Targeting phagocytosis for amyloid-β clearance: implications of morphology remodeling and microglia activation probed by bifunctional chimaeras

Youqiao Wang; Zeyi Wang; Ziyi Liu; Jinyan Li; Shuo Yang; Yuxuan Zhao; Yangmei Huang; Chenyang Liao; Yiqiu Zhang; Jiaojiao Zhao; Weilin Zhou; Binhua Zhou; Xin Yue; Qiang Zhou; Xianzhang Bu

doi:10.1038/s41467-025-63458-3

. 2025 Aug 30;16:8128. doi: 10.1038/s41467-025-63458-3

Targeting phagocytosis for amyloid-β clearance: implications of morphology remodeling and microglia activation probed by bifunctional chimaeras

Youqiao Wang ^1,^#, Zeyi Wang ^2,^#, Ziyi Liu ^1,^#, Jinyan Li ², Shuo Yang ², Yuxuan Zhao ¹, Yangmei Huang ², Chenyang Liao ¹, Yiqiu Zhang ¹, Jiaojiao Zhao ¹, Weilin Zhou ¹, Binhua Zhou ³, Xin Yue ^4,^✉, Qiang Zhou ^2,^✉, Xianzhang Bu ^1,^✉

PMCID: PMC12398596 PMID: 40885741

Abstract

Amyloid-β (Aβ), a key driver of Alzheimer’s disease (AD) pathogenesis, possesses diverse harmful and clearance-resistant structures that present substantial challenges to therapeutic development. Here, we demonstrate that modulating Aβ morphology, rather than Toll-like receptor 2 (TLR2)-dependent microglia activation, is essential for effective phagocytosis of Aβ species by microglia. By developing a bifunctional mechanistic probe (P2CSKn) designed to remodel Aβ and activate TLR2, we show it restructures soluble Aβ (sAβ) and fibrillar Aβ (fAβ) into less toxic hybrid aggregates (hPAβ). Critically, this structural remodeling protects microglia from Aβ toxicity while enabling robust phagocytosis. Moreover, although TLR2 activation mildly enhances Aβ uptake, it concurrently triggers detrimental inflammation that negates its benefits. Our findings establish morphological remodeling as the critical determinant of effective Aβ clearance and suggest a morphology-focused strategy for developing safe therapeutics for Aβ-related diseases.

Subject terms: Alzheimer's disease, Protein folding, Medicinal chemistry, Chemical modification

Researchers discover remodeling harmful Alzheimer disease-associated Amyloid-β species, rather than activating immune cells, enables their effective and safer removal in models, suggesting an alternative therapy design approach.

Introduction

Amyloidosis is a complex and intricate process consisting of a series of steps encompassing primary and secondary nucleation, elongation, and amyloid deposition, and it gives rise to diverse forms of amyloid species, including oligomers, protofibrils, mature fibrils, and amyloid plaques^1,2. These amyloid proteins share common characteristics of cross-β-structures that contribute to amyloidosis¹, and lead to the occurrence and development of amyloid-related diseases³, such as Alzheimer’s disease (AD)⁴, Parkinson’s Disease^5,6 and tumorigenesis^7,8. The β-amyloid (Aβ) proteins, generated through proteolytic cleavage of the amyloid precursor protein (APP) by β- and γ-secretase enzymes, have been regarded as a causal event triggering the onset and progression of Alzheimer’s disease (AD)^3,4. Moreover, recent research revealed that melanoma can produce and utilize Aβ proteins to facilitate its metastasis to the brain⁷. Drug development targeting Aβ protein has thus garnered sustained and substantial attention.

The considerable morphological diversity renders complex pathology of Aβ species⁹. Different Aβ variants generated by amyloidosis contribute to the neurotoxicity through interactions with distinct molecular targets^10,11, thus making it unlikely to achieve their modulation via a single molecule as for classical pathological targets. Currently, the predominant approaches targeting Aβ species evolve around targeting the amyloidosis process via small molecules known as “β-structure breaker”¹². However, the efficacy of these strategies is impeded, at least partially, by the morphological diversity, and by their dynamic equilibrium within the in vivo milieu¹⁰. On the other hand, the impaired ability of microglia to engulf Aβ is a pivotal mechanism driving the progression of AD^13,14. In the past decades, accumulated evidences indicated that anti-Aβ antibodies can bind to microglial FcR and/or Toll-like receptors¹⁵ to activate microglia and increase their phagocytic ability^16,17 to clear Aβ plaques in the AD brain. By targeting Aβ aggregates, protofibrils and plaques respectively, three anti-Aβ antibodies, Aducanumab¹⁸, Lecanemab¹⁹ and Donanemab²⁰ have demonstrated their ability to reduce Aβ plaque levels in early stage AD patients, and have received FDA approval for AD therapy²¹. Nonetheless, the development of anti-Aβ antibody drugs is a formidable process sparking ongoing discussions^21–25, and many other anti-Aβ antibodies have failed for various and complex reasons^26–28. Specifically, it has been found that antibodies targeting Aβ clearance may be associated with a phenomenon known as the “dust-raising” effect, whereby deposited Aβ is converted into more toxic soluble oligomers, and ultimately leads to accelerated neuronal degeneration^29–32. Moreover, treatment with anti-Aβ antibodies may induce a shift in microglia activation from M2 to M1 phenotype, and contribute to accelerated Aβ deposition and cognitive impairment^33–37. These findings collectively highlight the potential risks associated with immunotherapies targeting Aβ clearance. While restoring the phagocytosis activity of microglia by anti-Aβ antibodies has emerged as a major approach to engulf Aβ species³⁸, a delicate balance between augmenting Aβ phagocytosis and mitigating cytotoxicity and inflammatory responses needs to be achieved.

We have focused on two fundamental questions that hinder the rational drug design: (1) Can inducing Aβ species into uniform morphotypes overcome the cytotoxicity of Aβ species and avoid the “de-toxification/re-toxification” dilemma during plaque removal, and enhance their phagocytic clearance by microglia? (2) Does activation of microglia have synergy with structural remodeling of Aβ species to benefit their phagocytic clearance? Previously, we discovered that well-designed cyclopeptides can effectively reduce the morphological heterogeneity and impede the formation of toxic oligomers by stabilizing the monomer form of Aβ species^39–41.

In this work, we aim to dissect the contributions of Aβ morphology modulation and microglia activation to Aβ clearance by developing a chimeric probe (P2CSKn) that integrates an Aβ-remodeling module and a microglia-activating module. We demonstrate that P2CSKn, either through co-assembly with soluble Aβ species (sAβs) or reassembly of pre-formed Aβ fibrils (fAβs) into structurally distinct P2CSKn/Aβ hybrid aggregates (hPAβs), effectively reduces the morphological heterogeneity and mitigates the cytotoxicity of Aβ species towards microglia, and achieves augmented phagocytosis. Critically, we reveal that structural remodeling of Aβ ensures the high efficacy of phagocytosis of Aβ₄₂ species; moreover, while TLR2 activation achieves low to moderate enhancement for phagocytosis of Aβ₄₂ species, this benefit is counterbalanced by a critical inflammasome activation in vitro and in vivo. These mechanistic discoveries establish an alternative strategy for developing Aβ-targeting therapy in the future, by making Aβ conformation modulation a priority, with optional TLR2 co-activation restricted to localized delivery.

Results

Conctruction of structure-remodeling and microglia-activating chimaeras

To achieve efficient biochemical probes capable of modulating the morphology and phagocytosis of Aβ species, we proposed that chimaeras integrating a microglia-activating and an Aβ-recognizing module will be effective. Structurally, Aβ protein is an amphipathic molecule containing an acidic hydrophilic N-terminal, and two hydrophobic fragments in the middle and C-terminus that are ready to form stacking β-structure and contribute to amyloidosis^42,43. Cryogenic electron microscopy (cryo-EM) study has elucidated the structure of Aβ₄₂ fibrils in AD brains, where several salt bridges between D1 and K28’, D7 and R5, and E11 with H6 and H13 and intermolecular hydrophobic interactions were observed (Fig. 1a), stabilizing a kink in the N-terminal part of the β-sheet around Y10 and enabling the Aβ₄₂ stacking⁴⁴. Based on these observations, we proposed a chimaeras consists of Aβ species-recognizing motif, a salt bridge-breaking motif, and a long hydrophobic chain to break the stacks within fibrils and build hybrids hydrophobic interactions (Fig. 1a). Among the hydrophobic domain, residues 16–20 (KLVFF) fragment is critical for the β-sheet formation and self-assembly of Aβ species⁴⁵ and thus is used as a Aβ species-recognizing motif, with an L-1-naphthylalanine (nal) appended in the C-terminal to enhance the β-structure recognition (termed as KLVFFn)⁴⁶, to achieve a recognition and basic interaction towards Aβ species.

Fig. 1 — a The cryo-EM structures of fAβ₄₂ (pdb code: 5OQV) and proposed strategy for chimaera design. The image was generated by MOE 2015.10, and key interactions involved residues are shown. CT C-terminal, NT N-terminal. b The effects of fAβ₄₂ (0.5 μM) on the expression of mRNA of *Marco*, *Msr1* and *Mrc1* in BV2 cells after incubation with cells for different amount of time, as determined by rt-qPCR. c Effects of fAβ₄₂ treatment on the phagocytosis of fluorescent microsphere in BV2 cells, as measured by flow cytometry. d Effects of various TLR agonists on the expression of mRNA of *Marco*, *Msr1* and *Mrc1* in BV2 cells, as determined by rt-qPCR. LPS (1.0 μg/mL), Pam3CSK4 (1.0 μM), Pam2CSK4 (1.0 μM), Zymosan A (50 μg/mL), Flagellin (200 nM) were used. e Structures of designed chimaeras and reference compounds. RP, Random Peptide. Data are representative of three independent experiments in (b–d). Data are presented as mean ± SD, n = 3 independent samples in (b–d) using one-way ANOVA with Dunnett’s post hoc test. ns, not significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Exact p values were given in the Source Data file. Source data are provided as a Source Data file.

To integrate the micrglia-activating module, we first employed RNA-seq and found that series of phagosome-related genes were significantly donwregulated in BV2 cells treated with fAβ₄₂ (Supplementary Figs. 1a, b, 2a), especially for the scavenger receptor A1 (Msr1), scavenger receptor A2 (Marco), and macrophage mannose receptor 1 (Mrc1) with –log₂(FC) > 1, coincided with the attenuation of phagosome-associated pathway (Supplementary Fig. 1c). Rt-qPCR analysis further validated that their expression were significantly suppressed by 0.5 μM fAβ₄₂ in BV2 cells after 30 (Mrc1, Msr1) or 60 (Marco) min, respectively (Fig. 1b). Moreover, we observed that treatment of fAβ₄₂ (0.5 μM) time-dependently decreased the uptake of fluorescent microsphere in BV2 cells (Fig. 1c), implying that the Aβ₄₂ species impairs the phagocytic activity of BV2 cells.

Toll-like receptors (TLRs) play crucial roles in the innate immune response and are responsible for the phagocytosis of marcophages either by TLR2- and TLR4-dependent MyD88 signaling⁴⁷, or TLR5-dependent PI3K/Akt/mTORC1 pathway⁴⁸. We found that Tlr2, Tlr4 or Tlr5 was not significantly affected by fAβ₄₂ treatment (Supplementary Fig. 2b). In addition, different TLR agonists show diverse effects on the mRNA level of Mcr1, Marco and Msr1 respectively, where TLR4 agonist Lipopolysaccharides (LPS, 1.0 μg/mL), TLR2 agonists Pam3CSK4 (1.0 μM), Pam2CSK4 (1.0 μM), and zymosan A (50 μg/mL) significantly attenuated the expression of Mrc1 in BV2 cells. Conversely, Pam2CSK4 and zymosan A substantially upregulated Msr1 expression in BV2 cells, LPS, Pam2CSK4 upregulated Marco respectively, while TLR5 agonist flagellin (200 nM) failed to upregulate either Msr1 or Marco (Fig. 1d). Importantly, we noticed that TLR2 agonist Pam2CSK4 and Pam3CSK4 bear a KKKK C-terminal, and two and three hydrophobic palmitate chains respectively, that meet with the desires for the salt bridge-breaking and hydrophobic interrupting in chimaeras design. We then expected that introducing Pam2CSK4 or Pam3CSK4 into the chimaeras would allow us to integrate the Aβ-remodeling and the microglial-activating module into the chimaeras, and to probe their implications on the phagocytosis of Aβ species in both coupled and decoupled modes through spatial control strategies.

Several chimaeras (Fig. 1e) were then designed and synthesized using solid-phase synthesis as described in the Supplementary Method. Pam2CSK4 and its derivative Pam3CSK4 were used to couple with KLVFFn via a 6-aminohexanoic acid linker and generate P2CSKn (Pam2CSK4-ahx-KLVFF-nal) and P3CSKn (Pam3CSK4-ahx-KLVFF-nal), respectively. RhoB (Rhodamine B) labeled P2CSKn (RhoB-P2CSKn), P2KTKn (Pam2KTK4-ahx-KLVFF-nal) were also designed and synthesized by blocking the amino group of cys residue with RhoB, or replacing the cys-ser fragment with lys-thr and altering the palmitic acid chain. RhoB-P2CSKn is expected to serve as an optical probe of P2CSKn for morphology-related assays. In addition, a reference peptide with one palmitate chain and random peptide sequence (Pam-KCKSKVFLKFK-ahx-nal, RP), was also prepared as described in Supplementary Methods.

P2CSKn efficiently remodels different Aβ₄₂ species to monodisperse hybrid co-aggregates

P2CSKn fast co-assembles with sAβ₄₂ to form conformation-modified hybrid aggregates

We first evaluated the ability of designed chimaeras in promoting aggregation of Aβ species by observing the Tyndall effect of sAβ₄₂ before and immediately after the addition of designed chimaeras or reference compounds. As shown in Fig. 2a and Supplementary Fig. 3, definitive enhanced Tyndall effects were observed in the sAβ₄₂ with the addition of various concentrations of P2CSKn or RhoB-P2CSKn, indicating that these chimaeras rapidly promote sAβ₄₂ into large-size particles. P3CSKn and P2KTKn themselves show significant Tyndall effect but didn’t facilitate that of sAβ₄₂, implying a tendency of self-aggregate for P3CSKn and P2KTKn but incapable of recognizing the Aβ₄₂ species. Moreover, Pam2CSK4 showed much less activity than P2CSKn, indicating that the Pam2CSK4 motif partially facilitates the ability of P2CSKn to prompt Aβ₄₂ aggregation. Dynamic light scattering (DLS) analysis revealed that the two peaks around 3 and 12 nm, respectively, in the hydrodynamic diameter of sAβ₄₂ particle. With increasing concentration of P2CSKn, this distribution shifted to a single range around 100 nm, which did not occur in the presence of Pam2CSK4 or KLVFFn (Fig. 2b). Moreover, P2CSKn rapidly and concentration-dependently prevented the formation of Aβ₄₂ oligomers (Fig. 2c) as determined by dot-blot assay using an oligomer-specific antibody A11⁴⁹. By using RhoB-P2CSKn and soluble FITC-labeled Aβ₄₂ (sFITC-Aβ₄₂), we observed that both RhoB-P2CSKn and FITC-Aβ₄₂ co-located inside large particles generated after treating sFITC-Aβ₄₂ with RhoB-P2CSKn, and the Interactive Curve and 3D Surface Plot indicated that co-assembly occurred during the aggregation progress (Fig. 2d). As β-sheet rich conformations of Aβ species, which can be recognized by fluorescent dye Thioflavin T (ThT), are accumulated during their aggregations, the fluorescence intensity of ThT has been widely used as an indicator for aggregated grade of Aβ species⁵⁰. Interestingly, we observed that the addition of 5.0 eq. P2CSKn almost completely suppressed the increase of ThT fluorescence during the aging process of sAβ₄₂ (10 μM, Fig. 2e), implying that the native fibrillization was prevented and a conformation alteration occurred in the co-assembly process. Consistently, concentration-dependent decrease of β-sheet signal of sAβ₄₂ species was validated using Far-UV Circular Dichroism (CD) spectroscopy¹⁴ (Fig. 2f). Finally, we calculated the K_d (0.041 μM) of P2CSKn binding to sFITC-Aβ₄₂ (0.5 μM) by microscale thermophoresis (MST) assay (Fig. 2g). Together, P2CSKn can recognize and fast co-assemble with sAβ₄₂ to form conformation-modified P2CSKn/Aβ₄₂ hybrid aggregates (hPAβ₄₂).

Fig. 2 — a Tyndall Effects in sAβ₄₂ solution with or without gradient concentration of P2CSKn, Pam2CSK4 and KLVFFn, respectively. b Effects of different concentrations of P2CSKn, Pam2CSK4 and KLVFFn on the particle size of sAβ₄₂, assessed using DLS. c The effect P2CSKn and reference compound on the formation of Aβ₄₂ oligomers, time- (top) and concentration-(bottom) course effects were evaluated using Dot Blot with oligomer-specific antibody A11. d Representative morphology of RhoB-P2CSKn treated sFITC-Aβ₄₂ (12 h) captured using LSCM. The colocalization of RhoB (Rhodamine B) and FITC is analyzed using Image J (v 1.54 g). e ThT fluorescence intensity kinetics of sAβ₄₂ (10 μM) with or without P2CSKn (50 μM). 10 μM ThT was used, and the fluorescence was monitored before and after the addition of P2CSKn. The fluorescence intensity at the endpoint was measured for statistical analysis. f Far-UV CD spectra of sAβ₄₂ in the absence or presence of different concentrations of P2CSKn in 50.0 mM phosphate buffer (pH 7.4) without Cl^-. Curves were calibrated by subtracting the basic signal from phosphate buffer. g MST spectra of different concentrations of P2CSKn on sFITC-Aβ₄₂ (0.5 μM), data were processed by MO. Affinity Analysis v2.3, and the curve was fitted using the Hill model to afford K_d. Data are representative of three independent experiments with similar results in (a, c, d). Data are presented as mean ± SD, n = 4 independent samples in (e) using two-tailed t tests. ****p < 0.0001. Exact p values were given in the Source Data file. Source data are provided as a Source Data file.

P2CSKn re-assembles fAβ₄₂ to conformation and morphology-modified, and monodisperse hybrid co-aggregates

We next examined whether P2CSKn can influence the pre-formed fAβ₄₂. As shown in Fig. 3a, the fluorescence intensity of ThT in sAβ₄₂ (10 μM) species increased steady (lag time = 6.5 h) during the aging process, indicating the continuous formation of β-sheet-rich fAβ₄₂. Markedly, after the fluorescence intensity reached a plateau after 32 h, injection of 5.0 eq. P2CSKn to the solution sharply decreased the fluorescence (half-life = 2.0 hr), indicating that the β-sheet structure in fAβ₄₂ was disrupted by P2CSKn. This observation was further validated using Far-UV CD spectroscopy (Fig. 3b).

Next, we evaluated the alterations of different aggregative forms within the fAβ₄₂ upon P2CSKn treatment, using Photo-induced cross-linking of unmodified protein (PICUP)-Western Blot assay⁵¹ with antibody 4G8 recognizing residues 17–24 of Aβ species⁵². As depicted in Fig. 3c, oligomers within the fAβ₄₂ (20 μM) predominantly exhibited a molecular weight distribution ranging from around 40 to 170 kDa and remained relatively stable over time. Upon addition of 0.5 eq. of P2CSKn (10 μM), however, a reduction of high molecular weight oligomers (40–170 kDa) was observed, concurrently with an increase in 15–25 kDa oligomers in a time-dependent manner, suggesting a decomposition of these oligomers occurred by P2CSKn treatment. Interestingly, these newly appeared 15–25 kDa oligomers showed reduction with increasing concentrations of P2CSKn to 20 (1.0 eq.) and 40 μM (2.0 eq.), respectively. Conversely, new oligomers 40–130 kDa appeared and subsequently disappeared. These results indicated that P2CSKn modified these oligomers within fAβ₄₂, and promoted the formation of high molecular weight aggregates that were unable to migrate into the stacking gel during electrophoresis. Notably, such alterations were not observed in the Pam2CSK4- or KLVFFn-treated fAβ₄₂ (Supplementary Fig. 4). These data indicate that P2CSKn systematically perturbed the heterogeneity of aggregates within the Aβ assemblies.

In transmission electron microscopy (TEM) imaging (Fig. 3d), it is found that the 2.0 eq. P2CSKn-treated fAβ₄₂ (20 μM, 3 h) were present as stout and drop-like morphology, and become branched and stout sticks at 5 and 8 h, which were completely different from the untreated slender fAβ₄₂. For fFITC-Aβ₄₂, similar branched and stout sticks are observed upon 18-h-treatment with 2.0 eq. P2CSKn (Supplementary Fig. 5). These reshaped sticks of fAβ₄₂ eventually come into contact with each other, leading to the formation of plaques finally (18, 28 h) (Fig. 3d). Noticeably, there is a significant increase in the diameter of P2CSKn-treated fAβ₄₂ fibers after 3–5 h incubation, followed by a subsequent decrease over time (8, 18 and 28 h) (Fig. 3d). These observations suggest a dynamic process where P2CSKn initially facilitates the formation of relatively loose sticks during the re-assembling of fAβ₄₂, which then undergo further compaction with prolonged incubation. Moreover, upon injection of RhoB-P2CSKn (25 μM) into the pre-formed FITC-Aβ₄₂ aggregates (fFITC-Aβ₄₂, 5.0 μM), it is observed that RhoB-P2CSKn rapidly co-located with the FITC-Aβ₄₂ plaque within the individual plaques (P1-P6) or entire field of view, as revealed using continuous confocal imaging (Fig. 3e, f). The Interactive Curve and 3D Surface Plot also indicated that the original plaque was reshaped in a time-dependent manner (Fig. 3f and Supplementary Video 1), which did not occur with Pam2CSK4 and KLVFFn (Supplementary Fig. 6).

To further understand the nature of structural remodeling of fAβ₄₂ by P2CSKn, time-dependent hydrodynamic diameter distributions were analyzed by DLS in both whole mixtures and supernatants post-centrifugation (Fig. 3g). In the whole sample, three distinct populations (about 40, 700, and 6000 nm) were observed in fAβ₄₂-only control, indicative of a dynamic equilibrium among oligomers, protofibrils, and mature fibrils in fAβ₄₂ solution. Over 18.5 h, these populations dynamically equilibrated into species around 16, 136, and a dominant 6000 nm, suggesting spontaneous fibril fragmentation-reassembly cycles. In contrast, P2CSKn alone maintained monodisperse nanoparticles (around 66 nm) throughout incubation. Remarkably, co-incubation of equimolar P2CSKn and fAβ₄₂ induced immediate structural reorganization: Initial tri-modal distribution (around 40, 700 nm, and 9000 nm) shifted within 0.5 h to bimodal populations (around 40–900 nm range) with complete disappearance of >5000 nm aggregates. By 18.5 h, the system converged to a monodisperse state (around 218 nm), demonstrating P2CSKn-driven conversion of polymorphic fibrils into uniform co-assemblies. Centrifugation-resolved analysis revealed distinct supernatant dynamics. In fAβ₄₂ controls, the supernatant initially showed monomodal distributions (around 40–85 nm range) until 4.5 h, followed by emerging larger species (around 107 and 717 nm at 18.5 h), reflecting continuous fibril shedding. P2CSKn supernatants persistently exhibited unaltered about 66 nm particles. Strikingly, P2CSKn/fAβ₄₂ supernatants displayed time-dependent monomodal peaks (around 170–280 nm range), confirming effective sequestration of fAβ₄₂ into hybrid P2CSKn/fAβ₄₂ complexes. These data collectively demonstrate that P2CSKn rapidly remodels aggregated fAβ₄₂, ultimately generating monodisperse co-assemblies. This observation aligns with PICUP-Western Blot analyses showing the dynamic assembly process, TEM visualization of hybrid fibril nanostructures, and colocalization analysis fFITC-Aβ₄₂ with RhoB-P2CSKn.

Together, P2CSKn can efficiently reassemble fAβ₄₂ or fFITC-Aβ₄₂ to form conformation and morphology-modified, monodisperse hPAβ₄₂ or hPFITC-Aβ₄₂, respectively.

P2CSKn efficiently promotes a durable ingestion of Aβ₄₂ species in different cell model

We next investigated the efficiency of P2CSKn on the uptake of Aβ₄₂ species under various conditions. Briefly, sFITC-Aβ₄₂ and fFITC-Aβ₄₂ were pretreated with different equivalents of P2CSKn (30min), respectively, to obtain the P2CSKn-modified forms of FITC-Aβ₄₂ (hPFAβ₄₂). It is noted that in time-lapse analysis, 2.0 eq. P2CSKn was used for sFITC-Aβ₄₂, and a higher ratio of 8:1 (P2CSKn: fFITC-Aβ₄₂) was employed in the case of fFITC-Aβ₄₂ to ensure the formation of hPFAβ₄₂. Because we thought that using 2.0 eq. P2CSKn might be insufficient to reconstitute the fFITC-Aβ₄₂ under the same pre-incubation time (30 min) as that of sFITC-Aβ₄₂, as implicated in Fig. 3a.

In the mouse microglia cell line BV2 model, we first observed the accumulation of cellular FITC-Aβ₄₂ using a fluorescence microplate reader⁵³ upon P2CSKn treatment_. As shown in Fig. 4a, the intracellular uptake of sFITC-Aβ₄₂ (0.5 μM) reached saturation in about 2 hr which was followed by a gradual diminution. In contrast, the accumulation of 2.0 eq. P2CSKn-treated sFITC-Aβ₄₂ demonstrated a continuous augmentation throughout the duration of the experiments in general, surpassing the accumulation levels of untreated sFITC-Aβ₄₂ over time. We next observed that the extracellular Aβ₄₂ species were almost completely eliminated upon P2CSKn treatment, while untreated Aβ₄₂ species remained abundant (Supplementary Fig. 7). These results suggest that Aβ₄₂ species hinder the phagocytosis of BV2 cells, while P2CSKn treatment can restore the sustained phagocytosis and promote the clearance of Aβ₄₂ species.

The phagocytosis of sFITC-Aβ₄₂ (0.5 μM) in BV2 cells was therefore investigated using a flow cytometry assay. As shown in Fig. 4b, it is found that both the number of phagocytosis-activated BV2 cells (identified by count of cells internalized FITC-Aβ₄₂, Count), and the intracellular accumulation of FITC-Aβ₄₂ (identified by mean fluorescence intensity, MFI) exhibited an initial increase within ~ 360 min, followed by a decline over time. In contrast, pretreating sFITC-Aβ₄₂ with 2.0 eq. P2CSKn consistently and persistently augmented the amount of both BV2 cells internalizing FITC-Aβ₄₂ and the cellular accumulation of FITC-Aβ₄₂. In addition, the ability of P2CSKn to promote phagocytosis of fFITC-Aβ₄₂ was also investigated. Similar to the case of sFITC-Aβ₄₂, it is found that the number of phagocytosis-activated BV2 cells and the intracellular accumulation of fFITC-Aβ₄₂ exhibited an initial increase and followed by a decline (Fig. 4c), while pretreating fFITC-Aβ₄₂ with P2CSKn (8.0 eq.) efficiently promoted phagocytosis of fFITC-Aβ₄₂, emphasizing a robust activity of P2CSKn in promoting phagocytosis of FITC-Aβ₄₂. It is worth pointing out that, although TLR2 agonist Pam2CSK4 consistently and persistently increased the percentage of BV2 cells internalizing sFITC-Aβ₄₂ (Fig. 4b) or fFITC-Aβ₄₂ (Fig. 4c), the intracellular accumulation of FITC-Aβ₄₂ or fFITC-Aβ₄₂ was notably lower than that induced by P2CSKn treatment. For example, in sFITC-Aβ₄₂ group at 180 min, the count₁₈₀ of P2CSKn and Pam2CSK4 were similar (89.5 ± 1.2% and 63.6 ± 2.2% respectively); but the MFI₁₈₀ are significantly different (80549 ± 3085 and 18911 ± 310 respectively); in fFITC-Aβ₄₂ group at 180 min the count₁₈₀ of Pam2CSK4 (89.2 ± 0.3%) is even higher than P2CSKn (74.1 ± 2.7%), but its MFI₁₈₀ (24286 ± 520) are significantly lower than that of P2CSKn (37284 ± 2359). These results implied that while TLR2 activation by Pam2CSK4 or P2CSKn may activate the phagocytic activity of BV2 cells, and the formation of P2CSKn/Aβ₄₂ co-aggregates ensures an efficient uptake.

Next, we proceeded to utilize primary microglia to examine their persistence of phagocytic capacity towards sFITC-Aβ₄₂ (0.5 μM). Primary microglia from 10-month-old mice were extracted, purified and examined (Supplementary Fig. 8) according to the literatures⁵⁴. As illustrated in Fig. 4d, the cell population with engulfed sFITC-Aβ₄₂ initially increased to a moderate level (~40%) within about 360 min, and subsequently exhibited a significant decline; meanwhile, the intracellular accumulation of FITC-Aβ₄₂ was extremely low. Conversely, in P2CSKn (1.0 μM) pretreated sFITC-Aβ₄₂ group, the cell population with engulfed FITC-Aβ₄₂ rapidly increased and reached a saturation level within 1 h (approaching 100%), with a sustained increase in the intracellular accumulation of FITC-Aβ₄₂.

The above results confirmed that Aβ₄₂ species hinder the sustained phagocytosis in both BV2 and mouse primary microglia, which can be rescued by P2CSKn. We next calculated the EC₅₀ of P2CSKn using Count in promoting the phagocytosis of 0.5 μM sFITC-Aβ₄₂ and fFITC-Aβ₄₂ (183.6 and 1219 nM, respectively, Fig. 4e). To examine the selectivity of P2CSKn on Aβ phagocytosis, we utilized 1,1’-dioctadecyl-3,3,3’,3’-tetramethyl-indocarbocyanine perchlorate-labeled low-density lipoprotein (Dil-LDL, 1.0 μM) as a control. The results indicated that 0.5 μM P2CSKn didn’t enhance the phagocytosis of Dil-LDL, but promoted the engulfment of 0.5 μM sFITC-Aβ₄₂ (Supplementary Fig. 9), indicating a specificity of P2CSKn in Aβ phagocytosis.

To obtain details on the phagocytosis of P2CSKn-modified Aβ species, we used time-lapse confocal imaging to monitor the dynamic process. As shown in Fig. 4f and Supplementary Video 2–3, hybrid plaques of 1.0 eq. P2CSKn or RhoB-P2CSKn modified sFITC-Aβ₄₂ can be observed, and BV2 cells extend pseudopodia to capture these plaques, and subsequent shrinkage of the pseudopodia to finish the uptake of FITC-Aβ₄₂; in untreated fFITC-Aβ₄₂ group, this phenomenon was rarely observed, and more times is needed for visualization of cellular sFITC-Aβ₄₂ puncta (Supplementary Fig. 10), indicating that these hybrid plaques can be more readily recognized and up-taken by BV2 cells than the fFITC-Aβ₄₂. Moreover, the cellular localization of 0.5 μM P2CSKn-treated sFITC-Aβ₄₂ (0.5 μM) was observed, and it was found that intracellular fluorescence of FITC was exclusively co-localized with the Lysotracker Red (Fig. 4g), indicating that the hybrid aggregates were inside the lysosome after phagocytosis. Notably, when the BV2 cells were pretreated with lysosome inhibitor chloroquine (10 μM, 1 h), both lysosome amount and the level of phagocytosis of hybrid aggregates were significantly reduced (Fig. 4g), implying that the lysosome is involved in Aβ₄₂ phagocytosis. To evaluate the degradation behavior of Aβ₄₂ species after phagocytosis, sFITC-Aβ₄₂ (0.5 μM), fFITC-Aβ₄₂ (0.5 μM) and preformed hPFAβ₄₂ (0.5 μM P2CSKn + 0.5 μM sFITC-Aβ₄₂) were first incubated with BV2 cells for 180 min, respectively, then the cells were washed and further cultured for different periods of time, and cellular fluorescence was recorded. As shown in Fig. 4h, intracellular FITC-Aβ₄₂ in all hPFAβ₄₂, sFITC-Aβ₄₂ and fFITC-Aβ₄₂ groups was efficiently degraded, and this effect was almost completely abolished by pre-treatment of BV2 cells with chloroquine. These results indicated that different forms of Aβ₄₂ are degraded in lysosomes after phagocytosis.

Roles of P2CSKn in the phagocytosis of Aβ₄₂ species by BV2 cells

In the aforementioned study, we have elucidated the robust efficacy of P2CSKn in facilitating sustained phagocytosis of Aβ₄₂ species. Given that P2CSKn was designed as a multifunctional molecule, its effectiveness can be attributed to the synergistic interplay between morphological modulation and microglia activation. A series of experiments were then conducted to dissociate the impacts of each function of P2CSKn in the phagocytosis of Aβ₄₂ species.

BV2 cells can uptake different Aβ₄₂ species in TLR2-dependent manner

As P2CSKn contains a TLR2 agonist motif Pam2CSK4, we wondered whether TLR2 expression is indeed involved in the uptake of Aβ₄₂ species. TLR2-deficient BV2 cells were then generated using TLR2-siRNA, as validated using Western blot (Supplementary Fig. 11), the phagocytosis of different FITC-Aβ₄₂ forms by BV2-NC (cells transfected with null-siRNA control) and TLR2-knockdown BV2 cells was examined. In this assay, we used a short incubation time (5 and 30 min) to avoid potential damage to cells by Aβ₄₂ species. As shown in Fig. 5a, while the preformed hPFAβ₄₂ from 1.0 μM P2CSKn with 0.5 μM sFITC-Aβ₄₂ was much more rapidly taken up by the BV2-NC cells than that of sFITC-Aβ₄₂ and fFITC-Aβ₄₂, knocking down TLR2 dramatically reduced uptake of hPFAβ₄₂ (from ~65% to 8.64% after 30 min) as well as the uptake of sFITC-Aβ₄₂ (from ~27.3% to 12.6% after 30 min) and fFITC-Aβ₄₂ (from ~39.2% to 12.8% after 30 min). These results indicate that BV2 cells can uptake different FITC-Aβ₄₂ forms in a TLR2-dependent manner.

Fig. 5 — a The phagocytosis of 0.5 μM sFITC-Aβ₄₂ and fFITC-Aβ₄₂, as well as preformed hPFAβ₄₂ (0.5 μM sFITC-Aβ₄₂ pretreated with 1.0 μM P2CSKn) in TLR2-WT and TLR2-knockdown BV2 cells, as determined by Flow cytometry. NC: negative control. b Effects of P2CSKn (0.5 μM) and different forms of Aβ₄₂ (0.5 μM) on the mRNA expression of *iNOS* and *Tnfα* in BV2 cells, as determined using rt-qPCR. hPAβ₄₂ was performed by treating 0.5 μM sAβ₄₂ with 1.0 μM P2CSKn. c Protein levels of MARCO, MSR1, TLR2 and NLRP3 in BV2 cells treated with fAβ₄₂ (0.5 μM), 1.0 μM P2CSKn, Pam2CSK4 or KLVFFn, and 1.0 μM P2CSKn, Pam2CSK4 or KLVFFn-pretreated sAβ₄₂ (0.5 μM), respectively. Cell lysate was harvested after 12 hr incubation and examined using Western Blot. d mRNA expression of *Marco* and *Msr1* in BV2 cells treated with fAβ₄₂ (0.5 μM), 1.0 μM P2CSKn, Pam2CSK4 or KLVFFn, as well as 1.0 μM P2CSKn, Pam2CSK4 or KLVFFn-pretreated sAβ₄₂ (0.5 μM), for 12 hr, respectively. mRNA level of each gene was determined by rt-qPCR. e Immunofluorescence of Iba1 in TLR2-WT or TLR2-knockdown BV2 cells treated with P2CSKn (1.0 μM), 0.5 μM sAβ₄₂, fAβ₄₂, and preformed hPAβ₄₂ by treating 0.5 μM sAβ₄₂ with 1.0 μM P2CSKn. f Phagocytosis of fFITC-Aβ₄₂ in BV2 cells with or without pretreatment of 1.0 μM P2CSKn or Pam2CSK4, respectively, as determined using Flow cytometry. P2CSKn or Pam2CSK4 was added to the cells for 6 h then washed before addition of fFITC-Aβ₄₂ for 12 h incubation with cells. Data are representative of three independent experiments with similar results in (e), or three independent experiments in (a–d, f). Data are presented as mean ± SD, n = 3 independent samples in (a-d,f) using one-way ANOVA with Dunnett’s post hoc test (b–f) or two-tailed t tests (a). ns, not significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Exact p values were given in the Source Data file. Source data are provided as a Source Data file.

P2CSKn activates BV2 cells and upregulates scavenger receptor MARCO and MSR1

We next examined the impacts of P2CSKn on the transcriptome profile of BV2 cells using RNA-seq analysis, and found several sets of genes involved in the polarization and phagocytosis signaling were affected in BV2 cells. Partially, P2CSKn or preformed hPAβ₄₂ significantly shifted the polarization of BV2 cells from M2 to M1 phenotype (Supplementary Fig. 2c, d). Accordingly, increased mRNA of iNOS and Tnfα (Fig. 5b) upon P2CSKn (1.0 μM) or hPAβ₄₂ (1.0 μM P2CSKn + 0.5 μM sAβ₄₂) treatment, and the protein level of inflammasome marker NLRP3 (Fig. 5c) upon P2CSKn or Pam2CSK4 treatment, were observed.

Moreover, RNA-seq analysis revealed that the expression of several phagocytosis-associated genes (Supplementary Fig. 1d-h; Supplementary Fig. 2a), and the relevant KEGG pathways (Supplementary Fig. 1f, i) were enhanced by P2CSKn or hPAβ₄₂ treatment, especially for Marco and Msr1 (Supplementary Fig. 2a) which were negatively modulated by Aβ₄₂ species (Fig. 1b). The upregulated mRNA of Marco and Msr1 by P2CSKn or Pam2CSK4 in the absence or presence of Aβ₄₂ species was further confirmed using rt-qPCR (Fig. 5d), which is consistent with the significant increase in the protein levels of MARCO, and MSR1. Moreover, a small upregulation of TLR2 was observed using western blot (Fig. 5c) or RNA-seq analysis (Supplementary Fig. 2b). These results suggest that TLR2 activation by P2CSKn or hPAβ₄₂ upregulates MARCO and MSR1 as Pam2CSK4 (Fig. 1d). Additionally, immunofluorescence imaging of Iba1, a biomarker of activated microglia, was conducted in TLR2-WT or knockdown BV2 cells. We noticed a pronounced increase of Iba1 expression in the BV2-NC cells treated with either P2CSKn (1.0 μM) or hPAβ₄₂ (1.0 μM P2CSKn + 0.5 μM sAβ₄₂), compared to that of 0.5 μM sAβ₄₂ or fAβ₄₂ treated group; in contrast, knocking down TLR2 significantly attenuated Iba1 expression (Fig. 5e) induced by either P2CSKn (1.0 μM) or hPAβ₄₂ (1.0 μM P2CSKn + 0.5 μM sAβ₄₂). These results suggested that BV2 cells were activated by P2CSKn in a TLR2-dependent manner, wherein scavenger receptor Marco, and MSR1 were upregulated.

TLR2 activation by P2CSKn moderately enhances phagocytosis of Aβ₄₂ species

We have shown above that Aβ₄₂ species induce a notable decrease of a panel of phagosome-related genes (Fig. 1b and Supplementary Fig. 1a-c), and hinder their sustained phagocytosis by BV2 cells (Fig. 4a–d). We also have shown Pam2CSK4 (Fig. 1d) and P2CSKn (Fig. 5c, d) can lead to upregulation of the mRNA expression levels of Msr1 and Marco, and activated phagocytosis in BV2 cells (P2CSKn, Fig. 4a–d; Pam2CSK4, Fig. 4b, c). These results suggest that TLR2 activation by P2CSKn contributes to its ability to promote phagocytosis of Aβ₄₂ species. To isolate the contribution of morphology-modification by P2CSKn in assessing the impact of TLR2 activation on phagocytosis of Aβ₄₂ species, BV2 cells were pre-treated with 1.0 μM P2CSKn or Pam2CSK4 for 6 h first, and washed off before addition of fFITC-Aβ₄₂ (0.5 μM), to activate the cells but prevent potential re-assembly of fFITC-Aβ₄₂ by P2CSKn or Pam2CSK4. After 12 h incubation with BV2 cells, we observed that the phagocytosis of fFITC-Aβ₄₂ is increased almost identically by about 26% (Fig. 5f. From 17.0 ± 2.0% to 42.8 ± 2.9% refer to P2CSKn, or to 43.8 ± 4.1% refer to Pam2CSK4) in both P2CSKn and Pam2CSK4 treated groups. These results indicated that activating TLR2 by P2CSKn without modifying fFITC-Aβ₄₂ moderately contributes to the P2CSKn-enhanced phagocytosis of Aβ₄₂ species in BV2 cells.

Morphology remodeling of Aβ₄₂ species by P2CSKn efficiently enhances Aβ-TLR2 interaction

We have shown that the uptake of Aβ₄₂ species by BV2 cells is TLR2-dependent (Fig. 5a) and P2CSKn activates BV2 cells in a TLR2-dependent manner (Fig. 5e). Since the morphology of Aβ₄₂ species can be modified by P2CSKn (Figs. 2, 3), we wondered whether this modification affects the Aβ-TLR2 interaction, thus we analyzed the co-localization of RhoB-P2CSKn, sFITC-Aβ₄₂ and TLR2 in BV2 cells. As shown in Fig. 6a, in RhoB-P2CSKn (1.0 μM) pretreated sFITC-Aβ₄₂ (0.5 μM) group, the majority of TLR2 is co-localized with the RhoB-P2CSKn and FITC-Aβ₄₂, which is rarely observed in the untreated sFITC-Aβ₄₂ group under the current conditions. The interaction of hPAβ₄₂ with TLR2 was further confirmed using cellular thermal shift assay (CETSA), where hPAβ₄₂ (2.0 eq. P2CSKn with sAβ₄₂) concentration-dependently stabilized the TLR2 stability under heating (Fig. 6b). These results suggest that P2CSKn promotes the interaction between Aβ₄₂ and TLR2 via morphology manipulation.

Fig. 6 — a The cellular distribution of TLR2 and sFITC-Aβ₄₂ (0.5 μM) with or without pretreatment with RhoB-P2CSKn (1.0 μM). Images were recorded using LSCM and colocalization analysis was performed using Image J (v 1.54g). b CETSA of TLR2 in BV2 cells in the presence of 2.0 eq. P2CSKn pretreated sAβ₄₂ examined by Western blot. c Cell viability of BV2 cells in the presence of various concentrations of P2CSKn for 24 h, as determined using CCK8 assay. d Effects of different compounds (1.0 μM) and Aβ₄₂ (0.5 μM) on the apoptosis in BV2 cells. Green dots indicate fluorescence released by activated capases-3 from GreenNucr Caspase-3 Substrate. e Expression of cleaved forms of PARP and caspase-3 in BV2 cells treated with sAβ₄₂ or fAβ₄₂, and P2CSKn, and P2CSKn pretreated sAβ₄₂, respectively, as measured using Western Blot after 12 h incubation. f Expression of cleaved forms of PARP in BV2 cells treated with different Aβ₄₂ species as measured using Western Blot after 12 h incubation. Treatment & Wash group refers to P2CSKn preincubated with cells and washed off before sAβ₄₂ (0.5 μM) addition. g Phagocytosis of 0.5 μM sFITC-Aβ₄₂ with or without compounds (1.0 μM) in BV2 cells, as determined using Flow cytometry. Normal group refers to sFITC-Aβ₄₂ pretreated with P2CSKn or Pam2CSK4 for 30 min before addition to cells; Treatment & Wash group refers to P2CSKn or Pam2CSK4 preincubated with cells and washed off before sFITC-Aβ₄₂ (0.5 μM) addition. Cells were incubated with different FITC-Aβ₄₂ samples for 12 h. Data are representative of three independent experiments with similar results in (a, d), three independent experiments in (b, f, g), two independent experiments in (e). Data are presented as mean ± SD, n = 3 (b, f, g), n = 6 (c) independent samples using one-way ANOVA with Dunnett’s post hoc test (c, d, f) or two-tailed t tests (g). ns, not significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Exact p values were given in the Source Data file. Source data are provided as a Source Data file.

Morphology remodeling of Aβ₄₂ species by P2CSKn efficiently attenuates Aβ₄₂ species-induced apoptosis in BV2 cells

As we mentioned in the Introduction, we proposed that altering the conformation of Aβ species would be capable of attenuating its toxicity. We found that P2CSKn didn’t show apparent cytotoxicity towards BV2 cells at concentrations up to 25 μM (Fig. 6c–e), while both sAβ₄₂ and fAβ₄₂ induced significant apoptosis of BV2 cells identified using the upregulated cleaved caspase-3 (Fig. 6d, e) and cleaved PARP (Fig. 6d). The preformation of hPAβ₄₂ (1.0 μM P2CSKn + 0.5 μM sAβ₄₂) effectively blocked Aβ₄₂ species-induced apoptosis of BV2 cells (Fig. 6d, e). Remarkably, when BV2 cells were pre-treated with P2CSKn (1.0 μM) for 6 hr, and washed off before addition of sAβ₄₂ (0.5 μM) to activate BV2 cells together preventing the formation of hPAβ₄₂ (termed “treat & wash” group), the Aβ₄₂ species-induced apoptosis remained unchanged (Fig. 6f). These results indicate that the morphological modification of Aβ species by P2CSKn, but not TLR2 activation, prevents the Aβ₄₂ species-induced apoptosis.

Morphology remodeling by P2CSKn ensures the high efficacy of phagocytosis of Aβ₄₂ species in BV2 cells

To identify the contribution of morphology remodeling by P2CSKn to the high efficacy in Aβ phagocytosis, BV2 cells were treated with 1.0 μM P2CSKn or Pam2CSK4 for 6 h, respectively, then washed before addition of sFITC-Aβ₄₂ (termed “treat & wash”) group, and the phagocytosis was investigated. As depicted in Fig. 6g, the phagocytosis of FITC-Aβ₄₂ in treat & wash group (25.7 ± 2.7%) was reduced ~68% compared to that of 0.5 μM sFITC-Aβ₄₂ pretreated with 1.0 μM P2CSKn (normal group, 94.1 ± 0.3%). Moreover, although the normal group of P2CSKn (94.1 ± 0.3%) showed significantly higher activity than that of Pam2CSK4 (49.3 ± 4.1%), both P2CSKn and Pam2CSK4 in the treatment and wash group showed almost identical low phagocytotic efficacy (25.7 ± 2.7% vs. 24.0 ± 2.3%, respectively). These results suggest that preventing the formation of hPFAβ₄₂ hybrid aggregates greatly impairs the high efficacy of P2CSKn in promoting Aβ₄₂ species phagocytosis, the morphological modification of Aβ₄₂ species substantially contributes to the P2CSKn-enhanced uptake of Aβ species in BV2 cells.

Local injection of P2CSKn reduces Aβ burden in brains of APP/PS1 transgenic mice

To investigate the potency of P2CSKn in the native brain environment, we first injected P2CSKn or reference molecules mixed with Hoechst 33342 into the cortex of 13-month-old APP/PS1 transgenic mice, which have high expression of Aβ and extensive amyloid plaques. The Aβ plaque and activation status of microglia and astroglia were investigated using immunofluorescence 2- or 5-days post-injection. We used antibody OC and fluorescent dye Thioflavin S (TS) to identify all Aβ aggregation forms and Aβ aggregates with high abundance of β-structures, respectively, in the injected area identified by Hoechst 33342 (Supplementary Fig. 12a). As shown in Supplementary Fig. 13, the TS fluorescence was localized to the core of the aggregates which was identified by antibody OC, consistent with the literatures⁵⁵. Compared to the vehicle group, we observed a significant increase in the area positive for “OC” and “OC + TS-“ in the injection area (as indicated by Hoechst 33342) in the P2CSKn group at 2 d after injection, and a significant decrease at 5 d after injection. In the same set of mice, there was no notable change in the TS fluorescence intensity at 2 d after injection, but it exhibited a significant decrease at 5 d after injection. Since TS cannot recognize recombinant Aβ/P2CSKn co-aggregates, the increased staining in “OC” and “OC + TS-” is most likely caused by the formation of Aβ/P2CSKn co-aggregates during the disaggregation and reassembly of Aβ plaques, or by P2CSKn-induced elevated aggregation of soluble Aβ. The significant reduction in the staining of “OC” and TS fluorescence intensity at 5 d after injection suggests that P2CSKn facilitates the clearance of Aβ species. We also observed a significant increase in Iba1 staining associated with an increased OC staining; and this increase became much lower at 5 d after injection, together with a reduced OC staining. We then repeated the experiments in more mice and examined the levels of Aβ plaques (TS and OC), and microglia marker (Iba1) and astrocytic marker (GFAP), following P2CSKn treatment. We calculated the ratio of averaged fluorescence areas in the injection area over its adjacent areas (as a control), to evaluate compounds-induced changes. As shown in Fig. 7a, injection of P2CSKn significantly reduced the Aβ plaque identified by TS and antibody OC (INJ/ADJ_TS = 0.57 ± 0.10; INJ/ADJ_OC = 0.55 ± 0.13), while PBS (vehicle) injection (INJ/ADJ_TS = 1.0 ± 0.16; INJ/ADJ_OC = 1.0 ± 0.17) and Pam2CSK4 (INJ/ADJ_TS = 0.85 ± 0.11; INJ/ADJ_OC = 1.05 ± 0.28) didn’t induce significant alterations. These findings confirmed that P2CSKn effectively promotes the clearance of Aβ plaques in the cortex of APP/PS1 transgenic mice. Moreover, as a TLR2 agonist, Pam2CSK4 significantly increased Iba1 level (INJ/ADJ_Iba1=5.79 ± 1.71) compared to P2CSKn (INJ/ADJ_Iba1=3.09 ± 0.91), but had much less effect on reducing Aβ plaque level (INJ/ADJ_TS=0.85 ± 0.11; INJ/ADJ_OC=1.05 ± 0.28) also compared to P2CSKn (INJ/ADJ_TS=0.57 ± 0.10; INJ/ADJ_OC=0.55 ± 0.13).

Fig. 7 — a The effect of P2CSKn or Pam2CSK4 and PBS control on Aβ, Iba1 and GFAP level in the brain of APP/PS1 transgenic mice. Compounds or PBS were mixed with Hoechst33342 (for visualizing the injection area) and injected into the cortex, respectively. Staining and/or imaging were performed using TS, OC, antibodies against Iba1 or GFAP 5d post-injection, and the stained fluorescence area in the injection site was used for calculation. b P2CSKn (right hemisphere) or RP (left hemisphere) were mixed with Alexa Fluor™ 594 (for visualizing the injection area) and injected into the hippocampus. Staining and/or imaging were performed using TS, antibodies against cleaved caspase 3, Iba1 or GFAP 5 days after injection, and the stained fluorescence intensity in the injection site and its adjacent area are used for calculation. Results were presented as injected area/adjacent area, with RP for comparison. OC, Aβ fibril specific antibody; TS, Thioflavin S; AF594, Alexa Fluor™ 594; RP, Random Peptide (Pam-KCKSKVFLKFK-ahx-nal). Data presented as mean ± SD, n = 4 (a, PBS), n = 15 (a, P2CSKn); n = 5 (a, Pam2CSK4), n = 9 (b, P2CSKn), n = 9 (b, P2CSKn) mice using one-way ANOVA with Dunnett’s post hoc test (a) or two-tailed t tests (b). ns, not significant. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Exact p values were given in the Source Data file. Source data are provided as a Source Data file. The mouse and mouse head schema were sourced from the SciDraw website (https://scidraw.io/). Tyler, E., & Kravitz, L. (2020). mouse. Zenodo. 10.5281/zenodo.3925901. Petrucco, L. (2020). Mouse head schema. Zenodo. 10.5281/zenodo.3925903.

We further investigated whether the above changes can be observed in other brain regions and at another age in AD transgenic mice. Thus, we used 9-month-old APP/PS1 mice, and injected P2CSKn or reference random peptide (RP) mixed with Alexa Fluor^TM 594 (AF594) into the right and left hippocampus of mice, respectively (Supplementary Fig. 12b, Fig. 7b). The Aβ plaque burden (TS) and Iba1 and GFAP, and apoptosis level (cleaved caspase-3) were examined 5 days post-injection. We calculated ratios of the averaged fluorescence intensity in the injection area over its adjacent neighboring area (indicated by AF594). As shown in Fig. 7b, compared to the RP group (INJ/ADJ_TS = 1.12 ± 1.04), injection of P2CSKn significantly reduced Aβ plaque area (INJ/ADJ_TS = 0.13 ± 0.20), further confirming that P2CSKn effectively promotes the clearance of Aβ plaque in the hippocampus of APP/PS1 mice. It is worth noting that, comparing to the adjacent region, no significant increase in Iba1 or GFAP fluorescence, or cleaved caspase-3 was observed in the P2CSKn-injected area, with no significant difference found between P2CSKn and RP group (Fig. 7b). As we have shown elevated Iba1 level in the cortex of 13-month-old APPPS1 mice upon P2CSKn injection, these paradoxical results implicate an age-dependent or a region-specific acceleration of initial microglial responses. Moreover, we observed Aβ plaque, microgliosis, IL-1β secretion and NLRP3 expression in the hippocampus of 16-month-old APP/PS1 mice at 5- or 10-days after injection of P2CSKn or RP. The results further confirmed that P2CSKn significantly reduced Aβ plaque, but didn’t alter Iba1 (Supplementary Fig. 14a) or secreted IL-1β (Supplementary Fig. 14b) level, at both time points. However, we observed an increase in NLRP3 inflammasome expression in the hippocampus 10 d post P2CSKn injection using western blot (Supplementary Fig. 14c). The above findings collectively confirm that localized delivery of P2CSKn into either hippocampus or cortex significantly reduces Aβ plaque burden, and implicated a region-dependent immunomodulatory mechanism underlying Aβ clearance efficacy of P2CSKn.

To evaluate the translational potential of systemic P2CSKn delivery, we first assessed the stability of P2CSKn in the murine serum, observing progressive degradation over time with detectable levels of intact P2CSKn remaining after 24 hr incubation (Supplementary Fig. 15a). We then mapped its biodistribution in C57BL/6 J (n = 3) mice following intraperitoneal (i.p.) injection of RhoB-P2CSKn. Significantly higher cerebral accumulation of RhoB-P2CSKn than RhoB was observed 24 h after injection (Supplementary Fig. 15b), indicating a blood-brain barrier (BBB) penetration capacity of P2CSKn. Regrettably, in acute toxicity evaluation of P2CSKn (15 mg/kg, i.p. injection at 0 and 48 h), significant mortalities occurred in the APP/PS1 cohort (n = 9) with 2 fatalities within 48 hr after the first injection, and 5 additional fatalities after the second injection, resulting in only 2 surviving (Supplementary Fig. 16a). The mortality precluded systemic efficacy assessment of P2CSKn in the AD model.

The high mortality in AD models likely results from P2CSKn-induced amplification of systemic inflammatory responses. To test this possibility, we performed additional experiments. No mortality has been observed in wild-type mice (4 months, C57BL/6 J, n = 8) using the same experimental protocol as above (Supplementary Fig. 16a). We have observed hypoactivity in mice after both injections. In addition, in C57BL/6 J mice euthanized at 72 h post-injection, we observed multiple organ pathological alterations. Serum analysis revealed significant reductions in ALT, urea, and uric acid, alongside with elevated creatinine, with AST levels unaffected (Supplementary Fig. 16b). Importantly, histopathological analysis demonstrated hepatic architectural disorganization with hemorrhagic foci, renal tubular dilation, cardiac myofibril disarray featuring loose fiber alignment and focal fragmentation in P2CSKn-treated mice comparing to saline-injected controls (Supplementary Fig. 16c). In addition, we observed significant upregulation of TNF-α, IL-6, and IL-1β levels in the serum of C57BL/6 J mice at 2 or/and 4 hr post- injection, associated with significant upregulation of IL-10 level (Supplementary Fig. 16d). The above observations collectively indicate that P2CSKn-induced amplification of systemic inflammatory responses likely leads to organ injury and high mortality in the AD model mice. The absence of lethality in the young wild-type mice may reflect an age- and disease-dependent ability to cope with the TLR2-mediated body inflammatory cascades and potential toxicity.

Discussion

An effective clearance represents an important avenue for the development of therapeutics in Aβ-related diseases. Aβ uptake has been associated with microglia activation^56,57, plaque growth and subsequent microglia death^58,59, and the impaired ability of microglia to engulf Aβ^14,60, which attributes to reduced expression of phagocytic receptors, impaired enzymatic activity in lysosomes, and alterations in phagocytosis signaling pathways^61,62. In the current study, we observed that Aβ species induce significant apoptosis, reduce the expression of phagosome-relevant genes, and impair phagocytosis in BV2 cells. To gain novel insights to advance the development of therapeutics for treating Aβ-related diseases, we developed a chimaera P2CSKn integrating an Aβ-remodeling module and a microglia-activating module, and focused on assessing the importance of conformational regulation of Aβ species and activation of microglia on the clearance of amyloid proteins. We demonstrated that P2CSKn efficiently reduced the heterogeneity of Aβ species by either co-assembling with the soluble Aβ or re-assembling with the preformed Aβ fibrils to form conformation-modified P2CSKn/Aβ₄₂ hybrid aggregates and activated microglia in a TLR2-dependent manner. We provided compelling evidence that P2CSKn significantly enhances phagocytosis of Aβ species in a consistent and persistent manner in various in vitro and local-injection in vivo models.

We have predicted that remodeling the conformation of Aβ species will facilitate the blockade of their interactions with pathological targets, thereby mitigating the deleterious effects of Aβ species. As we have shown here, by forming monodisperse hPAβ₄₂, P2CSKn efficiently eliminates the Aβ oligomers and rescues the Aβ-induced apoptosis, and contributes to the long-lasting phagocytosis of Aβ proteins in the BV2 cells. It is worth pointing out that, by activating cells but preventing the morphology remodeling of Aβ₄₂ species, P2CSKn increases the phagocytosis by approximately only 26% (Fig. 5f) compared to controls, but decreases the phagocytosis by approximately 68% (Fig. 6g) compared to the preformed hPFAβ₄₂ group. These observations indicate that the morphology remodeling of Aβ₄₂ species by P2CSKn constitutes the most important process in the persistent phagocytosis of Aβ₄₂ species, which is likely attributed to the capacity of P2CSKn in enhancing Aβ-TLR2 interaction, and alleviation of Aβ₄₂ species-induced apoptosis.

On the other hand, though the TLR2-mediated the uptake of different Aβ₄₂ species (Fig. 5a), however, additional activation of TLR2 using its agonist seems to be risky. As we have shown here, the TLR2-activation capacity of P2CSKn enhances Aβ clearance efficiency at low to moderate levels. Mechanistically, TLR2 signaling counteracts the Aβ-induced suppression of scavenger receptors Msr1 and Marco, a mechanism we initially demonstrated using the TLR2 agonist Pam2CSK4. P2CSKn recapitulates this pathway in the BV2 cells, restoring phagocytic receptor expression and amplifying Aβ engulfment, indicating TLR2 activation as a potential enhancer of Aβ clearance. However, this benefit is counterbalanced by a critical liability. TLR2-driven microglia polarization towards the pro-inflammatory M1 phenotype, and the resulting sustained pro-inflammatory cytokine release, confer significant clinical risk. Indeed, while localized administration may mitigate long-term microgliosis in animals, systemic delivery reveals a pathological vulnerability evidenced as a boosted pro-inflammatory cascade and resulting damage of liver, kidney and heart. The apparent discrepancy between systemic and CNS toxicity profiles most likely stems from disease context and exposure differentials. While i.p. dosing of P2CSKn induced systemic toxicity, the low-dose CNS administration may limit the extent of P2CSKn exposure to brain cells and also avoids significant peripheral TLR2 engagement. On the other hand, the high mortality in aged APP/PS1 mice vs. no lethality in young wild-type mice after i.p injection of P2CSKn implicates an age-dependent and/or disease-related reduction in the ability to tolerate TLR2-mediated body inflammatory cascades and the resulting toxicity. For instance, it is likely that the contributions from neurovascular dysfunction (e.g., impaired BBB), primed glia and peripheral inflammation are synergistically enhanced by i.p. dosing of P2CSKn in AD mice.

Collectively, these mechanistic insights from the dual-action profile of P2CSKn chart a strategic course for future Aβ-targeted therapeutic development. First, the dominant contribution of Aβ morphological remodeling suggests that prioritizing molecular scaffolds that amplify this pathway while minimizing innate immune system engagement will be a rational choice. Second, should context-dependent TLR2 co-activation be proven necessary, such immune modulation needs to be architecturally constrained through delivery platforms to achieve spatial-temporal confinement. From this perspective, harnessing P2CSKn or its optimized derivatives to mitigate the pathogenesis of Aβ species would be promising in advancing relative therapeutic strategies. Much more work is needed, based on the proof-of-principle we have provided here, to devise much more improved drug-like molecules, devoid of the toxicity shown here, as future clinical testing candidates in the forthcoming research endeavors.

Methods

Ethical statement

Animal experiments have been approved by the Peking University Shenzhen Graduate School Animal Care and Use Committee (Permit Number: AP0011) and were in accordance with the ARRIVE guidelines on the Care and Use of Experimental Animals. The experimental mice were housed in an isolated SPF facility, and maintained at a temperature of ~25 °C in a humidity-controlled environment with a 12 h light/dark cycle, and with free access to standard food and water.

General

P2CSKn and other compounds are synthesized in our group as described in the Supplementary Methods. P2CSKn-based experiments comply with the community requirements (Box 1 in https://www.nature.com/articles/nchembio.1867). BV2 cells (authenticated by STR analysis) were purchased from Procell Life Science & Technology Co. Ltd. Aβ₄₂ (AG970) was purchased from Millipore, FITC-Aβ₄₂ were purchased from GL Biochem (Shanghai) Ltd. Tlr2 siRNAs were obtained from Synbio Technologies (Suzhou, China). Mouse anti-amyloid β oligomer-specific monoclonal antibody, A11 (Cat # AS10932) was from Agrisera. Mouse anti-Aβ 17-24 Antibody, 4G8 (Cat # SIG-39200) and mouse anti-Aβ 1-16 Antibody, 6E10 (Cat # SIG-39300) were from Biolegend. Rabbit anti-Toll-like Receptor 2 (Cat # 13744), rabbit anti-GAPDH antibody (Cat # 2118), rabbit anti-PARP antibody (Cat # 9542), rabbit anti-NLRP3 antibody (Cat # 15101), rabbit anti-caspase 3 antibody (Cat # 9662), and rabbit anti-cleaved caspase 3 antibody (Cat # 9664) were from Cell Signaling Technology. Rabbit anti-Iba1 antibody (Cat # ab178846), rabbit anti-GFAP (Cat # ab33922), rabbit anti-Toll-like Receptor 2 antibody (Cat # ab11864) and rabbit anti-MARCO antibody (Cat # ab259264) were from Abcam. Rabbit anti-MSR1 antibody (Cat # DF6694) was from Affinity. Rabbit Anti-Amyloid Fibrils OC Antibody (Cat # AB2286) was from Sigma-Aldrich. Rabbit anti-β tubulin antibody (10094-1-AP) and Rabbit anti-CD11b (FITC tag) antibody (FITC-65055) were from Proteintech. GreenNuc™ Caspase-3 Assay Kit for Live Cells was purchased from Beyotime Biotechnology. APPswe/PSEN1dE9 mice with a C57BL/6 background were originally obtained from the Jackson laboratory and bred subsequently at PKUSZ. Unless otherwise specified, all experimental mice in this study were male. APP/PS1 mice and wild-type littermate mice were genotyped by PCR analysis of genomic DNA.

Preparation of Aβ₄₂/FITC-Aβ₄₂ samples

Disassembly: The purchased Aβ₄₂ or FITC-Aβ₄₂ were dissolved in 1,1,1,3,3,3-Hexafluoro-2-propanol (HFIP) overnight and then lyophilized according to the literature⁶³. The lyophilized peptides were stored at −80 °C before use. sAβ₄₂ and sFITC-Aβ₄₂: Unless specially mentioned, the soluble Aβ₄₂ or FITC-Aβ₄₂ was prepared according to the following conditions. The disassembled and lyophilized Aβ₄₂ or FITC-Aβ₄₂ peptides were dissolved in 1% aqueous ammonia to 1 mg/mL and diluted to the final concentration with 50 mM PBS (for molecular level assays), or were dissolved in DMSO to 200 μM and diluted to the final concentration with DMEM (For cellular experiments). fAβ₄₂ and fFITC-Aβ₄₂: Unless specially mentioned, the pre-formed fibrillar Aβ₄₂ or FITC-Aβ₄₂ was obtained by incubating a 20 μM sAβ₄₂ or sFITC-Aβ₄₂ solution at 37 °C for 36h, and diluted with 50 mM PBS or DMEM corresponding to molecular or cellular assays, respectively, to the indicated concentration before use. hPAβ₄₂ and hPFAβ₄₂: Unless specially mentioned, the hybrid aggregates of P2CSKn with Aβ₄₂ or FITC-Aβ₄₂ were prepared by incubating 2.0 eq. P2CSKn with 0.5 μM sAβ₄₂ or fFITC-Aβ₄₂ at 37 °C for 30 min before use.

Cell culture

TLR-WT or knockdown BV2 cells or primary microglia cells were cultured in Dulbecco’s modified Eagle medium (DMEM) (Gibco) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Gibco) and 1% penicillin/streptomycin at 37 °C with 5% CO₂.

Tlr2 siRNA transfection in BV2 cells

BV2 cells were plated in 6-well plates (100,000 cells per well). Opti-MEM (990 μL), siRNA (5.0 μL, final concentration: 20 nM) and Lipofectamine 2000 (5.0 μL) were mixed for 5.0 min and the mixture was transfected to BV2 cells for 8.0 h and the cells were washed with DMEM and cultured with DMEM (10% FBS) for another 16 h. The cells were collected for validation of TLR2 knockdown by Western blot or further experiments. Tlr2 siRNA (si-2) (sequence 5’ to 3’, sense: GCGAACUCCUAUCCUUUACdTdT, antisense: GUAAAGGAUAGGAGUUCGCdTdT) and negative control were from Synbio Technologies.

Mouse primary microglia cells

Isolation and culture of primary microglia cells from ten-month-year-old-C57/BL6 mice were performed according to the protocol from the Bio-protocol website⁵⁴. In brief, the skull was cut carefully peeled off without damaging the brain tissue. The brain was gently lifted using sterilized curved forceps and placed into a large dish that had been pre-filled with D-Hank’s solution, followed by removing the inner layer of the meninges. The brain was transferred to a large dish containing 0.05% trypsin (with EDTA), and minced into a paste-like consistency for enzymatic digestion culture in a CO₂ incubator. After termination of trypsin digestion by DMEM medium containing 10% FBS, the mixture was centrifuged (200 × g, 8 min) using a nylon cell filter (70 μm). The cell pellet was then resuspended with DMEM medium containing 10% FBS in a culture flask and cultured for 3 days. Then the supernatant of culture medium was centrifuged (200 × g, 8 min), and mixed with an equal volume of fresh DMEM medium containing 10% FBS, and added back to the cells. The cells were cultured for 10 days with repeated medium change procedures during this period. When the confluence of the mixed glial cells reached about 90%, the cells were suspended on a shaker at 20 × g for 2 hours. The suspension was centrifuged (200 × g, 8 min) and the cells were resuspended with DMEM medium containing 10% FBS, and cultured for 3 days. The shaking and centrifuge procedure were repeated and the supernatant was mixed with an equal volume of fresh DMEM medium containing 10% FBS, and added back to the cells, cultured for 8 days (repeat this medium change procedure during this period). The cells were expanded for subsequent experiments after determination of cell purity by flow cytometry using FITC-CD11b antibody. The FACS gating strategy to detect CD11b-FITC signals is shown in Supplementary Fig. 17b.

Tyndall effect

sAβ₄₂ (500 μL, in PBS buffer) and different compounds (500 μL, in PBS buffer) with gradient concentrations were mixed rapidly in glass vial, where a red laser beam was penetrated and imaged immediately.

Dynamic light scattering

sAβ₄₂ (20 μM, 500 μL, in PBS buffer) and different compounds (500 μL, in PBS buffer) with gradient concentrations were mixed rapidly and centrifuged (200 × g, 10 min), the supernatant was taken for particle size test by DynaPro Plate Reader II (Wyatt, America). fAβ₄₂ (final concentration: 20 μM, 500 μL, in PBS buffer) and P2CSKn (final concentration: 20 μM) were incubated for gradient time points. The whole sample or the supernatant was taken for particle size test by DynaPro Plate Reader II.

Thioflavin fluorescence kinetics

sAβ₄₂ (10 μM) with or without P2CSKn and ThT (5.0 μM, in 50 mM PBS) were added into the well of 384-black-transparent plate with lid (Thermo Fisher Scientific, Nunc 142761). To observe the effect of P2CSKn on preformed β-sheet, sAβ₄₂ (10 μM) and ThT (5 μM, in 50 mM PBS) were incubated as above for 32 h before P2CSKn (final 50 μM) was injected into the well. The plate was incubated at 37 °C, and the fluorescence intensity was measured every hour by using excitation and emission wavelengths of 444 and 484 nm, respectively.

Circular dichroism

Analytical stock solutions (10 mM) of P2CSKn were prepared in ethylene glycol and diluted to the final concentrations in 50 mM phosphate buffer (pH 7.4, excluding NaCl and KCl). The disassembled and lyophilized Aβ₄₂ powder was dissolved in ethylene glycol to 200 μM, and diluted to 10 μM in 50 mM phosphate buffer (pH 7.4, excluding NaCl and KCl) to afford sAβ₄₂ form, which was either incubated with different concentrations of P2CSKn at 37 °C for 24 h, or incubated at 37 °C for 36 h to form fAβ₄₂ before addition of different concentrations of P2CSKn for another 12 h incubation, respectively. CD measurements were performed on the Chirascan CD spectrophotometer (Applied Photophysics, UK) at room temperature in a 1 mm optical path length cell without dilution, and the spectra were recorded over a wavelength range of 190–260 nm at a 1-nm bandwidth, 0.5-nm step size, and 0.5-s per point. The spectra of the blank solutions were auto-subtracted. Final analysis of the data was conducted with Chirascan software.

Microscale thermophoresis

P2CSKn (10 mM) was dissolved in DMSO and diluted to the test concentration with 50 mM PBS. sFITC-Aβ₄₂ (0.5 μM) was dissolved in 50 mM PBS and incubated with different compounds respectively at room temperature for 10 min. Samples were loaded into standard/premium-treated capillaries. Measurements were performed using 20% MST power with laser off/on times of 4 s and 5 s, respectively with Monolith NT.115 (NanoTemper Technologies GmbH). The data were processed by MO. Affinity Analysis 2.3.0 software.

Dot blot of Aβ₄₂ oligomer

10 μL sAβ₄₂ (final concentration: 4.0 μM) with 10 μL indicated final concentration P2CSKn in PBS in the tubes were incubated at 37 °C for indicated time. For each sample, 180 μL of PBS (10 mM) was added, samples were spotted and absorbed on the surface of a nitrocellulose (NC) membrane by Bio-Dot^® Microfiltration System (Bio-rad). Then the NC membrane was blocked by non-fat milk (5%, w/v, TBST) for 2 h at 25 °C. The membrane was incubated overnight at 4 °C with 1:1000 diluted anti-Aβ oligomer antibody A11 (Agrisera). The membrane was then washed three times with TBST for 10 min each and incubated horseradish peroxidase-(HRP-) conjugated goat anti-mouse antibody (1:10000) in non-fat milk (5%, w/v, TBST) for 2 h at room temperature. The membrane was then washed three times with TBST (0.5% Tween 20) for 10 min each and imaged by electrochemiluminescence (ECL) solution with Tanon 5200.

Laser scanning confocal microscopy

Fluorescent imaging of samples by LSCM was recorded using FV3000 (Olympus), at Ex/Em of 577/590 nm for LysoTracker Red, 546/570 nm for RhoB, 488/520 nm for FITC, 500/530 nm for GreenNuc Caspase-3 substrate, 652/672 nm for Dylight 649 conjugated-secondary antibody of TLR2, 555/568 nm for Dylight 549 conjugated-secondary antibody of Iba1. sFITC-Aβ₄₂ co-assembly with RhoB-P2CSKn: Freshly prepared sFITC-Aβ₄₂ (final concentration: 0.5 μM) and RhoB-P2CSKn (final concentration: 0.5 μM) were incubated for 12 h at 37 °C. The morphology of hybrid aggregates of RhoB-P2CSKn and FITC-Aβ₄₂ was captured by FV3000. The colocalization was analyzed by Image J software (v 1.54g). fFITC-Aβ₄₂ re-assembly by RhoB-P2CSKn: Freshly fFITC-Aβ₄₂ (final concentration: 5.0 μM) was added onto the glass bottom dish to find appropriate microscope vision, and then RhoB-P2CSKn or Pam2CSK4 or KLVFFn (final concentration: 25 μM, PBS) was injected at a corner of the dish. Continuous confocal recording (every 1 min) was conducted immediately after the injection by LSCM. The colocalization of RhoB-P2CSKn in FITC-Aβ₄₂ plaque (Fluorescence intensity of RhoB within FITC area) was analyzed by Photoshop and Image J software (v 1.54g). Dynamic phagocytosis of FITC-Aβ₄₂: Freshly prepared sFITC-Aβ₄₂ (0.5 μM) was incubated with either P2CSKn (0.5 μM) or RhoB-P2CSKn (0.5 μM) in DMEM for 0.5 h at 37 °C, respectively, to afford corresponding hybrid aggregates. BV2 cells were plated in glass-bottom dishes (20,000 cells per dish) and treated with hybrid aggregates or fFITC-Aβ₄₂ (0.5 μM). The phagocytosis process was recorded immediately after administration by LSCM for every minute for 1 h. Cellular colocalization of FITC-Aβ₄₂/P2CSKn/Lysosome: BV2 cells plated in glass bottom dishes (20,000 cells per dish) was pretreated with or without 10 μM chloroquine for 1 h before the treatment of the P2CSKn/FITC-Aβ₄₂ sample, which was prepared by mixing P2CSKn (0.5 μM) with freshly sFITC-Aβ₄₂ (0.5 μM) and incubated for 0.5 h at 37 °C. After 2 h incubation of cells with P2CSKn/FITC-Aβ₄₂ sample, LysoTracker Red (50 nM) was added into the dishes for incubation 30 min. The intracellular localization of P2CSKn/FITC-Aβ₄₂ was captured by FV3000. Cellular colocalization of FITC-Aβ₄₂/RhoB-P2CSKn/TLR2: Freshly sFITC-Aβ₄₂ (0.5 μM) with or without RhoB-P2CSKn (1.0 μM) incubated for 0.5 h at 37 °C first, and was added to and incubated with BV2 cells plated in glass-bottom dishes (20,000 cells per dish) for 1 h. The cells were then washed with PBS three times and fixed with 4% paraformaldehyde for 30 min at room temperature. Then the cells were washed three times with PBS and permeabilized and blocked with QuickBlock^TM Blocking Buffer for Immunol Staining (Beyotime, China) for 1 h at room temperature. Then the cells were washed three times with PBS and incubated with primary antibody TLR2 (Cell Signaling Technology, 1:1000) at 4 °C overnight. After washing with PBS, cells were incubated with secondary antibody Dylight 649 goat anti-rat (Thermo Scientific, 1:500) at for 1 h at room temperature. The fluorescence from different makers in cells was imaged by FV3000.

Transmission electron microscopy

20 μM fAβ₄₂ or fFITC-Aβ₄₂ were treated with or without P2CSKn (40 μM, PBS) for the indicated time, then the mixtures were spotted onto the surface of carbon-coated copper grid for 3 min. The copper grid was stained with phosphotungstic acid (1%, 10 μL) for 2 min, and excess staining solution was dried off with filter paper. Images were recorded by a transmission electron microscope HT7700 (Hitachi).

Western blot

The harvested protein samples were electrophoresed on Tris-Glycine gels (15%, for PICUP samples; 10%, for cellular proteins). Proteins were transferred onto a PVDF membrane (Millipore). The membrane was blocked by non-fat milk (5%, w/v, TBST) for 2 hr at 25 °C. The membrane was incubated overnight with the primary antibody. The membrane was then washed three times with TBST (0.5% Tween 20) for 10 min each and incubated with HRP-conjugated antibody (1:10000) in non-fat milk (5%, w/v, TBST) for 2 h at room temperature. The membrane was then washed three times with TBST (0.5%Tween 20) for 10 min each and imaged using ECL solution. For TLR2/MARCO/MSR1/NLRP3/PARP/Caspase 3 Proteins level: The freshly prepared fAβ₄₂ (0.5 μM) or sAβ₄₂ (0.5 μM) were incubated with or without P2CSKn (1.0 μM), Pam2CSK4 (1.0 μM) or KLVFFn (1.0 μM) for 0.5 h at 37 °C, and were added to the BV2 cells for 12 h incubation. After that, cells were washed with PBS and lysed with RIPA. Cell lysate was collected, and cell debris, together with precipitated and aggregated proteins, was removed by centrifuging samples at 12000 × g for 25min at 4 °C. After the addition of Loading buffer, the samples were boiled at 95 °C for denaturation and analyzed by Western blot. For extracellular Aβ₄₂ species determination: BV2 cells were plated in 6-well plate (20,000 cells per well), and treated with 0.5 μM sAβ₄₂, fAβ₄₂ and P2CSKn (0.5 μM, 1.0 μM, 2.0 μM) pretreated sAβ₄₂ (0.5 μM, 30 min at 37 °C) for 6 h incubation. The culture supernatant was collected and lyophilized. Lyophilized supernatant was resuspended in PBS (100 μL). Supernatant solution (20 μL) was added with loading buffer (5.0 μL) and boiled at 95 °C for denaturation and analyzed by western blot using 6E10 antibody. The uncropped and unprocessed scans of the blots are presented in the Source Data file.

Photo-induced cross-linking of unmodified proteins

The freshly prepared fAβ₄₂ were incubated with or without P2CSKn, Pam2CSK4 or KLVFFn in gradient concentrations for different times in 50 mM PBS at 37 °C. After each time interval, ammonium persulfate (APS, 1.0 μL, 40 mM) and Tris (2,2’-bipridyl) dichlororuthenium (II) (1.0 μL, 2.0 mM) were added to the prepared samples (18 μL) in a clear PCR tube. The mixture was then irradiated immediately for 1s using a 200 W incandescent lamp at a distance of 10cm from the bottom of the PCR tube, and was quenched immediately by adding 5.0μL 5 × loading buffer containing 5% β-mercaptoethanol. The mixture was boiled at 95 °C for denaturation and analyzed by Western blot.

Cellular thermal shift assay

BV2 cells with 70 to 80% confluence in 15-cm culture dish were treated with P2CSKn, or preformed hPAβ₄₂ with different concentrations (ratio 2:1) or vehicle (DMSO) for 2 h. Cells were harvested and washed once with PBS, then suspended in 1 mL of PBS supplemented with proteinase and phosphatase inhibitors (Beyotime). 80 μL cell suspension was distributed into eight 0.2 mL PCR tubes respectively, and heated at their designated temperatures for 5 min in AB 96-well thermal cycler. Immediately after heating, tubes were removed and incubated in the ice for 5 min. After that, tubes were immediately snap-frozen in liquid nitrogen to lyse the cells, five freeze and thaw cycles in liquid nitrogen were performed. The tubes were vortexed briefly after each thawing. Cell lysate was collected, and cell debris, together with precipitated and aggregated proteins, was removed by centrifuging samples at 12,000 × g for 25min at 4 °C. Loading buffer was added to the samples and boiled at 95 °C for denaturation and analyzed by Western blot.

Immunofluorescence of cellular Iba1

0.5 μM freshly prepared sAβ₄₂ or fAβ₄₂ were incubated with different compounds (1.0 μM) or vehicle (DMSO) for 0.5 h at 37 °C first, and each sample or vehicle (DMSO) was added and incubated with BV2 cells plated in glass bottom dishes (20,000 cells per dish) for 12 h, respectively. Then the cells were washed with PBS and fixed with 4% paraformaldehyde for 30 min at room temperature. The cells were washed again by three times with PBS, and permeabilized and blocked with QuickBlock^TM Blocking Buffer for Immunol Staining (Beyotime, China) for 1 h at room temperature, and were further washed three times with PBS and incubated with primary antibody Iba1 (Cell Signaling Technology, 1:1000) at 4 °C overnight. After washing with PBS, cells were incubated with DAPI and secondary antibody Dylight 549 goat anti-mouse (Thermo Scientific, 1:500) for 1 h at room temperature. Cells were imaged by FV3000.

Cell viability and apoptosis assay

BV2 cells were plated in 96-well plate (5500 cells per well) and treated with P2CSKn in gradient concentrations for 24 h. Then CCK-8 solution (10 μL) was added to each well and incubated for 2.0 h. The absorbance (OD value) at 450 nm was measured by a microplate reader. Cell viability (%) = [Absorbance (P2CSKn)-Absorbance (DMEM)]/ Absorbance (Blank)-Absorbance (DMEM)] × 100. Caspase-3 activation in living cells: BV2 cells were plated in glass-bottom dishes (20,000 cells per dish), 0.5 μM freshly prepared sAβ₄₂ or fAβ₄₂ with or without pretreatment of different compounds (1.0 μM, 0.5 h at 37 °C) was added to cells for 12 h incubation at 37 °C. Then cells were washed twice with DMEM, and GreenNuc Caspase-3 Substrate (5 μM) was added to the cells for 30 min. Afterward, the cells were washed again twice with DMEM, and PBS was added to the cells. Images were immediately captured using a laser confocal microscope (Ex/Em: 500 nm/530 nm). Caspase-3 and PARP activation by western blot: BV2 cells were plated in 6-well plates (100,000 cells per well). 0.5 μM fAβ₄₂ or sAβ₄₂, or preformed hPAβ₄₂ by treating 0.5 μM sAβ₄₂ with P2CSKn (0.5, 1.0, 2.0 μM; 0.5 h at 37 °C), were subjected to the cells for 12 h incubation. Cells were collected, and the protein level of PARP, Caspase 3 and their cleaved forms were analyzed by western blot.

Real-time quantitative PCR

Freshly fAβ₄₂ (0.5 μM) or sAβ₄₂ (0.5 μM) with or without P2CSKn (0.5 μM) were incubated for 0.5 h at 37 °C and administered to the cells for 12 h. Total RNA was extracted from cell lysate with RNAiso Plus Kit (Takara). DNA removal and reverse transcription reactions were performed using HiScript II Q RT SuperMix for qPCR (Vazyme, China). Quantitative real-time PCR was performed by CFX96 (Bio-rad) using Hieff® qPCR SYBR Green Master Mix (No Rox) (Yeasen, China). Primers sequences (5’ to 3’): Gapdh: Forward: ATTCAACGGCACAGTCAAGG, Reverse: GCAGAAGGGGCGGAGATGA; Tnfα: Forward: CTGAACTTCGGGGTGATCGG, Reverse: GGCTTGTCACTCGAATTTTGAGA; iNOS: Forward: GTTCTCAGCCCAACAATACAAGA, Reverse: GTGGACGGGTCGATGTCAC; Marco: Forward: ATGGCACCAAGGGAGACAAAGG, Reverse: GCCTGGTTTTCCAGCATCACCT; Msr1: Forward: CGCACGTTCAATGACAGCATCC, Reverse: GCAAACACAAGGAGGTAGAGAGC; Mrc1: Forward: GTTCACCTGGAGTGATGGTTCTC, Reverse: AGGACATGCCAGGGTCACCTTT; Tlr2: Forward: ACAGCAAGGTCTTCCTGGTTCC, Reverse: GCTCCCTTACAGGCTGAGTTCT.

RNA sequence

BV2 cells with 70 to 80% confluence in 6-cm culture dish were treated with P2CSKn, fAβ₄₂, hPAβ₄₂ or vehicle (DMSO) for 12 h. The extracted RNA samples were sent to Beijing Tsingke Biotech Co. Ltd. for sequencing analysis. The mRNA library was assessed for quality and quantity using the Agilent 2100 bioanalyzer. The qualified library was amplified on cBot to generate the cluster on the flowcell. And the amplified flowcell was sequenced single end on the Illumina Novaseq X plus platform. The sequencing data were filtered with SOAPnuke (v1.5.2). Bowtie2 (v2.2.5) was applied to align the clean reads to the reference coding gene set, then the expression level of the gene was calculated by StringTie (v2.1.2). Heatmaps and volcano plots were plotted by https://www.bioinformatics.com.cn, an online platform for data analysis and visualization. Significance threshold was set as: Genes with absolute fold change ≥2 and nominal p < 0.05 were considered differentially expressed without multiple testing correction. Differential expression analysis was performed using the DESeq2(v1.4.5). To gain insight into the change of phenotype, GO (http://www.geneontology.org/) and KEGG enrichment analysis of annotated different expressed gene was performed by Phyper (https://en.wikipedia.org/wiki/Hypergeometric_distribution) based on Hypergeometric test. The raw sequencing data have been uploaded to the National Genomics Data Center (NGDC) Genome Sequence Archive (GSA) data platform, and the data have been reviewed and archived (Project Number: PRJCA023516; GSA Number: CRA014895).

Flow cytometry for phagocytosis determination

Generally, BV2 cells were plated in 6-well plates (100,000 cells per well), and the harvested cells were digested with cold trypsin and washed with PBS, and analyzed by flow cytometer (Beckman, CytoFlex S) immediately. The results were processed by FlowJo (V10). For time-course investigation: Cells were treated with 0.5 μM sFITC-Aβ₄₂, fFITC-Aβ₄₂, or 2.0 eq. compound (P2CSKn or Pam2CSK4) pretreated sFITC-Aβ₄₂ (0.5 μM), or 8.0 eq. compound (P2CSKn or Pam2CSK4) pretreated fFITC-Aβ₄₂ (0.5 μM), for designated time before harvest. For concentration-dependent response determination: 0.5 μM sFITC-Aβ₄₂ or fFITC-Aβ₄₂ were pretreated with gradient concentrations of P2CSKn for 30 min at 37 °C, and added to BV2 cells for 12 h incubation before harvest. For phagocytosis of Dil-LDL: sFITC-Aβ₄₂ (0.5 μM) or Dil-LDL (0.5 μM) with or without pretreated with P2CSKn (1.0 μM, 30 min at 37 °C) were incubated in BV2 cells for 1 h before harvest. The FACS gating strategy to detect FITC signals (Figs. 1c, 4b-e, 5a, f, 6g, Supplementary Fig. 9) is shown in Supplementary Fig. 17a.

Dissociating impact of morphology remodeling and microglia activation of P2CSKn on phagocytosis or apoptosis resistance

Generally, BV2 cells were plated in 6-well plates (100,000 cells per well), and different compounds were added for 6.0 h incubation before the addition of Aβ₄₂ species, and cells were harvested for further examinations. For phagocytosis of fFITC-Aβ₄₂: P2CSKn (1.0 μM, DMEM) or Pam2CSK4 (1.0 μM, DMEM) was added to the cells for 6.0 h incubation, and then washed with DMEM, and then the cells were treated with fFITC-Aβ₄₂ (0.5 μM) for 12 h incubation before harvest (Treat & wash group). Phagocytosis of fFITC-Aβ₄₂ was analyzed by a flow cytometer (Beckman, CytoFlex S) immediately. Results were processed by FlowJo (V10). For phagocytosis of sFITC-Aβ₄₂: P2CSKn (1.0 μM, DMEM) or Pam2CSK4 (1.0 μM, DMEM) was added to the cells for 6.0 h first, then washed with DMEM, and then the cells were treated with sFITC-Aβ₄₂ (0.5 μM) for 12 h before harvest (Treat & wash group). Cells incubated with sFITC-Aβ₄₂ (0.5 μM), or 1.0 μM P2CSKn or Pam2CSK4 (30 min, 37 °C)-pretreated sFITC-Aβ₄₂ (0.5 μM, Normal group) were used as reference. Phagocytosis of fFITC-Aβ₄₂ was analyzed by a flow cytometer (Beckman, CytoFlex S) immediately. Results were processed by FlowJo (V10). For apoptosis resistance: P2CSKn (1.0 μM) was added to the cells for 6 h and then washed with DMEM, then sAβ₄₂ (0.5 μM) was added for a further 12 h incubation (Treat & wash group). Cells treated with 0.5 μM fAβ₄₂ or sAβ₄₂, or preformed hPAβ₄₂ (0.5 μM sAβ₄₂ treated with 1.0 μM P2CSKn for 30 min at 37 °C), were used as references. Cells were collected, and the protein levels of PARP and its cleaved form were analyzed by western blot.

Cellular FITC-Aβ₄₂ determination by fluorescent microplate reader

For intracellular accumulation. BV2 cells were plated in a 96-well plate (7500 cells per well) and treated with 0.5 μM sFITC-Aβ₄₂ or preformed hPFAβ₄₂ for different times. Cells were then washed with PBS and stained with Hoechst 33342, and the fluorescence intensity of FITC and Hoechst 33342 were measured by microplate reader Synergy H1 (BioTek). Intracellular accumulation of FITC-Aβ₄₂ was represented as the fluorescence ratio of FITC with Hoechst 33342 according to the following formula: [FITC(Samples)-FITC(Blank)]/Hoechst 33342 (Samples) × 100%. For degradation: BV2 cells were plated in a 96-well plate (7500 cells per well) and treated with or without chloroquine (10 μM) for 1 h and then washed with PBS. 0.5 μM sFITC-Aβ₄₂, fFITC-Aβ₄₂, and preformed hPFAβ₄₂ (0.5 μM sFITC-Aβ₄₂ + 1.0 μM P2CSKn) were added to cells and incubated for 180 min. The Aβ₄₂ species were then washed off, and cells were further cultured in DMEM with 10% FBS for different times. The cells were washed with PBS and stained with Hoechst 33342. The fluorescence intensity of FITC and Hoechst 33342 was measured by the microplate reader Synergy H1 (BioTek). Intracellular accumulation of FITC-Aβ₄₂ was represented as the fluorescence ratio of [FITC(Samples)-FITC(Blank)] to Hoechst 33342, and signals at 0 min were designated as 100%. The results were processed by Graphpad (V8.0.2).

Serum stability of P2CSKn

Whole blood samples were collected from retro-orbital blood of mice (C57BL/6 J, 4-month old) used in serum separator tubes. Samples were centrifuged (1000 g, 20 min) to collect serum. P2CSKn (100 μM, saline) was added into the serum and incubated at 37 °C for gradient time points. Then, trifluoroacetic acid was added to the serum, and the mixture was immediately vortexed. The supernatant after centrifugation procedure (12000 × g, 10 min) was analyzed using Rigol L-3000 HPLC System. The stability of P2CSKn in serum is calculated using the standard concentration-response (mAU s) curve of pure P2CSKn.

In vivo investigation in local-injection APP/PS1 model

Surgery: (I) APP/PS1 mice (13 months, C57BL/6 background) were anesthetized with sodium pentobarbital (50 mg/kg) and placed on a stereotaxic apparatus (RWD, China). Skull was revealed via a midline sagittal incision. P2CSKn or reference compound (100 μM, 300 nL, saline) was mixed with Hoechst 33342 and unilaterally injected (50 nL per min) into the cerebral cortex with coordinates of AP + 1.0, ML ± 1.0, DV -1.0. The injection pipette was kept in place for 5 min and then slowly withdrawn. (II) APP/PS1 mice (9 or 16 months, C57BL/6 background) were anesthetized with sodium pentobarbital (50 mg/kg) and placed on a stereotaxic apparatus (RWD, China). P2CSKn (100 μM, 300 nL, saline) was mixed with Alexa Fluor™ 594 dye and unilaterally injected into the right dorsal hippocampus (coordinates: AP −2.2 mm, ML + 1.5 mm, DV −1.9 mm), while RP (100 μM, 300 nL) mixed with the same dye was injected into the left dorsal hippocampus (coordinates: AP −2.2 mm, ML −1.5 mm, DV −1.9 mm). The injection pipette was kept in place for 5 min to allow sufficient diffusion before slowly withdrawing. For all mice, anesthetic depth was monitored by the absence of pedal withdrawal reflex to ensure that all animals remained fully unconscious throughout the procedure. Additional dosing was administered as needed to maintain a sufficient anesthesia level during surgery. Meloxicam (5.0 mg/kg) was given subcutaneously 30 min prior to surgery, and once every 12 h after surgery for 3 consecutive days to reduce pain sensation. Histological staining: (I) for cortex, 2 or 5 days after injection, mice were deeply anesthetized with an overdose of sodium pentobarbital (100 mg/kg, i.p.) and perfused with 1x PBS, followed by 4% paraformaldehyde (PFA), through the heart. Brains were removed and fixed in 4% PFA overnight, then transferred to 20% and 30% sucrose solution until they sank to the bottom, each for 24 h to allow sufficient dehydration. After being embedded in Optical Cutting Temperature (OCT, Tissue-Cut), the brains were sectioned into coronal sections (40 μm) using a cryostat (Leica, Germany). Sections were permeabilized and blocked with 0.5% Triton-PBS containing 5% goat serum for 30 min at 37 °C; and then incubated with primary antibodies prepared in blocking solution overnight at 4 °C; and then washed with PBS, incubated with respective secondary antibodies for 2 hours and cover-slipped with Vectashield® mounting medium. (II) For hippocampus, 5- or 10-days post-injection, mice underwent a similar procedure; in addition, sections were stained with DAPI for 15 min following secondary antibody incubation and prior to mounting. Hoechst 33342 (for cortex) or Alexa Fluor™ 594 (for hippocampus) fluorescence was used to identify the injection area. Thioflavin S (TargetMol Chemicals) was used for Aβ staining. The sources and dilutions of primary antibodies were as follows: Rabbit anti-amyloid fibril OC antibody (Millipore, 1:500), anti-GFAP (Abcam, 1: 500), and anti-Iba1 (Abcam, 1:500), goat anti-rabbit IgGs, cleaved caspase-3 (Cell signaling, 1:500). The secondary antibodies were goat anti-rabbit IgG, conjugated to Alexa Fluor 647 (for 13-months group, 1:500) or Alexa Fluor 488 (for 9-months group, 1:400) from Life Technologies. Imaging and calculation: The region of interest (ROI) was determined by following the procedures. (1) Stereotaxic intracranial injections were performed using test compounds mixed with Hoechst or Alexa Fluor™ 594; (2). Brain sections were imaged to identify the injection zone (INJ; >80% maximal fluorescent intensity of relative dye) and an adjacent zone (ADJ; >200 μm from the INJ border, <10% fluorescent dye signal). (3) Analyses deliberately excluded tissues within 50 μm from the needle track to minimize injection artifacts. (4) Both ROI INJ and ROI ADJ regions were sampled for quantification of target proteins within these anatomically matched regions. Images were acquired using the Nikon AR1 confocal system with LAS AF software (Nikon Corporation). The fluorescence intensity or area of the images was quantified using Image J software (v 1.54g). The unilateral ratio of fluorescence intensity (or area) was calculated as Fluorescence Intensity (or area) of Injection Area / Fluorescence Intensity (or area) of Adjacent Area. Subsequently, fluorescence intensity ratios between the left and right hemisphere groups were compared using appropriate statistical methods. Measurement of NLRP3 level and secreted IL-1β level in injected APP/PS1 mice brain: The mice (16 months, n = 5 per group) underwent a similar procedure, but were injected with P2CSKn or RP at two different positions in the hippocampus. 5- or 10-days post-treatment, mice were anesthetized and hippocampal tissues were removed. The hippocampal tissues were divided into two parts (each containing an injection area), one part was fixed in 4% PFA overnight, then transferred to 20% and 30% sucrose solution until they sank to the bottom, each for 24 hr to allow sufficient dehydration for histological staining as above; another part was mechanically homogenized with lysis buffer containing with protease inhibitors and centrifuged (12000 × g, 20 min, 4 °C). The supernatant was collected for NLRP3 or IL-1β detection. The protein level of NLRP3 was determined by Western blot. And IL-1β concentrations in hippocampus tissues were quantified using enzyme-linked immunosorbent assay (ELISA) kits for mice according to the manufacturer’s instructions.

In vivo investigation in i.p.-injection mice model

Biodistribution of RhoB-P2CSKn: Four-month-old C57BL/6 J mice (both sexes, randomly assigned) were divided into 2 groups (n = 3 per group). Each group received i.p. injections of Rhodamine B or RhoB-P2CSKn. respectively. at a dose of 15 mg/kg. Mice were humanely euthanized by cervical dislocation under anesthesia 24 h after injection, followed by tissue harvesting of major organs. Specimens were immediately subjected to fluorescence imaging using a small animal in vivo imaging system NightOWL II B 983 (Berthold). Survival Analysis: APP/PS1 mice were divided into 2 groups (n = 9 per group, 9-month old) and C57BL/6 J mice were divided into 2 groups (n = 8 per group, 4-month old). Mice of each group received i.p injections of saline or P2CSKn (15 mg/kg) every 48 hours, with a total of two injections. Survival data were recorded and processed by Graphpad (V8.0.2). Biochemical indexes measurement and Hematoxylin and eosin (HE) staining: Each group of four-month-old C57BL/6 J mice (n = 5) received intraperitoneal injections of saline or P2CSKn (15 mg/kg) every 48 hours, with a total of two administrations. The mice were humanely euthanized, and tissues and whole blood were isolated on the fifth day of drug administration. The whole blood samples were collected from the retro-orbital blood of mice. The samples were centrifuged (1000 g, 20 min) to collect the serum. The biochemical indices of the serum were detected by ZVAST Bio. Corporation using the corresponding ELISA kits. The tissues of the mice were fixed with 4% paraformaldehyde, dehydrated with ethanol, immersed in xylene, embedded in paraffin, and cut into longitudinal sections (4.0 μm). The paraffin-embedded sections were stained with hematoxylin and eosin (HE) according to the manufacturer’s instructions (ZVAST Bio. Corporation). The images were captured by microscope (EVOS, ThermoFisher). The results were processed by Graphpad (V8.0.2).

Statistical analysis

Data are presented as mean ± standard deviation (SD) of different experiments. Statistical tests were calculated with GraphPad Prism (GraphPad Software, version 8.0.2). Significant differences were identified using two-tailed Student’s t test for pairwise comparisons or analysis of one-way ANOVA followed by Dunnett’s post hoc test for comparisons among multiple groups. A significant difference was considered when the p value was less than 0.05.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Supplementary information

Supplementary Information^{(7.9MB, pdf)}

41467_2025_63458_MOESM2_ESM.pdf^{(80.5KB, pdf)}

Description of Additional Supplementary Files

Supplementary Movie 1^{(2.3MB, avi)}

Supplementary Movie 2^{(277KB, avi)}

Supplementary Movie 3^{(729.3KB, avi)}

Reporting Summary^{(249.1KB, pdf)}

Transparent Peer Review file^{(3.3MB, pdf)}

Source data

Source Data^{(29.4MB, xlsx)}

Acknowledgements

This project was supported by the National Natural Science Foundation of China (NSFC) (No. 81973174 and 82373713 to X. Bu, No.82272743 to X. Yue, No. 22167006 to B. Zhou), Shenzhen-Hong Kong Institute of Brain Science-Shenzhen Fundamental Research Institutions (2022SHIBS0004 to Q. Zhou), Guangdong Basic and Applied Basic Research Foundation (2022A1515012221 to X. Bu). 96-well plate icon (by Marcel Tisch https://twitter.com/MarcelTisch, licensed under CC0 https://creativecommons.org/publicdomain/zero/1.0/) and microplate reader machine icon (by KeHan https://github.com/kehantan, licensed under CC0 https://creativecommons.org/publicdomain/zero/1.0/) are from Bioicons (Fig. 4a). The mouse schematic picture in the Fig. 7a and Supplementary Fig. 13 (originated from Ethan Tyler, Lex Kravitz, 10.5281/zenodo.3925901) and mouse head schema in Fig. 7b (originated from Luigi Petrucco, 10.5281/zenodo.3925903) from SciDraw.

Author contributions

X. Bu and Q. Zhou designed the study; Y. Wang, Z. Wang, Z. Liu, J. Li, S. Yang, Y. Zhao, Y. Huang, C. Liao, Y. Zhang, J. Zhao, W. Zhou, performed the experiments; Y. Wang, Z. Wang, J. Li, Z. Liu, Y. Huang, B. Zhou and X. Yue performed data calculation and analysis. X. Bu, Q. Zhou, and X. Yue supervised and interpreted the data, and wrote the manuscript.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.

Data availability

Unless otherwise stated, all data supporting the results of this study can be found in the article, supplementary, and source data files. The RNA-Seq data generated in this study have been deposited in the Genome Sequence Archive (GSA) database under accession code CRA014895. The mass spectrometry data generated in this study is provided in the Supplementary Information. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

These authors contributed equally: Youqiao Wang, Zeyi Wang, Ziyi Liu.

Change history

11/18/2025

A Correction to this paper has been published: 10.1038/s41467-025-66496-z

Contributor Information

Xin Yue, Email: yuex2504@jnu.edu.cn.

Qiang Zhou, Email: zhouqiang@pkusz.edu.cn.

Xianzhang Bu, Email: phsbxzh@mail.sysu.edu.cn.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-025-63458-3.

References

1.Iadanza, M. G., Jackson, M. P., Hewitt, E. W., Ranson, N. A. & Radford, S. E. A new era for understanding amyloid structures and disease. Nat. Rev. Mol. Cell Biol.19, 755–773 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Ke, P. C. et al. Half a century of amyloids: past, present and future. Chem. Soc. Rev.49, 5473–5509 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Chuang, E., Hori, A. M., Hesketh, C. D. & Shorter, J. Amyloid assembly and disassembly. J. Cell Sci.131, jcs189928 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Chiti, F. & Dobson, C. M. Protein misfolding, functional amyloid, and human disease. Annu. Rev. Biochem.75, 333–366 (2006). [DOI] [PubMed] [Google Scholar]
5.Lashuel, H. A., Overk, C. R., Oueslati, A. & Masliah, E. The many faces of alpha-synuclein: from structure and toxicity to therapeutic target. Nat. Rev. Neurosci.14, 38–48 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Villar-Pique, A., Lopes da Fonseca, T. & Outeiro, T. F. Structure, function and toxicity of alpha-synuclein: the Bermuda triangle in synucleinopathies. J. Neurochem.139, 240–255 (2016). [DOI] [PubMed] [Google Scholar]
7.Kleffman, K. et al. Melanoma-secreted amyloid beta suppresses neuroinflammation and promotes brain metastasis. Cancer Discov.12, 1314–1335 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Palanikumar, L. et al. Protein mimetic amyloid inhibitor potently abrogates cancer-associated mutant p53 aggregation and restores tumor suppressor function. Nat. Commun.12, 3962 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Yoo, J., Lee, J., Ahn, B., Han, J. & Lim, M. H. Multi-target-directed therapeutic strategies for Alzheimer’s disease: controlling amyloid-β aggregation, metal ion homeostasis, and enzyme inhibition. Chem. Sci.16, 2105–2135 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Panza, F., Lozupone, M., Logroscino, G. & Imbimbo, B. P. A critical appraisal of amyloid-beta-targeting therapies for Alzheimer disease. Nat. Rev. Neurol.15, 73–88 (2019). [DOI] [PubMed] [Google Scholar]
11.Tu, S., Okamoto, S., Lipton, S. A. & Xu, H. Oligomeric Aβ-induced synaptic dysfunction in Alzheimer’s disease. Mol. Neurodegener.9, 48 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Francioso, A. et al. β-sheet interfering molecules acting against β-amyloid aggregation and fibrillogenesis. Bioorg. Med. Chem.23, 1671–1683 (2015). [DOI] [PubMed] [Google Scholar]
13.Wang, X. A bridge between the innate immunity system and amyloid-β production in Alzheimer’s disease. Neurosci. Bull.37, 898–901 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Sarlus, H. & Heneka, M. T. Microglia in Alzheimer’s disease. J. Clin. Invest.127, 3240–3249 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Ries, M. & Sastre, M. Mechanisms of Aβ clearance and degradation by glial cells. Front. Aging Neurosci.8, 160 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Wilcock, D. M. et al. Microglial activation facilitates Aβ plaque removal following intracranial anti-Aβ antibody administration. Neurobiol. Dis.15, 11–20 (2004). [DOI] [PubMed] [Google Scholar]
17.Fuller, J. P., Stavenhagen, J. B. & Teeling, J. L. New roles for Fc receptors in neurodegeneration-the impact on Immunotherapy for Alzheimer’s Disease. Front. Neurosci.8, 235 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Heidebrink, J. L. & Paulson, H. L. Lessons learned from approval of aducanumab for Alzheimer’s disease. Annu. Rev. Med.75, 99–111 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Hossain, M. F., Husna, A. U. & Kharel, M. Use of lecanemab for the treatment of Alzheimer’s disease: a systematic review. Brain Behav.14, e3592 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Høilund-Carlsen, P. F. et al. Donanemab, another anti-Alzheimer’s drug with risk and uncertain benefit. Ageing Res. Rev.99, 102348 (2024). [DOI] [PubMed] [Google Scholar]
21.Zhang, Y., Chen, H., Li, R., Sterling, K. & Song, W. Amyloid beta-based therapy for Alzheimer’s disease: challenges, successes and future. Signal Transduct. Target. Ther.8, 248 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Wang, Y. et al. Lessons from anti-amyloid-beta immunotherapies in Alzheimer disease: aiming at a moving target. Neurodegener. Dis.17, 242–250 (2017). [DOI] [PubMed] [Google Scholar]
23.Plascencia-Villa, G. & Perry, G. In Handbook of Clinical Neurology (ed Espay, A. J.) (Elsevier, 2023).
24.Haass, C. & Selkoe, D. J. Soluble protein oligomers in neurodegeneration: lessons from the Alzheimer’s amyloid β-peptide. Nat. Rev. Mol. Cell Biol.8, 101–112 (2007). [DOI] [PubMed] [Google Scholar]
25.Karran, E. & De Strooper, B. The amyloid hypothesis in Alzheimer disease: new insights from new therapeutics. Nat. Rev. Drug. Discov.21, 306–318 (2022). [DOI] [PubMed] [Google Scholar]
26.Yamada, K. et al. Abeta immunotherapy: intracerebral sequestration of Abeta by an anti-Abeta monoclonal antibody 266 with high affinity to soluble Abeta. J. Neurosci.29, 11393–11398 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Demattos, R. B. et al. A plaque-specific antibody clears existing beta-amyloid plaques in Alzheimer’s disease mice. Neuron76, 908–920 (2012). [DOI] [PubMed] [Google Scholar]
28.Gardberg, A. S. et al. Molecular basis for passive immunotherapy of Alzheimer’s disease. Proc. Natl Acad. Sci. USA104, 15659–15664 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Liu, Y. H., Giunta, B., Zhou, H. D., Tan, J. & Wang, Y. J. Immunotherapy for Alzheimer disease: the challenge of adverse effects. Nat. Rev. Neurol.8, 465–469 (2012). [DOI] [PubMed] [Google Scholar]
30.Liu, Y. H. et al. An N-terminal antibody promotes the transformation of amyloid fibrils into oligomers and enhances the neurotoxicity of amyloid-beta: the dust-raising effect. J. Neuroinflammation12, 153 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Lian, Y. et al. Clarity on the blazing trail: clearing the way for amyloid-removing therapies for Alzheimer’s disease. Mol. Psychiatry29, 297–305 (2023). [DOI] [PubMed] [Google Scholar]
32.Liu, Y. H. et al. Clearance of amyloid-beta in Alzheimer’s disease: shifting the action site from center to periphery. Mol. Neurobiol.51, 1–7 (2015). [DOI] [PubMed] [Google Scholar]
33.Cai, W., Wu, T. & Chen, N. The amyloid-beta clearance: from molecular targets to glial and neural cells. Biomolecules13, 313 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Wilcock, D. M. et al. Diverse inflammatory responses in transgenic mouse models of Alzheimer’s disease and the effect of immunotherapy on these responses. ASN Neuro3, 249–258 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Pan, X. D. et al. Microglial phagocytosis induced by fibrillar β-amyloid is attenuated by oligomeric β-amyloid: implications for Alzheimer’s disease. Mol. Neurodegener.6, 45 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Fakhoury, M. Microglia and astrocytes in alzheimer’s disease: implications for therapy. Curr. Neuropharmacol.16, 508–518 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Franco-Bocanegra, D. K., McAuley, C., Nicoll, J. A. R. & Boche, D. Molecular mechanisms of microglial motility: changes in ageing and Alzheimer’s disease. Cells8, 639 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Loeffler, D. A. Antibody-mediated clearance of brain amyloid-β: mechanisms of action, effects of natural and monoclonal anti-aβ antibodies, and downstream effects. J. Alzheimers Dis. Rep.7, 873–899 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Wen, G. et al. Stabilizing the monomeric amyloid-beta peptide by tyrocidine A prevents and reverses amyloidogenesis without the accumulation of oligomers. Chem. Commun.53, 3886–3889 (2017). [Google Scholar]
40.Wen, G. et al. Stabilizing amyloid-beta peptide by the N-terminus capture is capable of preventing and eliminating amyloid-beta oligomers. Chem. Commun.53, 7673–7676 (2017). [Google Scholar]
41.Chen, D. et al. Dissociation of haemolytic and oligomer-preventing activities of gramicidin S derivatives targeting the amyloid-beta N-terminus. Chem. Commun.53, 13340–13343 (2017). [Google Scholar]
42.Ahmed, M. et al. Structural conversion of neurotoxic amyloid-β 1-42 oligomers to fibrils. Nat. Struct. Mol. Biol.17, 561–567 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Luhrs, T. et al. 3D structure of Alzheimer’s amyloid- (1-42) fibrils. Proc. Natl Acad. Sci. USA102, 17342–17347 (2005). [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Gremer, L. et al. Fibril structure of amyloid-β(1–42) by cryo–electron microscopy. Science358, 116–119 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Chafekar, S. M. et al. Branched KLVFF tetramers strongly potentiate inhibition of beta-amyloid aggregation. Chembiochem8, 1857–1864 (2007). [DOI] [PubMed] [Google Scholar]
46.Fleming, S. & Ulijn, R. V. Design of nanostructures based on aromatic peptide amphiphiles. Chem. Soc. Rev.43, 8150–8177 (2014). [DOI] [PubMed] [Google Scholar]
47.Landreth, G. E. & Reed-Geaghan, E. G. Toll-like receptors in Alzheimer’s disease. Curr. Top. Microbiol. Immunol.336, 137–153 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Ifuku, M. et al. Activation of Toll-like receptor 5 in microglia modulates their function and triggers neuronal injury. Acta Neuropathol. Commun.8, 159 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Kayed, R. et al. Common structure of soluble amyloid oligomers implies common mechanism of pathogenesis. Science300, 486–489 (2003). [DOI] [PubMed] [Google Scholar]
50.Reinke, A. A. & Gestwicki, J. E. Insight into amyloid structure using chemical probes. Chem. Biol. Drug Des.77, 399–411 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Fancy, D. A. & Kodadek, T. Chemistry for the analysis of protein-protein interactions: rapid and efficient cross-linking triggered by long wavelength light. Proc. Natl Acad. Sci. USA96, 6020–6024 (1999). [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Kayed, R. et al. Fibril specific, conformation dependent antibodies recognize a generic epitope common to amyloid fibrils and fibrillar oligomers that is absent in prefibrillar oligomers. Mol. Neurodegener.2, 18 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]
53.Raha, S., Ghosh, A., Dutta, D., Patel, D. R. & Pahan, K. Activation of PPARalpha enhances astroglial uptake and degradation of beta-amyloid. Sci. Signal.14, eabg4747 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Chen, X., Zhang, Y., Sadadcharam, G., Cui, W. & Wang, J. H. Isolation, Purification, and Culture of Primary Murine Microglia Cells. Bio Protoc.3, e314 (2013). [Google Scholar]
55.Klementieva, O. et al. Pre-plaque conformational changes in Alzheimer’s disease-linked Aβ and APP. Nat. Commun.8, 14726 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Meyer-Luehmann, M. et al. Rapid appearance and local toxicity of amyloid-beta plaques in a mouse model of Alzheimer’s disease. Nature451, 720–724 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Prinz, M., Jung, S. & Priller, J. Microglia Biology: One Century of Evolving Concepts. Cell179, 292–311 (2019). [DOI] [PubMed] [Google Scholar]
58.Baik, S. H., Kang, S., Son, S. M. & Mook-Jung, I. Microglia contributes to plaque growth by cell death due to uptake of amyloid beta in the brain of Alzheimer’s disease mouse model. Glia64, 2274–2290 (2016). [DOI] [PubMed] [Google Scholar]
59.Mildner, A. et al. Distinct and non-redundant roles of microglia and myeloid subsets in mouse models of Alzheimer’s disease. J. Neurosci.31, 11159–11171 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Ni, J. et al. An impaired intrinsic microglial clock system induces neuroinflammatory alterations in the early stage of amyloid precursor protein knock-in mouse brain. J. Neuroinflammation16, 173 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Floden, A. M. & Combs, C. K. Microglia demonstrate age-dependent interaction with amyloid-beta fibrils. J. Alzheimers Dis.25, 279–293 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
62.Njie, E. G. et al. Ex vivo cultures of microglia from young and aged rodent brain reveal age-related changes in microglial function. Neurobiol. Aging33, e1–e12 (2012). [DOI] [PubMed] [Google Scholar]
63.Teoh, C. L. et al. Chemical fluorescent probe for detection of abeta oligomers. J. Am. Chem. Soc.137, 13503–13509 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information^{(7.9MB, pdf)}

41467_2025_63458_MOESM2_ESM.pdf^{(80.5KB, pdf)}

Description of Additional Supplementary Files

Supplementary Movie 1^{(2.3MB, avi)}

Supplementary Movie 2^{(277KB, avi)}

Supplementary Movie 3^{(729.3KB, avi)}

Reporting Summary^{(249.1KB, pdf)}

Transparent Peer Review file^{(3.3MB, pdf)}

Source Data^{(29.4MB, xlsx)}

Data Availability Statement

[CR1] 1.Iadanza, M. G., Jackson, M. P., Hewitt, E. W., Ranson, N. A. & Radford, S. E. A new era for understanding amyloid structures and disease. Nat. Rev. Mol. Cell Biol.19, 755–773 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR2] 2.Ke, P. C. et al. Half a century of amyloids: past, present and future. Chem. Soc. Rev.49, 5473–5509 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR3] 3.Chuang, E., Hori, A. M., Hesketh, C. D. & Shorter, J. Amyloid assembly and disassembly. J. Cell Sci.131, jcs189928 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR4] 4.Chiti, F. & Dobson, C. M. Protein misfolding, functional amyloid, and human disease. Annu. Rev. Biochem.75, 333–366 (2006). [DOI] [PubMed] [Google Scholar]

[CR5] 5.Lashuel, H. A., Overk, C. R., Oueslati, A. & Masliah, E. The many faces of alpha-synuclein: from structure and toxicity to therapeutic target. Nat. Rev. Neurosci.14, 38–48 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR6] 6.Villar-Pique, A., Lopes da Fonseca, T. & Outeiro, T. F. Structure, function and toxicity of alpha-synuclein: the Bermuda triangle in synucleinopathies. J. Neurochem.139, 240–255 (2016). [DOI] [PubMed] [Google Scholar]

[CR7] 7.Kleffman, K. et al. Melanoma-secreted amyloid beta suppresses neuroinflammation and promotes brain metastasis. Cancer Discov.12, 1314–1335 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR8] 8.Palanikumar, L. et al. Protein mimetic amyloid inhibitor potently abrogates cancer-associated mutant p53 aggregation and restores tumor suppressor function. Nat. Commun.12, 3962 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR9] 9.Yoo, J., Lee, J., Ahn, B., Han, J. & Lim, M. H. Multi-target-directed therapeutic strategies for Alzheimer’s disease: controlling amyloid-β aggregation, metal ion homeostasis, and enzyme inhibition. Chem. Sci.16, 2105–2135 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR10] 10.Panza, F., Lozupone, M., Logroscino, G. & Imbimbo, B. P. A critical appraisal of amyloid-beta-targeting therapies for Alzheimer disease. Nat. Rev. Neurol.15, 73–88 (2019). [DOI] [PubMed] [Google Scholar]

[CR11] 11.Tu, S., Okamoto, S., Lipton, S. A. & Xu, H. Oligomeric Aβ-induced synaptic dysfunction in Alzheimer’s disease. Mol. Neurodegener.9, 48 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR12] 12.Francioso, A. et al. β-sheet interfering molecules acting against β-amyloid aggregation and fibrillogenesis. Bioorg. Med. Chem.23, 1671–1683 (2015). [DOI] [PubMed] [Google Scholar]

[CR13] 13.Wang, X. A bridge between the innate immunity system and amyloid-β production in Alzheimer’s disease. Neurosci. Bull.37, 898–901 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR14] 14.Sarlus, H. & Heneka, M. T. Microglia in Alzheimer’s disease. J. Clin. Invest.127, 3240–3249 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR15] 15.Ries, M. & Sastre, M. Mechanisms of Aβ clearance and degradation by glial cells. Front. Aging Neurosci.8, 160 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR16] 16.Wilcock, D. M. et al. Microglial activation facilitates Aβ plaque removal following intracranial anti-Aβ antibody administration. Neurobiol. Dis.15, 11–20 (2004). [DOI] [PubMed] [Google Scholar]

[CR17] 17.Fuller, J. P., Stavenhagen, J. B. & Teeling, J. L. New roles for Fc receptors in neurodegeneration-the impact on Immunotherapy for Alzheimer’s Disease. Front. Neurosci.8, 235 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR18] 18.Heidebrink, J. L. & Paulson, H. L. Lessons learned from approval of aducanumab for Alzheimer’s disease. Annu. Rev. Med.75, 99–111 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR19] 19.Hossain, M. F., Husna, A. U. & Kharel, M. Use of lecanemab for the treatment of Alzheimer’s disease: a systematic review. Brain Behav.14, e3592 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR20] 20.Høilund-Carlsen, P. F. et al. Donanemab, another anti-Alzheimer’s drug with risk and uncertain benefit. Ageing Res. Rev.99, 102348 (2024). [DOI] [PubMed] [Google Scholar]

[CR21] 21.Zhang, Y., Chen, H., Li, R., Sterling, K. & Song, W. Amyloid beta-based therapy for Alzheimer’s disease: challenges, successes and future. Signal Transduct. Target. Ther.8, 248 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR22] 22.Wang, Y. et al. Lessons from anti-amyloid-beta immunotherapies in Alzheimer disease: aiming at a moving target. Neurodegener. Dis.17, 242–250 (2017). [DOI] [PubMed] [Google Scholar]

[CR23] 23.Plascencia-Villa, G. & Perry, G. In Handbook of Clinical Neurology (ed Espay, A. J.) (Elsevier, 2023).

[CR24] 24.Haass, C. & Selkoe, D. J. Soluble protein oligomers in neurodegeneration: lessons from the Alzheimer’s amyloid β-peptide. Nat. Rev. Mol. Cell Biol.8, 101–112 (2007). [DOI] [PubMed] [Google Scholar]

[CR25] 25.Karran, E. & De Strooper, B. The amyloid hypothesis in Alzheimer disease: new insights from new therapeutics. Nat. Rev. Drug. Discov.21, 306–318 (2022). [DOI] [PubMed] [Google Scholar]

[CR26] 26.Yamada, K. et al. Abeta immunotherapy: intracerebral sequestration of Abeta by an anti-Abeta monoclonal antibody 266 with high affinity to soluble Abeta. J. Neurosci.29, 11393–11398 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR27] 27.Demattos, R. B. et al. A plaque-specific antibody clears existing beta-amyloid plaques in Alzheimer’s disease mice. Neuron76, 908–920 (2012). [DOI] [PubMed] [Google Scholar]

[CR28] 28.Gardberg, A. S. et al. Molecular basis for passive immunotherapy of Alzheimer’s disease. Proc. Natl Acad. Sci. USA104, 15659–15664 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR29] 29.Liu, Y. H., Giunta, B., Zhou, H. D., Tan, J. & Wang, Y. J. Immunotherapy for Alzheimer disease: the challenge of adverse effects. Nat. Rev. Neurol.8, 465–469 (2012). [DOI] [PubMed] [Google Scholar]

[CR30] 30.Liu, Y. H. et al. An N-terminal antibody promotes the transformation of amyloid fibrils into oligomers and enhances the neurotoxicity of amyloid-beta: the dust-raising effect. J. Neuroinflammation12, 153 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR31] 31.Lian, Y. et al. Clarity on the blazing trail: clearing the way for amyloid-removing therapies for Alzheimer’s disease. Mol. Psychiatry29, 297–305 (2023). [DOI] [PubMed] [Google Scholar]

[CR32] 32.Liu, Y. H. et al. Clearance of amyloid-beta in Alzheimer’s disease: shifting the action site from center to periphery. Mol. Neurobiol.51, 1–7 (2015). [DOI] [PubMed] [Google Scholar]

[CR33] 33.Cai, W., Wu, T. & Chen, N. The amyloid-beta clearance: from molecular targets to glial and neural cells. Biomolecules13, 313 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR34] 34.Wilcock, D. M. et al. Diverse inflammatory responses in transgenic mouse models of Alzheimer’s disease and the effect of immunotherapy on these responses. ASN Neuro3, 249–258 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR35] 35.Pan, X. D. et al. Microglial phagocytosis induced by fibrillar β-amyloid is attenuated by oligomeric β-amyloid: implications for Alzheimer’s disease. Mol. Neurodegener.6, 45 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR36] 36.Fakhoury, M. Microglia and astrocytes in alzheimer’s disease: implications for therapy. Curr. Neuropharmacol.16, 508–518 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR37] 37.Franco-Bocanegra, D. K., McAuley, C., Nicoll, J. A. R. & Boche, D. Molecular mechanisms of microglial motility: changes in ageing and Alzheimer’s disease. Cells8, 639 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR38] 38.Loeffler, D. A. Antibody-mediated clearance of brain amyloid-β: mechanisms of action, effects of natural and monoclonal anti-aβ antibodies, and downstream effects. J. Alzheimers Dis. Rep.7, 873–899 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR39] 39.Wen, G. et al. Stabilizing the monomeric amyloid-beta peptide by tyrocidine A prevents and reverses amyloidogenesis without the accumulation of oligomers. Chem. Commun.53, 3886–3889 (2017). [Google Scholar]

[CR40] 40.Wen, G. et al. Stabilizing amyloid-beta peptide by the N-terminus capture is capable of preventing and eliminating amyloid-beta oligomers. Chem. Commun.53, 7673–7676 (2017). [Google Scholar]

[CR41] 41.Chen, D. et al. Dissociation of haemolytic and oligomer-preventing activities of gramicidin S derivatives targeting the amyloid-beta N-terminus. Chem. Commun.53, 13340–13343 (2017). [Google Scholar]

[CR42] 42.Ahmed, M. et al. Structural conversion of neurotoxic amyloid-β 1-42 oligomers to fibrils. Nat. Struct. Mol. Biol.17, 561–567 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR43] 43.Luhrs, T. et al. 3D structure of Alzheimer’s amyloid- (1-42) fibrils. Proc. Natl Acad. Sci. USA102, 17342–17347 (2005). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR44] 44.Gremer, L. et al. Fibril structure of amyloid-β(1–42) by cryo–electron microscopy. Science358, 116–119 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR45] 45.Chafekar, S. M. et al. Branched KLVFF tetramers strongly potentiate inhibition of beta-amyloid aggregation. Chembiochem8, 1857–1864 (2007). [DOI] [PubMed] [Google Scholar]

[CR46] 46.Fleming, S. & Ulijn, R. V. Design of nanostructures based on aromatic peptide amphiphiles. Chem. Soc. Rev.43, 8150–8177 (2014). [DOI] [PubMed] [Google Scholar]

[CR47] 47.Landreth, G. E. & Reed-Geaghan, E. G. Toll-like receptors in Alzheimer’s disease. Curr. Top. Microbiol. Immunol.336, 137–153 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR48] 48.Ifuku, M. et al. Activation of Toll-like receptor 5 in microglia modulates their function and triggers neuronal injury. Acta Neuropathol. Commun.8, 159 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR49] 49.Kayed, R. et al. Common structure of soluble amyloid oligomers implies common mechanism of pathogenesis. Science300, 486–489 (2003). [DOI] [PubMed] [Google Scholar]

[CR50] 50.Reinke, A. A. & Gestwicki, J. E. Insight into amyloid structure using chemical probes. Chem. Biol. Drug Des.77, 399–411 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR51] 51.Fancy, D. A. & Kodadek, T. Chemistry for the analysis of protein-protein interactions: rapid and efficient cross-linking triggered by long wavelength light. Proc. Natl Acad. Sci. USA96, 6020–6024 (1999). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR52] 52.Kayed, R. et al. Fibril specific, conformation dependent antibodies recognize a generic epitope common to amyloid fibrils and fibrillar oligomers that is absent in prefibrillar oligomers. Mol. Neurodegener.2, 18 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR53] 53.Raha, S., Ghosh, A., Dutta, D., Patel, D. R. & Pahan, K. Activation of PPARalpha enhances astroglial uptake and degradation of beta-amyloid. Sci. Signal.14, eabg4747 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR54] 54.Chen, X., Zhang, Y., Sadadcharam, G., Cui, W. & Wang, J. H. Isolation, Purification, and Culture of Primary Murine Microglia Cells. Bio Protoc.3, e314 (2013). [Google Scholar]

[CR55] 55.Klementieva, O. et al. Pre-plaque conformational changes in Alzheimer’s disease-linked Aβ and APP. Nat. Commun.8, 14726 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR56] 56.Meyer-Luehmann, M. et al. Rapid appearance and local toxicity of amyloid-beta plaques in a mouse model of Alzheimer’s disease. Nature451, 720–724 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR57] 57.Prinz, M., Jung, S. & Priller, J. Microglia Biology: One Century of Evolving Concepts. Cell179, 292–311 (2019). [DOI] [PubMed] [Google Scholar]

[CR58] 58.Baik, S. H., Kang, S., Son, S. M. & Mook-Jung, I. Microglia contributes to plaque growth by cell death due to uptake of amyloid beta in the brain of Alzheimer’s disease mouse model. Glia64, 2274–2290 (2016). [DOI] [PubMed] [Google Scholar]

[CR59] 59.Mildner, A. et al. Distinct and non-redundant roles of microglia and myeloid subsets in mouse models of Alzheimer’s disease. J. Neurosci.31, 11159–11171 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR60] 60.Ni, J. et al. An impaired intrinsic microglial clock system induces neuroinflammatory alterations in the early stage of amyloid precursor protein knock-in mouse brain. J. Neuroinflammation16, 173 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR61] 61.Floden, A. M. & Combs, C. K. Microglia demonstrate age-dependent interaction with amyloid-beta fibrils. J. Alzheimers Dis.25, 279–293 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR62] 62.Njie, E. G. et al. Ex vivo cultures of microglia from young and aged rodent brain reveal age-related changes in microglial function. Neurobiol. Aging33, e1–e12 (2012). [DOI] [PubMed] [Google Scholar]

[CR63] 63.Teoh, C. L. et al. Chemical fluorescent probe for detection of abeta oligomers. J. Am. Chem. Soc.137, 13503–13509 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Targeting phagocytosis for amyloid-β clearance: implications of morphology remodeling and microglia activation probed by bifunctional chimaeras

Youqiao Wang

Zeyi Wang

Ziyi Liu

Jinyan Li

Shuo Yang

Yuxuan Zhao

Yangmei Huang

Chenyang Liao

Yiqiu Zhang

Jiaojiao Zhao

Weilin Zhou

Binhua Zhou

Xin Yue

Qiang Zhou

Xianzhang Bu

Abstract

Introduction

Results

Conctruction of structure-remodeling and microglia-activating chimaeras

Fig. 1. Design of multifunctional chimaera.

P2CSKn efficiently remodels different Aβ42 species to monodisperse hybrid co-aggregates

P2CSKn fast co-assembles with sAβ42 to form conformation-modified hybrid aggregates

Fig. 2. P2CSKn co-assembles with sAβ42 to form conformation-modified hybrid aggregates.

P2CSKn re-assembles fAβ42 to conformation and morphology-modified, and monodisperse hybrid co-aggregates

Fig. 3. P2CSKn re-assembles fAβ42 to conformation- and morphology-modified hybrid aggregates.

P2CSKn efficiently promotes a durable ingestion of Aβ42 species in different cell model

Fig. 4. P2CSKn promotes ingestion of Aβ42 species in different cell models.

Roles of P2CSKn in the phagocytosis of Aβ42 species by BV2 cells

BV2 cells can uptake different Aβ42 species in TLR2-dependent manner

Fig. 5. The effects of BV2 cell activation on P2CSKn-facilitated phagocytosis.

P2CSKn activates BV2 cells and upregulates scavenger receptor MARCO and MSR1

TLR2 activation by P2CSKn moderately enhances phagocytosis of Aβ42 species

Morphology remodeling of Aβ42 species by P2CSKn efficiently enhances Aβ-TLR2 interaction

Fig. 6. The effects of morphology remodeling of Aβ42 species by P2CSKn on phagocytosis.

Morphology remodeling of Aβ42 species by P2CSKn efficiently attenuates Aβ42 species-induced apoptosis in BV2 cells

Morphology remodeling by P2CSKn ensures the high efficacy of phagocytosis of Aβ42 species in BV2 cells

Local injection of P2CSKn reduces Aβ burden in brains of APP/PS1 transgenic mice

Fig. 7. P2CSKn enhances Aβ phagocytosis in brains of APP/PS1 transgenic mice.

Discussion

Methods

Ethical statement

General

Preparation of Aβ42/FITC-Aβ42 samples

Cell culture

Tlr2 siRNA transfection in BV2 cells

Mouse primary microglia cells

Tyndall effect

Dynamic light scattering

Thioflavin fluorescence kinetics

Circular dichroism

Microscale thermophoresis

Dot blot of Aβ42 oligomer

Laser scanning confocal microscopy

Transmission electron microscopy

Western blot

Photo-induced cross-linking of unmodified proteins

Cellular thermal shift assay

Immunofluorescence of cellular Iba1

Cell viability and apoptosis assay

Real-time quantitative PCR

RNA sequence

Flow cytometry for phagocytosis determination

Dissociating impact of morphology remodeling and microglia activation of P2CSKn on phagocytosis or apoptosis resistance

Cellular FITC-Aβ42 determination by fluorescent microplate reader

Serum stability of P2CSKn

In vivo investigation in local-injection APP/PS1 model

In vivo investigation in i.p.-injection mice model

Statistical analysis

Reporting summary

Supplementary information

Source data

Acknowledgements

Author contributions

Peer review

Peer review information

Data availability

Competing interests

Footnotes

P2CSKn efficiently remodels different Aβ₄₂ species to monodisperse hybrid co-aggregates

P2CSKn fast co-assembles with sAβ₄₂ to form conformation-modified hybrid aggregates

Fig. 2. P2CSKn co-assembles with sAβ₄₂ to form conformation-modified hybrid aggregates.

P2CSKn re-assembles fAβ₄₂ to conformation and morphology-modified, and monodisperse hybrid co-aggregates

Fig. 3. P2CSKn re-assembles fAβ₄₂ to conformation- and morphology-modified hybrid aggregates.

P2CSKn efficiently promotes a durable ingestion of Aβ₄₂ species in different cell model

Fig. 4. P2CSKn promotes ingestion of Aβ₄₂ species in different cell models.

Roles of P2CSKn in the phagocytosis of Aβ₄₂ species by BV2 cells

BV2 cells can uptake different Aβ₄₂ species in TLR2-dependent manner

TLR2 activation by P2CSKn moderately enhances phagocytosis of Aβ₄₂ species

Morphology remodeling of Aβ₄₂ species by P2CSKn efficiently enhances Aβ-TLR2 interaction

Fig. 6. The effects of morphology remodeling of Aβ₄₂ species by P2CSKn on phagocytosis.

Morphology remodeling of Aβ₄₂ species by P2CSKn efficiently attenuates Aβ₄₂ species-induced apoptosis in BV2 cells

Morphology remodeling by P2CSKn ensures the high efficacy of phagocytosis of Aβ₄₂ species in BV2 cells

Preparation of Aβ₄₂/FITC-Aβ₄₂ samples

Dot blot of Aβ₄₂ oligomer

Cellular FITC-Aβ₄₂ determination by fluorescent microplate reader