evaluation of W20-LT’s cognitive benefits, we performed the NOR and Y-maze paradigms. In the NOR test, the AD-PBS group displayed a diminished discrimination index, reflecting a clear impairment in recognition memory. Conversely, W20-LT administration significantly elevated the discrimination index compared to the vehicle-treated AD mice (Fig. 3G), indicating improved object-recognition performance. Similarly, in the Y-maze test, mice treated with W20-LT spent significantly more time in the “Novel Arm” and showed a higher number of entries compared to AD-PBS mice (Fig. 3H, J, K), while the total distance traveled during the same testing period showed no significant differences among groups (Fig. 3I), suggesting that the observed improvements were not confounded by locomotor hyperactivity or motor deficits. Fig. 3W20-LT treatment ameliorates cognitive deficits in APP/PS1 mice. A Schematic timeline of the experimental design and behavioral testing. Eight-month-old male APP/PS1 (AD) and wild-type (WT) littermates were divided into five groups (n = 8 per group): WT-PBS, WT-W20-LT, AD-PBS, AD-W20, and AD-W20-LT. Mice received intracerebroventricular (ICV) injections (0.5 µg/mouse) every 3 days for 1 month (10 doses total). Behavioral assessments (Morris Water Maze, Novel Object Recognition, and Y-maze) were conducted over 2 weeks following a 5-day rest period after cannula removal. B Representative swimming trajectories of mice on day 5 of the training phase in the Morris Water Maze (MWM) test. C Statistical analysis of the escape latency to find the hidden platform during the 5-day training period. D–F Probe trial performance on day 7 after removing the platform. Percentage of time spent in the target quadrant (D), number of platform crossings (E), and swimming speed (F) were recorded. G Discrimination index in the Novel Object Recognition (NOR) test, calculated as (number of touches on new object/total touches) × 100%. H Representative movement trajectories in the Y-maze test. I–K Quantitative analysis of the Y-maze test showing total distance traveled (I), time spent in the Novel Arm (J), and number of entries into the Novel Arm (K). All data are presented as mean ± SD (n = 8 mice per group). Statistical comparisons were performed using Kruskal–Wallis test followed by Mann-Whitney U test (C, E) or one-way ANOVA followed by Tukey’s post hoc test (D, F, G, I–K). P values in C indicate comparisons between AD-PBS and AD-W20-LT groups at corresponding time points. Exact P values for key comparisons are displayed in the figure; for complete statistical details, please refer to Table S1 W20-LT treatment ameliorates cognitive deficits in APP/PS1 mice. A Schematic timeline of the experimental design and behavioral testing. Eight-month-old male APP/PS1 (AD) and wild-type (WT) littermates were divided into five groups (n = 8 per group): WT-PBS, WT-W20-LT, AD-PBS, AD-W20, and AD-W20-LT. Mice received intracerebroventricular (ICV) injections (0.5 µg/mouse) every 3 days for 1 month (10 doses total). Behavioral assessments (Morris Water Maze, Novel Object Recognition, and Y-maze) were conducted over 2 weeks following a 5-day rest period after cannula removal. B Representative swimming trajectories of mice on day 5 of the training phase in the Morris Water Maze (MWM) test. C Statistical analysis of the escape latency to find the hidden platform during the 5-day training period. D–F Probe trial performance on day 7 after removing the platform. Percentage of time spent in the target quadrant (D), number of platform crossings (E), and swimming speed (F) were recorded. G Discrimination index in the Novel Object Recognition (NOR) test, calculated as (number of touches on new object/total touches) × 100%. H Representative movement trajectories in the Y-maze test. I–K Quantitative analysis of the Y-maze test showing total distance traveled (I), time spent in the Novel Arm (J), and number of entries into the Novel Arm (K). All data are presented as mean ± SD (n = 8 mice per group). Statistical comparisons were performed using Kruskal–Wallis test followed by Mann-Whitney U test (C, E) or one-way ANOVA followed by Tukey’s post hoc test (D, F, G, I–K). P values in C indicate comparisons between AD-PBS and AD-W20-LT groups at corresponding time points. Exact P values for key comparisons are displayed in the figure; for complete statistical details, please refer to Table S1 W20-LT Reduces Amyloid Pathology and Modulates Autophagy-Lysosomal Pathways in APP/PS1 Mice Next, we examined the changes of amyloid pathology after W20-LT treatment by assessing the levels of ThS+ plaque deposition in the dorsal hippocampus. Extensive plaque deposition was observed in AD-PBS mice, while W20-LT treatment significantly reduced the fluorescence intensity and density of ThS+ plaques (Fig. 4A, B). To further characterize the changes in distinct Aβ conformations, we performed Dot blot analysis using the oligomer-specific antibody A11 (Kayed et al. 2003) and the fibril-specific antibody OC. W20-LT treatment significantly decreased the levels of A11-positive toxic oligomers, indicating a reduction in the neurotoxic oligomeric pool. Interestingly, a significant reduction was also observed in OC-positive fibrils after W20-LT treatment (Fig. 4C–E). Moreover, ELISA quantification of hippocampal homogenates showed that W20-LT significantly reduced the levels of soluble Aβ40, soluble Aβ42, and insoluble Aβ42 compared to those in the AD-PBS group (Fig. 4F–H). To investigate the impact of W20-LT on the autophagy-lysosome pathway (ALP), we analyzed key protein markers by Western blotting (Fig. 4I). In AD-PBS mice, we observed an accumulation of p62 (SQSTM1), a classical marker commonly associated with impaired autophagy-lysosomal processing (Nixon 2013). W20-LT treatment significantly reduced p62 levels (Fig. 4J) and markedly increased the LC3B-II/I ratio (Fig. 4K), consistent with enhanced autophagy-related activity. Moreover, AD-PBS mice exhibited elevated levels of the lysosomal structural protein LAMP1 and the protease Cathepsin D (CatD), which was consistent with lysosomal accumulation and dysfunction commonly reported in AD pathology (Nixon 2013). Importantly, increased CatD abundance in this context was not interpreted as evidence of more efficient functional degradation, but rather as part of lysosomal stress and compensatory lysosomal accumulation under chronic Aβ burden. W20-LT treatment significantly downregulated these markers (Fig. 4L, M), suggesting reduced lysosomal stress and improved lysosomal homeostasis. Collectively, these results showed that W20-LT treatment reduced amyloid burden and was associated with changes in autophagy-lysosomal pathway-related markers in APP/PS1 mice. Fig. 4W20-LT treatment reduces amyloid pathology and modulates autophagy-lysosomal pathway markers in APP/PS1 mice. After behavioral testing, brain tissues were collected; the left hemisphere was fixed for paraffin sectioning, and the right hippocampus was homogenized for protein analysis. A Representative immunohistochemical images showing Thioflavin S (ThS)-positive senile plaques (green) in the hippocampal area. NeuN staining (red) was used for anatomical localization. Scale bar = 400 μm. B Quantitative analysis of ThS fluorescence intensity from panel A. C Representative dot blot images detecting A11-positive soluble amyloid oligomers and OC-positive amyloid fibrils in brain homogenates from different treatment groups. β-Actin was used as a loading control. D, E Quantification of A11-positive soluble oligomers (D) and OC-positive fibrils (E) from panel C. F–H ELISA quantification of soluble Aβ40 (F), soluble Aβ42 (G), and insoluble Aβ42 (H) levels in brain homogenates. I Representative Western blot images showing the levels of autophagy-related proteins (p62, LC3B) and lysosome-related proteins (LAMP1, CatD) in brain homogenates. β-Actin served as the loading control. J–M Quantitative analysis of p62 (J), LC3B-II/I ratio (K), LAMP1 (L), and CatD (M) protein levels from panel I. All data are presented as mean ± SD (n = 6 mice per group). Statistical comparisons were performed using the Kruskal-Wallis test followed by the Mann–Whitney U test (B, D, E–H, M) or one-way ANOVA followed by Tukey’s post hoc test (J–L). Exact P values for key comparisons are displayed in the figure; for complete statistical details, please refer to Table S1 W20-LT treatment reduces amyloid pathology and modulates autophagy-lysosomal pathway markers in APP/PS1 mice. After behavioral testing, brain tissues were collected; the left hemisphere was fixed for paraffin sectioning, and the right hippocampus was homogenized for protein analysis. A Representative immunohistochemical images showing Thioflavin S (ThS)-positive senile plaques (green) in the hippocampal area. NeuN staining (red) was used for anatomical localization. Scale bar = 400 μm. B Quantitative analysis of ThS fluorescence intensity from panel A. C Representative dot blot images detecting A11-positive soluble amyloid oligomers and OC-positive amyloid fibrils in brain homogenates from different treatment groups. β-Actin was used as a loading control. D, E Quantification of A11-positive soluble oligomers (D) and OC-positive fibrils (E) from panel C. F–H ELISA quantification of soluble Aβ40 (F), soluble Aβ42 (G), and insoluble Aβ42 (H) levels in brain homogenates. I Representative Western blot images showing the levels of autophagy-related proteins (p62, LC3B) and lysosome-related proteins (LAMP1, CatD) in brain homogenates. β-Actin served as the loading control. J–M Quantitative analysis of p62 (J), LC3B-II/I ratio (K), LAMP1 (L), and CatD (M) protein levels from panel I. All data are presented as mean ± SD (n = 6 mice per group). Statistical comparisons were performed using the Kruskal-Wallis test followed by the Mann–Whitney U test (B, D, E–H, M) or one-way ANOVA followed by Tukey’s post hoc test (J–L). Exact P values for key comparisons are displayed in the figure; for complete statistical details, please refer to Table S1 W20-LT Treatment Attenuates Synaptic Loss and Neuronal Damage in the Hippocampus of APP/PS1 Mice The progressive erosion of synaptic integrity and the depletion of neurons are established as the primary structural drivers underlying the clinical manifestation of cognitive decline in AD (Terry et al. 1991). To evaluate the potential neuroprotective effects of W20-LT, we examined synaptic integrity and neuronal viability in the hippocampus. Immunostaining for the postsynaptic marker PSD95 revealed a severe depletion of synaptic puncta in the hippocampus of AD-PBS mice, while W20-LT treatment significantly increased PSD95 fluorescence intensity in APP/PS1 mice (Fig. 5A, D). Consistently, staining for the neuronal marker NeuN demonstrated a marked reduction in neuronal density in AD-PBS mice. W20-LT treatment significantly attenuated this loss, resulting in increased NeuN fluorescence intensity and cell density (Fig. 5B, E, F). Moreover, we also performed Nissl staining to visualize cellular morphology in the CA1 region. The AD-PBS or AD-W20 group exhibited widespread neuronal shrinkage and loss in this region (West et al. 2004). In contrast, W20-LT treatment significantly ameliorated the area of healthy Nissl-positive cells (Fig. 5C, G). These histological observations were corroborated by Western blot analysis of hippocampal homogenates, which confirmed that W20-LT treatment significantly upregulated the protein levels of both PSD95 and NeuN compared to the AD-PBS group (Fig. 5H–J). Collectively, these results suggested that W20-LT treatment mitigated synaptic loss and neuronal damage in the hippocampus of APP/PS1 mice. Fig. 5W20-LT treatment attenuates synaptic loss and neuronal damage in the hippocampus of APP/PS1 mice. A–C Representative images of immunohistochemical staining for PSD95 (green, A; scale bar = 25 μm) and NeuN (red, B; scale bar = 200 μm) in the hippocampus, and Nissl staining in the hippocampal CA1 region (C; scale bar = 100 μm) from different treatment groups. D–G Quantitative analysis performed using ImageJ software for: hippocampal PSD95 fluorescence intensity (D), whole hippocampal NeuN fluorescence intensity (E), NeuN-positive cell density in the CA1 region (F), and the percentage of Nissl-positive cell area in the CA1 region (G). H Representative Western blot images showing the protein levels of PSD95 and NeuN in hippocampal homogenates. β-Actin served as the loading control. I, J Quantitative analysis of PSD95 (I) and NeuN (J) protein levels from panel H using ImageJ. All data are presented as mean ± SD (n = 6 mice per group). Statistical comparisons were performed using one-way ANOVA followed by Tukey’s post hoc test (D, F, G, I) or the Kruskal-Wallis test followed by the Mann–Whitney U test (E, J). Exact P values for key comparisons are displayed in the figure; for complete statistical details, please refer to Table S1 W20-LT treatment attenuates synaptic loss and neuronal damage in the hippocampus of APP/PS1 mice. A–C Representative images of immunohistochemical staining for PSD95 (green, A; scale bar = 25 μm) and NeuN (red, B; scale bar = 200 μm) in the hippocampus, and Nissl staining in the hippocampal CA1 region (C; scale bar = 100 μm) from different treatment groups. D–G Quantitative analysis performed using ImageJ software for: hippocampal PSD95 fluorescence intensity (D), whole hippocampal NeuN fluorescence intensity (E), NeuN-positive cell density in the CA1 region (F), and the percentage of Nissl-positive cell area in the CA1 region (G). H Representative Western blot images showing the protein levels of PSD95 and NeuN in hippocampal homogenates. β-Actin served as the loading control. I, J Quantitative analysis of PSD95 (I) and NeuN (J) protein levels from panel H using ImageJ. All data are presented as mean ± SD (n = 6 mice per group). Statistical comparisons were performed using one-way ANOVA followed by Tukey’s post hoc test (D, F, G, I) or the Kruskal-Wallis test followed by the Mann–Whitney U test (E, J). Exact P values for key comparisons are displayed in the figure; for complete statistical details, please refer to Table S1 W20-LT Treatment Mitigates Astrogliosis and Microglial Activation in the Hippocampus of APP/PS1 Mice Chronic neuroinflammation, driven by sustained glial reactivity and the release of inflammatory signaling molecules, fuels a destructive feed-forward loop that promotes progressive neuronal attrition (Heneka et al. 2015). Furthermore, the accumulation of undegraded aggregates within the lysosomal system is a known trigger for stimulating the inflammatory response in glial cells (Halle et al. 2008; Sheedy et al. 2013). We therefore investigated whether W20-LT could modulate the neuroinflammatory environment. Immunostaining for GFAP (an astrocyte marker) and Iba1 (a microglial marker) revealed extensive astrogliosis and microglial activation in the dorsal hippocampus of AD-PBS mice, characterized by hypertrophic cell bodies and increased coverage area. In contrast, W20-LT treatment significantly reduced the percentage of GFAP-positive and Iba1-positive areas compared to the AD-PBS group, indicating a suppression of aberrant gliosis (Fig. 6A, B, D, E). Moreover, we utilized Imaris software for 3D morphological reconstruction to provide a more detailed characterization of the microglial phenotype (Fig. 6C). While microglia in AD-PBS mice displayed an “activated” amoeboid morphology with retracted processes and enlarged somata, W20-LT treatment promoted a shift toward a more ramified morphology, characterized by a significantly increased number of branch points (Fig. 6F) and reduced cell body volume (Fig. 6G) compared to the AD-PBS group. Western blot analysis confirmed the downregulation of GFAP and Iba1 protein levels in the W20-LT group (Fig. 6H–J). Furthermore, ELISA quantification of hippocampal lysates revealed that W20-LT treatment significantly decreased the concentrations of the pro-inflammatory cytokines TNF-α (Fig. 6K) and IL-1β (Fig. 6L) compared to AD-PBS mice. Collectively, these results suggest that W20-LT dampened the neuroinflammatory response in the hippocampus, possibly in association with reduced Aβ burden and alleviated lysosomal stress. Fig. 6W20-LT treatment mitigates astrocytosis and microglial activation in the hippocampus of APP/PS1 mice. A, B Representative immunohistochemical images of GFAP-positive astrocytes (A) and Iba1-positive microglia (B) in the hippocampus. Scale bar = 200 μm. C Representative 3D reconstruction images of microglial morphology. Scale bar = 50 μm. D, E Quantitative analysis of glial activation. The percentages of GFAP-positive area (D) and Iba1-positive area (E) in the hippocampus was quantified using ImageJ software. F, G Microglial morphology analysis using Imaris 9.9 software to determine the number of branch points (F) and cell body volume (G). H Representative Western blot images showing the protein levels of GFAP and Iba1 in hippocampal homogenates. β-Actin served as the loading control. I, J Quantitative analysis of GFAP (I) and Iba1 (J) protein levels from panel H using ImageJ. K, L Protein levels of TNF-α (K) and IL-1β (L) in hippocampal homogenates were determined by ELISA kit. All data are presented as mean ± SD (n = 6 mice per group). Statistical comparisons were performed using one-way ANOVA followed by Tukey’s post hoc test (D, F, G, I, J, K) or Kruskal-Wallis test followed by Mann–Whitney U test (E, L). Exact P values for key comparisons are displayed in the figure; for complete statistical details, please refer to Table S1 W20-LT treatment mitigates astrocytosis and microglial activation in the hippocampus of APP/PS1 mice. A, B Representative immunohistochemical images of GFAP-positive astrocytes (A) and Iba1-positive microglia (B) in the hippocampus. Scale bar = 200 μm. C Representative 3D reconstruction images of microglial morphology. Scale bar = 50 μm. D, E Quantitative analysis of glial activation. The percentages of GFAP-positive area (D) and Iba1-positive area (E) in the hippocampus was quantified using ImageJ software. F, G Microglial morphology analysis using Imaris 9.9 software to determine the number of branch points (F) and cell body volume (G). H Representative Western blot images showing the protein levels of GFAP and Iba1 in hippocampal homogenates. β-Actin served as the loading control. I, J Quantitative analysis of GFAP (I) and Iba1 (J) protein levels from panel H using ImageJ. K, L Protein levels of TNF-α (K) and IL-1β (L) in hippocampal homogenates were determined by ELISA kit. All data are presented as mean ± SD (n = 6 mice per group). Statistical comparisons were performed using one-way ANOVA followed by Tukey’s post hoc test (D, F, G, I, J, K) or Kruskal-Wallis test followed by Mann–Whitney U test (E, L). Exact P values for key comparisons are displayed in the figure; for complete statistical details, please refer to Table S1 Discussion The efficient removal of Aβ oligomers continues to pose a significant bottleneck in the quest for therapies capable of altering the course of AD. While Aβ accumulation is a central pathological event, current immunotherapies face significant challenges. Traditional monoclonal antibodies (mAbs) rely on Fcγ receptor-mediated phagocytosis to clear aggregates, a process that can trigger pro-inflammatory signaling, microglial exhaustion, and adverse events such as Amyloid-Related Imaging Abnormalities (ARIA) (Salloway et al. 2014) Furthermore, even when internalized, Aβ aggregates often resist degradation within the acidic lysosomal environment, leading to lysosomal storage dysfunction and persistent cellular stress. Accordingly, a pressing demand exists for bifunctional approaches capable of catalyzing internal degradation alongside extracellular neutralization, without compromising the metabolic threshold of the microglial machinery. Herein, we developed and validated W20-LT, a novel tripartite fusion protein designed to bridge the gap between Aβ targeting, microglial uptake, and autophagic degradation. Our results showed that W20-LT was retained within the brain parenchyma following ICV administration, recognized Aβ oligomers, and promoted their intracellular processing in microglia through a p62-LIR-associated autophagy pathway. Consequently, W20-LT treatment in APP/PS1 mice significantly reduced amyloid burden, mitigated neuroinflammation, protected synaptic integrity, and ultimately reversed cognitive deficits. These findings support W20-LT as a promising proof-of-concept “targeted protein degrader” strategy for AD. A critical innovation of our design lies in its distinct mechanism, which synergizes microglial targeting with autophagic degradation to resolve the bottleneck of inefficient clearance. By fusing the p62-LIR motif and the Tuftsin peptide to the scFv, W20-LT functions conceptually similarly to PROTACs or LYTACs, but is designed to harness the autophagy-lysosomal pathway to promote the clearance of extracellularly derived aggregates. While hijacking the autophagic pathway for substrate clearance is a paradigm previously explored in oncology and Huntington’s disease research (Li et al. 2019; Takahashi et al. 2019), the current study extends this protein engineering strategy to an Alzheimer’s disease-related setting. Mechanistically, this “clearance-coupled-degradation” process relies on the coordinated action of the C-terminal Tuftsin and the N-terminal p62-LIR motif. Consistent with previous reports that Tuftsin stimulates phagocytosis via Neuropilin-1 (Nissen et al. 2013; Teesalu et al. 2009), our in vitro data confirmed that W20-LT significantly enhanced the microglial uptake of Aβ oligomers compared to the unmodified scFv. However, enhanced uptake alone is insufficient if the degradation machinery is stalled. Crucially, once internalized, the p62-LIR motif of W20-LT is proposed to facilitate LC3-associated sequestration of Aβ cargo into autophagosome-related compartments, as supported by the elevated LC3B-II/I ratio. This pattern is consistent with enhanced autophagy-associated processing and may help relieve the burden on the lysosomal system, as reflected by reduced LAMP1 and CatD levels. Importantly, increased CatD abundance in the AD-PBS group was interpreted here as being more consistent with lysosomal stress and compensatory lysosomal accumulation under chronic Aβ burden, rather than unequivocal evidence of more efficient functional degradation (Nixon 2013; Whyte et al. 2017). This improvement in intracellular homeostasis, likely related to more effective handling of toxic aggregates, may contribute to the observed phenotypic shift in microglia. In vivo analysis using Imaris revealed that W20-LT treatment reversed the reactive, hypertrophic phenotype seen in AD mice, promoting a return to a more ramified, neuroprotective state. Concomitantly, this morphological shift was accompanied by a marked down-regulation of pro-inflammatory mediators (TNF-α and IL-1β), supporting the view that W20-LT promotes amyloid handling while attenuating inflammatory responses. Regarding comparative efficacy, our study suggests an advantage of the “active degrader” strategy over “passive neutralization” under the low-dose conditions tested. We employed a low-dosage regimen (0.5 µg, every 3 days) to stringently test therapeutic efficiency. Within the current experimental parameters, the native W20 antibody did not yield measurable enhancements in either cognitive performance or disease pathology relative to the AD-PBS cohort. This lack of efficacy contrasts with previous reports where high doses of W20 were effective (Zhao et al. 2014). This discrepancy suggests that, as a passive blocker, W20 may be constrained by stoichiometric requirements and may require higher concentrations to saturate toxic oligomers. In contrast, W20-LT, by promoting intracellular clearance rather than merely neutralizing its targets, achieved clear therapeutic effects at the same low dosage. Moreover, no measurable behavioral impairment was observed in healthy mice under the conditions tested, as pairwise comparisons indicated no significant disparities between WT-PBS and WT-W20-LT cohorts. Across the entire battery of behavioral tests, the performance of W20-LT-treated wild-type mice remained indistinguishable from that of the vehicle control group. Collectively, these results suggest that W20-LT combines the Fc-sparing features of scFvs with an efficient degradation-oriented design. Despite the potential of W20-LT demonstrated here, several limitations of this study should be acknowledged. First, efficacy was evaluated only in male APP/PS1 mice, which limits generalizability across sex, disease stage, and model context. Given the higher prevalence of AD in females and known sex-dependent differences in microglial biology (Mielke et al. 2014), future studies should include female cohorts. Second, the current study relied on a single amyloid-model background, and validation in additional AD models will be necessary to determine the broader applicability of this strategy. Third, although the in vitro degradation assays, the W20-nsT control, and the changes in LC3B, p62, LAMP1, and CatD support involvement of the autophagy-lysosomal pathway, these data do not by themselves definitively establish restored autophagic flux or direct intracellular cargo routing in vivo. Fourth, the treatment window was relatively short and relied on indwelling ICV cannulation, leaving the durability, long-term tolerability, and translational feasibility of W20-LT to be established. To improve translational potential, future studies could explore less invasive delivery strategies, such as coupling W20-LT to blood-brain barrier-penetrating shuttles, including transferrin receptor binders (Burgio et al. 2023), or using vector-based expression approaches. Finally, safety evaluation in this study remained limited, and broader assessment of immunogenicity, off-target effects, and long-term toxicity will be important in future development. In summary, we engineered W20-LT, a multifunctional fusion protein that converts a passive scFv into an active clearance-oriented construct for amyloid handling. By simultaneously enhancing microglial uptake via Tuftsin and promoting autophagy-associated intracellular clearance via p62-LIR, W20-LT may help interrupt the cycle of accumulation and inflammation. Our findings provide preclinical support for the concept that redirecting pathogenic cargo toward an autophagy-associated pathway may help alleviate lysosomal burden and improve microglial functional stability, thereby supporting further investigation of this strategy in AD. Supplementary Information Below is the link to the electronic supplementary material.Supplementary file1 (PDF 319 KB) Supplementary file1 (PDF 319 KB)