Active immunization with bacteriophage AP205 VLPs results in reduced amyloid load and microgliosis in 5xFAD female mice

Abstract

Alzheimer’s disease (AD) is the leading cause of dementia worldwide and the accumulation of amyloid β (Aβ) oligomers in the brain parenchyma is one of the main characteristics of AD. Pyroglutamate-modified amyloid β (pE₃Aβ) forms are highly pathogenic components of Aβ plaques and are viable targets for disease-modifying strategies. Active immunization using virus-like particles (VLPs) represents a promising therapeutic approach to combating AD. Using RNA bacteriophage-based VLPs, we developed and tested a vaccine targeting pE₃Aβ in 5xFAD female mice. This study utilized the bacteriophage AP205 VLP platform to generate candidate compounds for active immunization. AP205 VLPs were modified to display short pE₃Aβ peptides. At 2 months of age, 5xFAD female mice received four immunizations with either pE₃Aβ VLPs or AP205 VLPs. Blood titers were assessed biweekly for the first 45 days and then every 2 months using ELISA. Behavioral tests, including open field, spontaneous alternation, Morris water maze, and elevated zero maze (EZM), were performed at 6 and 8 months of age. Immunohistochemical analyses evaluated levels of Aβ42, pE₃Aβ, glial fibrillary acidic protein (GFAP), and ionized calcium-binding adapter molecule-1 (Iba-1). pE₃Aβ VLPs and AP205 VLPs did not alter cognitive or locomotor performance in 5xFAD mice. The working memory of 8-month-old pE₃Aβ VLP-treated mice was better than it was at 6-months of age. A moderate reduction in Aβ42 pathology and microglial activation was observed in both vaccinated groups. VLP-based vaccine administration showed no behavioral improvements in 5xFAD mice but demonstrated modest effects on Aβ42 load and microgliosis.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00210-025-04739-y.

Introduction

Approximately 60 million people worldwide are affected by dementia, with Alzheimer’s disease (AD) being the most prevalent cause (Nichols et al. 2022). AD is characterized by the extracellular deposition of beta-amyloid (Aβ) protein as plaques and the formation of neurofibrillary tangles composed of phosphorylated tau protein (Serrano-Pozo et al. 2011). In less than 1% of AD cases, genetic mutations in the Presenilin (PSEN) 1, PSEN 2, and amyloid-β protein precursor genes are implicated (DeTure and Dickson 2019).

The production of Aβ is a critical step in AD pathogenesis, triggering a cascade of reactions that lead to the formation of neurofibrillary tangles, neuroinflammation, neuronal damage, and eventual neuronal death and white matter loss in the brain (Chadha et al. 2021).

Over recent decades, both active and passive immunization strategies have been extensively studied (Usman et al. 2021). Early treatment is essential, as Aβ accumulation and neuronal loss can occur years before cognitive symptoms manifest (Holmes et al. 2008). Monoclonal antibodies, such as lecanemab (Dyck et al. 2023) and donanemab (Sims et al. 2023), have shown promise in slowing cognitive decline by targeting specific forms of Aβ. However, these therapies are costly, require frequent administration, and are associated with significant side effects such as vasogenic edema (Song et al. 2022).

Vaccination, which stimulates endogenous antibody production, offers a more economically feasible and less invasive alternative. Virus-like particles (VLPs)—hollow, non-infectious protein structures derived from viruses (Roldão et al. 2010; Pumpens et al. 2016; Jennings and Bachmann 2008)—have been extensively used as carriers for antigens, including Aβ and tau epitopes, see Lim et. al for a recent review (Lim et al. 2024). For example, CAD106, a second-generation VLP-based Aβ active immunotherapy, halts Aβ deposition in humans by using bacteriophage Qβ VLPs to present Aβ1–6 peptides on the surface (Winblad et al. 2012). Even more recently norovirus VLPs were used to produce bi-valent Aβ and tau targeting vaccines, able to significantly eradicate Aβ plaques and tau tangles and even enhance the cognitive abilities of 3xTg mice (Feng et al. 2024). Therefore, the decoration of VLPs with multiple copies of Aβ and/or tau epitopes proved to be advantageous in modulating the intensity of the immune response elicited by the vaccine.

Building on the success of donanemab, which targets the pyroglutaminated N-terminus of Aβ (pE₃Aβ), there is interest in developing vaccines that generate autoantibodies against pE₃Aβ. pE₃Aβ is extremely pathogenic and is the major constituent of Aβ plaques in the AD brain (Wirths et al. 2010). Recent studies have shown that a pE₃Aβ-targeted Multi-TEP platform vaccine can significantly reduce Aβ plaque area (Zagorski et al. 2023).

In this study, we used bacteriophage AP205 VLPs genetically fused with N-pyroglutamated Aβ_3–9 peptide (pE₃Aβ VLPs) to explore the efficacy of active immunization.

Methods

Chemicals and antibodies

The antibodies used in the study are listed in Table 1. Fluoromount™ aqueous mounting medium (cat. no. F4680), paraformaldehyde (PFA, cat. no. P6148), phosphate-buffered saline (PBS, cat. no. P3813), Tween® 20 (cat. no. P2287), and normal goat serum (cat. no. S26-M) were purchased from Sigma-Aldrich (USA). AddaVax™ adjuvant (cat. no. vac-adx-10) was obtained from InvivoGen (France).

Table 1.

List of antibodies used in the study (Ms: mouse; Rb: rabbit)

Antibody	Manufacturer	Concentration	Cat. no	RRID
Ms anti-Aβ	Biolegend (The Netherlands)	1:1000	803001	AB_2564653
Rb anti-GFAP	Abcam (UK)	1:500	ab68428	AB_1209224
Rb anti-Iba-1	Wako (Japan)	1:500	019–19741	AB_839504
Rb anti-pE3	Synaptic systems (Germany)	1:500	218003	AB_2056424
Goat anti-Ms AlexaFluor®488	Abcam (UK)	1:500	ab150113	AB_2576208
Goat anti-Rb AlexaFluor®488	Abcam (UK)	1:500	ab150077	AB_2630356
Goat anti-Rb AlexaFluor®594	Abcam (UK)	1:500	ab150084	AB_2734147

Animals and husbandry

Seven-week-old female BALB/cOlaHsd mice (Envigo, The Netherlands) were used for primary immunization studies. For vaccine efficiency tests, 7-week-old female transgenic 5xFAD mice (B6SJL-Tg(APPSwFlLon,PSEN1*M146L*L286V)6799Vas/Mmjax, cat. no 034840, The Jackson Laboratories, USA) were employed. The hemizygotes used did not carry the retinal degeneration allele Pde6b^rd1. Mice were housed in standard conditions, with a 12-h light/dark cycle (lights off from 19:30 to 07:30), at 25 ± 1 °C and 50–60% humidity. They were group-housed in ventilated cages containing aspen wood-chip bedding, polycarbonate tunnels, and wood wool.

All procedures were performed in compliance with ARRIVE guidelines (Percie du Sert et al. 2020) and approved by the Animal Ethics Committee of the Food and Veterinary Service (permits no.136 and no.147).

Treatment schedule

The experimental design is illustrated in Fig. 1. Blood samples were collected from all mice on experimental days (D) 0, 14, 28, 42, 95, and 174 for antibody titer comparison. Vaccine candidates pE₃Aβ VLPs and AP205 VLPs, formulated with AddaVax™ vaccine adjuvant, or PBS with AddaVax™ adjuvant as a control, were administered on D1, D15, D29, and D103. Behavioral tests commenced one week after the final vaccination when the mice were 6 and 8 months old. The EZM test was performed only once, on D168. On D175, the animals were deeply anesthetized, and transcardial perfusion was conducted, followed by tissue collection for subsequent biochemical analyses.

Fig. 1.

Experimental design of the study. Blood samples were collected from the mice’s tail vein on experimental days (D) 0, 14, 28, 42, 95, and 103. Antibody titers were determined using ELISA. MWM: Morris water maze

Design of the vaccine candidate

The terminal N-end of the bacteriophage AP205 coat protein (Kirsteina et al. 2020) was genetically fused with the Aβ peptide 3-EFRHDSG-9 (Lowe et al. 2021) to produce pE₃Aβ VLPs. A Tobacco Etch Virus (TEV) protease cleavage linker sequence was introduced upstream of the Aβ peptide to ensure the release of free glutamine at the N-terminus. The expression plasmid was commercially synthesized by BioCat GmBH (Germany).

Production of recombinant proteins

E. coli strain BL21 (DE3) cells containing the TEV protease plasmid (cat. no. 8827, Addgene) were transformed with plasmids encoding the TEV-Aβ(3–9)-AP205 coat protein. Additional cells were transformed with a plasmid harboring the human glutamic cyclase (hQC) gene. Individual colonies were inoculated into 200 mL of LB medium supplemented with 30 μg/mL kanamycin and incubated overnight at 37 °C without agitation. The overnight cultures were transferred to 2.0 L of 2xTY medium and grown at 37 °C until the OD600 reached 0.6 isopropylthio-β-galactoside was added to a final concentration of 1 mM to induce protein expression, and the cultures were incubated for an additional 16 h at 20 °C. Cells were harvested by centrifugation and stored at −80 °C for further downstream applications.

Purification of chimeric VLPs

Four grams of wet cells were lysed by sonication, and the clarified lysate was precipitated overnight with 40% saturated ammonium sulfate. The precipitate was collected by centrifugation, and dissolved in 4 mL PBS containing 0.2% Tween 20, 0.5 M urea, and 1 mM phenylmethylsulfonyl fluoride. The clarified lysate was loaded onto a Sepharose 4FF column (120 mL, XK 16/70, Cytiva) equilibrated with PBS, and connected to an Akta PrimePlus chromatography system (GE Healthcare). Fractions (3 mL) were collected at a flow rate of 0.5 mL/min. Samples containing VLPs were pooled and subjected to ion exchange chromatography on a Fracto-DEAE column (20 mL, XK 16/20, Cytiva) pre-equilibrated with PBS. Bound proteins were eluted with a 100 mL NaCl gradient (0.15–1 M) in PBS at a flow rate of 3 mL/min, and fractions were collected. The purest fractions were pooled and dialyzed against PBS.

Purification of human glutaminyl cyclase

For purification, cell lysates were prepared in a lysis buffer (20 mM Tris–HCl, pH 8.0, and 300 mM NaCl) using sonication. The supernatant was passed through a HisTrap™ FF crude column (1 mL, Cytiva) equilibrated with lysis buffer containing 10 mM imidazole. Bound protein was eluted with 0.5 M imidazole in the lysis buffer. Final polishing was performed using a Superdex 200 gel filtration column (XK 16/70, Cytiva) equilibrated with PBS.

In vitro pyroglutaminylation of Aβ3–9 N-terminal glutamine

A total of 200 μg chimeric Q₃Aβ VLPs and 40 μg hQC were mixed and incubated on an end-over-end rotator at 37 °C for 2 h. The protein mixture was subsequently purified using the Superdex 200 gel filtration column (XK 16/70, Cytiva) equilibrated with PBS.

Electron microscopy

Electron microscopy was performed following established protocols (Liekniņa et al. 2019). Purified VLP samples (1.0 mg/mL) were adsorbed onto carbon-formvar-coated copper grids, negatively stained with a 1% aqueous solution of uranyl acetate, and visualized using a JEM-1230 electron microscope (JEOL Ltd., Japan) operated at 100 kV.

Mass spectrometry

Before mass spectrometry (MS) analysis, disulfide bonds linking the VLP coat protein monomers were disrupted by incubating the samples with 10 mM dithiothreitol overnight at 4 °C. Pyroglutaminylation of the N-terminal glutamine was assessed using an Autoflex MALDI-TOF mass spectrometer (Bruker Daltonics, Billerica, MA, USA). The sample for MS analysis consisted of 1 μL of protein (1.0 mg/mL in PBS) mixed with 1 μL of 0.1% TFA and 1 μL of matrix solution containing 15 mg/mL 2,5-dihydroxyacetophenone in 20 mM ammonium citrate and 75% ethanol. The mixture was transferred to a target plate, dried, and analyzed. Pyroglutaminated protein exhibited a molecular weight of approximately 20 Da smaller than the non-pyroglutaminated form (14,828.8 vs. 14,848.8 Da), consistent with the expected value of 17 Da within experimental error.

Animal studies

Seven-week-old female BALB/cOlaHsd mice were used for vaccine antigenicity studies (n = 10). Vaccines with the squalene-based adjuvant AddaVax™ (InvivoGen, vac-adx-10, France) were mixed in a 1:1 (v/v) ratio immediately before injection. Each animal received three doses of the vaccine (25 µg) or control, administered 14 days apart. Animals were weighed and monitored for potential vaccination-induced side effects. At the end of the experiment, the animals were humanely euthanized under deep surgical isoflurane anesthesia (5%), and a cardiac puncture was performed to collect final blood samples.

Eight-week-old female 5xFAD mice were used for the vaccine efficiency experiment. Three immunization groups were established: Group A (PBS, n = 13), Group B (AP205 VLP, n = 12), and Group C (pE₃Aβ VLP, n = 14). Fifty microliters of PBS, AP205 VLPs, or pE₃Aβ VLPs (0.5 mg/mL) were mixed with 50 µL of AddaVax™ adjuvant. The protein dose per mouse was 25 µg, and the total injection volume was 100 µL. Mice were immunized subcutaneously, with each mouse receiving four injections (Fig. 1).

Behavioral tests

Tracking and analysis

Noldus EthoVision software (version 11.5, Wageningen, Netherlands) was used to video-record and track animal behavior, and analyses were performed using the same software.

Open field test (OFT)

To evaluate the effects of the vaccine on locomotor function and anxiety, the OFT was conducted on experimental days 110 (D110) and 166 (D166). The OFT apparatus consisted of four white square boxes (50 cm W × 50 cm L × 20 cm H). A Basler video camera connected to EthoVision XT 11.5 was positioned above the OFT boxes to track animal movements. Mice were placed in their corresponding boxes, facing the wall, and allowed to explore freely for 10 min.

After the test, the following parameters were assessed:

• Total distance traveled,

• Mean and maximum speed,

• Number of crossings into the center zone of the arena,

• Time spent in the center zone of the arena.

Spontaneous alternation test (SAT)

To assess whether vaccination influenced the willingness of 5xFAD mice to explore novel environments, the SAT was performed on D111 and D167. The test used a Y-shaped maze with three black arms positioned at 120° angles. Mice were placed in one of the arms and allowed to explore freely for 8 min.

Behavior was recorded using EthoVision XT 11.5, and tracking data were analyzed afterward. The number of successful arm entries and the alternation index were calculated using the formula:

$$\left(S,p,o,t,a,n,e,o,u,s,,A,l,t,e,r,n,a,t,i,o,n,s,,\left(n\right)\right)/\left(T,o,t,a,l,,N,u,m,b,e,r,,o,f,,A,r,m,,E,n,t,r,i,e,s,-,2\right)\times 100$$

Elevated zero maze (EZM) test

To investigate anxiety-like behavior following active vaccination, the EZM test was conducted. The maze consisted of a circular white plexiglass ring (70 cm diameter, 5 cm width), elevated 50 cm above the ground and divided into four equal quadrants. Two quadrants were open (without walls), and the other two were enclosed by 15 cm-high walls. A 4 mm-high lip was added to the edges of the open quadrants to prevent mice from falling.

Mice were placed in one of the open quadrants and allowed to explore undisturbed for 5 min. The maze was cleaned with 70% ethanol between trials to ensure consistency.

The following parameters were recorded:

• Time spent in open quadrants (arms),

• Number of entries into open arms.

Morris water maze (MWM) test

To assess spatial learning and memory changes in 5xFAD mice, the MWM test was conducted on D112–D116 and D169–D173. The MWM apparatus consisted of a circular pool (180 cm in diameter and 75 cm in height, Ugo Basile, Italy) and a polycarbonate platform (10 cm in diameter and 30 cm in height). The pool was filled with warm water (24 ± 2 °C) so that the platform was submerged 1 cm below the water surface. Using EthoVision XT 11.5 software coupled with a video-tracking device, the pool was divided into four quadrants: northwest, northeast, southwest, and southeast.

Mice underwent four days of training, with four trials per day, to learn the platform location. Each trial started with the mouse being placed in the pool facing the wall from a different quadrant. The sequence of starting quadrants varied daily to avoid pattern recognition. After being placed in the pool, mice had 60 s to locate the platform, climb onto it, and remain there for 10 s. If a mouse failed to find the platform, the experimenter gently guided it to the platform, where it stayed for 10 s.

On D116 and D173, a probe trial was conducted 24 h after the final training session. During the probe trial, the platform was removed, and each mouse was placed in the pool for 60 s. During training, escape latency was recorded, while in the probe trial, the time spent in the target quadrant and the number of platform zone crossings were measured.

Immunofluorescence

Following behavioral tests, animals were anesthetized with a ketamine and xylazine mixture (100 and 10 mg/kg, respectively). Transcardial perfusion was performed with ice-cold PBS, followed by perfusion with 4% PFA. Brains were extracted, fixed overnight in 4% PFA, and then transferred to a 30% sucrose solution in PBS. Sections (30 µm thick) were prepared using a freezing vibratome (CM1850, Leica Biosystems, USA) at −25 °C, based on coordinates from the mouse brain atlas: from + 1.18 to –2.3 mm anterior–posterior to the bregma (Paxinos and Watson 2007). Three random coronal sections within every 200 µm of these coordinates were used for immunofluorescence assays.

Sections were rinsed three times for 5 min in 1X PBS containing 0.3% Tween-20 (PBST). Antigen retrieval was performed by incubating sections in 0.01 M sodium citrate buffer (pH 6.0) at 95 °C for 10 min. To block nonspecific binding, sections were incubated for 1 h at room temperature in a blocking solution of 10% normal goat serum (NGS) in PBST. Sections were then incubated overnight at 4 °C in 5% NGS-PBST containing primary antibodies (6E10 and Iba-1 at 1:1000, GFAP at 1:500).

The next day, sections were rinsed in PBST (three times for 5 min each), followed by incubation with secondary antibodies (goat anti-mouse or goat anti-rabbit immunoglobulins, both at 1:500) in PBST for 1 h at room temperature. Nuclei were stained with Hoechst solution (1:1000) for 5 min, rinsed with dH2O, and mounted on gelatin-coated histological slides. Slides were air-dried and coverslipped.

Double immunofluorescence

Double immunofluorescence was performed using Iba-1/6E10 and pE3/6E10 antibodies. Antigen retrieval was performed in a citrate buffer for 20 min at 90 °C, followed by blocking in 10% NGS-PBST for 1 h. Sections were incubated overnight in 5% NGS-PBST containing antibodies against 6E10 (1:1000) and Iba-1 (1:1000) or pE3 (1:500). The following day, sections were rinsed in PBST and incubated with secondary antibodies as described above. Nuclei were stained with Hoechst solution, and slides were coverslipped similarly to the regular immunohistochemistry protocol.

Digitalization and quantification

Slides were digitized promptly by a blinded investigator using a Nikon Eclipse Ti microscopy system (Nikon Europe, The Netherlands). Immunohistochemical data were analyzed using ImageJ2 software (version 1.54f, USA). A region of interest (ROI) of equal size was used for all samples. Quantifications included the number of 6E10-positive (6E10⁺) plaques, pE3⁺ plaques, and Iba-1⁺ cells. The area of 6E10⁺ and pE3⁺ plaques was measured using Icy software. GFAP staining intensity was analyzed via densitometry using the Roadbard calibration method and expressed in arbitrary units. Three sections per animal were used for quantification.

Evaluation of vaccine-induced antibody responses

ELISA plates (Greiner, Germany) were coated with 1 µg/well of avidin (cat. no. 21121, ThermoFisher, Belgium) or N-terminally truncated (3–42) pyroglutaminylated or non-pyroglutaminylated Aβ peptides (cat. no. AS-29907–01; cat. no. AS-63715–01, Anaspec) in carbonate buffer (pH 9.5) overnight at 4 °C. Plates were washed and incubated with avidin-coated biotinylated Aβ peptide (QFRHDG or pyrFRHDG, Metabion, Germany) diluted in PBS (10 µg/mL) for 1 h at 37 °C. After blocking with BSA, serum samples were diluted 1:100 and serially diluted two-fold. Plates were washed with PBST and dH2O between steps. Serum antibody binding was visualized using rabbit anti-mouse IgG HRP conjugate (1:5000, cat. no. A9044, Sigma-Aldrich, USA) for 1 h at 37 °C and developed with colorimetric O-phenylenediamine dihydrochloride detection (cat. no. P6912-100TAB, Sigma, USA). Optical density (OD) was measured at 492 nm using an ELISA plate reader (BDSL Immunoskan MS, Finland). IgG antibody responses against the AP205 VLP carrier were evaluated using the same protocol, except the ELISA plate was coated with recombinant AP205 coat protein. A positive signal was defined as OD492 values above those from wells probed with serum samples from PBS-treated animals.

Statistical analysis

Statistical analyses were performed using GraphPad Prism (version 10.3.1, GraphPad, USA). Nonparametric two-tailed Mann–Whitney tests and unpaired t-tests were used to compare antibody titers. Behavioral and immunohistochemical data were analyzed using two-way ANOVA followed by Holm-Sidak’s post hoc tests. Data are presented as mean ± standard deviation (SD), and P-values ≤ 0.05 were considered statistically significant.

Results

Vaccine production and characterization

A vaccine candidate was created using the bacteriophage AP205 VLP-based platform: chimeric pE₃Aβ VLPs that regularly and densely display the pE₃Aβ3–9 peptide on their surface. A genetic fusion technique was employed to develop the vaccine candidate, eliminating the need for chemical linkers. To obtain free glutamine, a TEV protease cleavage link was introduced into the plasmid upstream of the Aβ3–9 peptide. This link cleaves the protein sequence in the chimera, before glutamine, which can then be further treated with glutaminyl cyclase, which works both with glutamate and glutamine yielding pyroglutamate in both cases. However, TEV protease cleaves more efficiently before glutamine, which was therefore included in the construct. The TEV protease cleavage sequence is processed in vivo by co-expressing TEV-Q₃Aβ VLP with TEV protease. Virus-like particle assembly occurs only with the cleaved chimeric protein, resulting in homogeneous particles. A small amount of functional TEV protease is sufficient to efficiently release chimeric proteins from the TEV protease sequence. Reduced production of this protease was achieved by minimizing the antibiotic ampicillin load, which improved the stability of the pRK793 plasmid in the producer cells. Protein production was sufficiently high and yielded completely soluble material (Fig. 2a). pE₃Aβ VLPs were purified using straightforward chromatographic methods, including gel filtration on Sepharose 4FF, ion exchange on Fracto-DEAE, and another gel filtration on Superdex200 after pyroglutaminylation. The final sample’s purity was assessed using SDS-PAGE. Virus-like particle formation was confirmed by native agarose gel electrophoresis, which revealed a single discrete band characteristic of VLPs (Fig. 2b), and electron microscopy (Fig. 2c). Figure 2d shows the detailed process of pE₃Aβ VLP production.

Fig. 2.

Characterization of production and purification of pE₃Aβ VLPs. (a) Determination of protein production level (lane 1), solubility (lanes 2 and 3: soluble and insoluble fractions), and purity of the final vaccine material (lane 4). 15% SDS-PAGE gel stained with Coomassie blue. M: Pierce™ Unstained Protein MW Marker #26,610. (b, c) Characterization of virus-like particle formation using 1.2% native agarose gel stained with EtBr and electron microscopy. (d) Schematic representation of production of pE₃Aβ VLPs

Immunological characterization of the vaccine

The immunogenicity and potential side effects of the vaccine candidate were initially tested in BALB/c mice. The vaccine elicited a strong immunological response specific to the pE₃ form of Aβ_3–42 (Fig. 3a). Comparable antibody titers against the pE₃ form were also observed in 5xFAD mice, and these titers remained high throughout the study. No adverse post-immunization side effects were observed.

Fig. 3.

Characterization of pE3Aβ VLP-induced antibodies in female mice. (a) Comparison of antibody titers in immunized BALB/c mice against pE₃ Aβ3–42 and Aβ3–42 peptides. (b) Vaccine-induced specific antibody titer levels in 5xFAD mice throughout the experiment. (c) Anti-AP205 VLP-specific antibody titer levels in 5xFAD mice throughout the experiment. ****P < 0.0001, Mann–Whitney two-tailed t-test

Since the potential immune response of 5xFAD mice to the vaccine was not previously known, blood serum samples were regularly collected, and antibody titers against the pE₃Aβ peptide were assessed. Two months after the second booster dose, a slight decrease in antibody titers was observed (Fig. 3b, c). Consequently, the mice were immunized a fourth time, after which specific antibody titers increased again and remained sufficiently high until the end of the experiment.

Effects of the vaccine candidate on locomotor activity in 5xFAD mice

OFT was used to assess locomotor activity and exploratory behavior in 5xFAD mice. Locomotor activity measures included the total distance covered within 10 min, mean speed, and maximum speed. To evaluate exploratory behavior, we calculated the cumulative time spent in the center of the OFT arena and the number of center zone crossings.

One-way ANOVA revealed no significant differences in locomotor activity among groups of 6- and 8-month-old female 5xFAD mice. Specifically, there were no differences in the total distance moved (Fig. 4a, d), mean speed (Fig. 4b and e), or maximum speed (Fig. 4c and f). Similarly, no significant changes were observed in the number of center zone crossings (Fig. 5a and c) or time spent in the center (Fig. 5b and d) of 6- and 8-month-old female 5xFAD mice.

Fig. 4.

Locomotor performance of 6- and 8-month-old female 5xFAD mice in the OFT. Locomotor parameters included total distance moved (a, d), mean speed (b, e), and maximum speed (c, f) at 6 and 8 months of age, respectively. Data are presented as mean ± S.D., n = 12–13. Statistical analysis: one-way ANOVA followed by Holm-Sidak’s post hoc test

Fig. 5.

Anxiety assessment of female 5xFAD mice in the OFT and EZM test. OFT parameters included center zone crossings (a, c) and time spent in the center zone (b, d) at 6 and 8 months of age, respectively. For the EZM, the number of entries into open arms (e) and time spent in the open arms (f) were recorded at 8 months. Data are presented as mean ± S.D., n = 12–13. Statistical analysis: one-way ANOVA followed by Holm-Sidak’s post hoc test

The EZM test was employed as an additional method for assessing anxiety-related behavior in 8-month-old 5xFAD mice. No significant differences were observed in the number of entries into the open arms (Fig. 5e) or in the time spent in the open arms (Fig. 5f).

Spontaneous alternation

To assess hippocampal working memory, we calculated the number of alternations using the spontaneous alternation test. A higher number of alternations indicates more effective working memory.

There were no significant differences in the percentage of alternations among groups of 6-month-old 5xFAD mice (Fig. 6a) or 8-month-old 5xFAD mice (Fig. 6b). Within-group comparisons revealed no significant changes between 6 and 8 months of age in the PBS and AP205 VLP groups (Fig. 6c, d). However, in the pE₃Aβ VLP group, a significantly higher percentage of alternations was observed at 8 months compared to 6 months (P < 0.05, Fig. 6e).

Fig. 6.

Performance of 6- and 8-month-old female 5xFAD mice in the SAT. The percentage of alternations was assessed at 6 months (a) and 8 months (b). Comparisons within groups (PBS, AP205 VLP, and pE₃Aβ VLP) are shown in (c–e), respectively. Data are presented as mean ± S.D., n = 12–13. Statistical analysis: one-way ANOVA followed by Holm-Sidak’s post hoc test (A, B) or paired t-test (C–E). *P < 0.05 for pE₃Aβ VLP at 6 months vs. 8 months

Spatial learning and memory

The Morris water maze (MWM) is a hippocampal-dependent task used to assess spatial learning and memory. Mice were trained for four consecutive days to locate a hidden platform, with escape latency and swim speed recorded during each trial. On the fifth day, a probe trial was conducted where the platform was removed, and the time spent in the platform quadrant as well as the number of platform zone crossings were recorded.

Escape latencies during the four training days did not differ significantly between 6-month-old 5xFAD mice in the PBS, AP205 VLP, and pE₃Aβ VLP groups (Fig. 7a). Holm-Sidak’s post-hoc analysis only revealed a significant decrease in escape latency on training day 1 of the pE₃Aβ VLP group compared to the AP205 VLP group (P < 0.0001) at the 6-month time point. Maximum swimming speed increased over the training days for all groups of 6-month-old mice but did not differ significantly between groups (Fig. 7c).

Fig. 7.

Spatial learning and memory performance of 6- and 8-month-old female 5xFAD mice in the MWM test. Escape latency (a, b), maximum swimming speed (c, d) and time in quadrant (e, f) were measured at 6 and 8 months of age, respectively. Data are presented as mean ± S.D., n = 12–13. Statistical analysis: two-way repeated measures ANOVA followed by Holm-Sidak’s post hoc test. ***P < 0.0001 for pE₃Aβ VLP vs. PBS at 6 m

At 8 months of age, no significant differences in escape latency were observed between groups (Fig. 7b). Maximum swimming speed between groups also did not differ over the training days at this time point (Fig. 7d).

The probe trial, conducted 24 h after the final training session at 6 months revealed significant differences in the time spent in the NW vs. SE quadrant in pE₃Aβ VLP group mice (Fig. 7e) and no changes in the number of platform crossings (data not shown). Similarly, no significant differences were found between groups at 8 months in the number of platform crossings (data not shown). However, multiple significant changes were observed in the time spent in different quadrants (Fig. 7f): a) in the AP205 VLP group, where mice spent significantly longer time in the NE quadrant compared to the NW, target and SE quadrant and b) in the pE₃Aβ VLP group, where mice spent significantly longer time in the NE quadrant compared to the target one.

Effects of active immunization on Aβ plaque deposition in the hippocampus and neocortex

Extracellular deposition of Aβ plaques is a key neuropathological event in Alzheimer’s disease (AD) and occurs early in 5xFAD mice (Paxinos and Watson 2007). To evaluate the effects of active immunization on AD-like neuropathology, brain sections from 8-month-old 5xFAD female mice treated with PBS, AP205 VLP, or pE₃Aβ VLP were analyzed for plaque visualization.

Representative images of Aβ plaques in the hippocampal dentate gyrus (DG) and neocortex (NCtx) are shown in Fig. 8a and Fig. 8b, respectively, with their quantification in Fig. 8c and Fig. 8e. One-way ANOVA revealed no significant reduction in the density of amyloid plaques in the AP205 VLP and pE₃Aβ VLP groups compared to the PBS group (Fig. 8b). However, the area covered by Aβ plaques differed significantly between groups (F₂,₃₃ = 14.16, P < 0.0001, Fig. 8d).

Fig. 8.

Reduced density of and area populated by Aβ plaques in 8-month-old female 5xFAD mice treated with AP205 VLP and pE₃Aβ VLP. Representative images of Aβ plaques in the hippocampal dentate gyrus (DG, a) and neocortex (NCtx, b). Scale bar = 200 µm. Quantification of Aβ plaques in the DG (c) and NCtx (e), and plaque area measurements in the DG (D) and NCtx (f). Data are presented as mean ± S.D., n = 11–13. Statistical analysis: one-way ANOVA followed by Holm-Sidak’s post hoc test. *P < 0.01, **P < 0.001, ***P < 0.0001 vs. PBS

Holm-Sidak’s post hoc test showed a significant decrease in Aβ plaque area in the AP205 VLP and pE₃Aβ VLP groups compared to the PBS group (P = 0.0002 and P < 0.0001, respectively).

In the neocortex, the density of Aβ plaques was significantly altered between groups (F₂,₃₃ = 13.72, P < 0.0001, Fig. 8b and Fig. 8e). The density of Aβ-positive inclusions was significantly lower in the AP205 VLP and pE₃Aβ VLP groups compared to the PBS group (P < 0.0001 and P = 0.0003, respectively). The area taken up by Aβ plaques in the neocortical region also differed significantly between groups (F₂,₃₃ = 8.76, P = 0.0009, Fig. 8f). Holm-Sidak’s test confirmed a marked reduction in Aβ plaque area in both treatment groups (P = 0.0019 for AP205 VLP vs. PBS and P = 0.0009 for pE₃Aβ VLP vs. PBS).

We then double-stained Aβ42 and pyroglutamate-modified plaques with 6E10 and anti-pE3 antibodies, respectively, and calculated the density of pE₃Aβ⁺ plaques and the area they occupy between groups (Fig. 9). Naturally, these plaques co-localized with 6E10-stained Aβ plaques. The area occupied by pE₃Aβ⁺ plaques in the hippocampal DG was significantly higher in the AP205 VLP group (P = 0.0042, Fig. 9d). Besides that, neither the pE₃Aβ⁺ density (Fig. 9c and 9e) nor the area they occupied (Fig. 9d and f) was significantly different between any of the other groups.

Fig. 9.

Co-localization of pyroglutamated and Aβ42 plaques in 8-month-old female 5xFAD mice treated with AP205 VLP and pE₃Aβ VLP. Representative images show pE3 (red) and 6E10 (green) staining in the hippocampal DG (a) and NCtx (b). Scale bar = 200 µm. Image excerpts (red) show pE3 and 6E10 co-staining at 200 × magnification. Quantification of pE₃Aβ plaque density in the DG (c) and NCtx (e), and percentage of area covered by plaques in the DG (d) and NCtx (f). Data are presented as mean ± S.D., n = 11–13. Statistical analysis: one-way ANOVA followed by Holm-Sidak’s post-hoc test. **P < 0.01 vs. PBS

Effects of active immunization on astroglial and microglial activation

Astroglial activation, or astrogliosis, occurs when astrocytes undergo morphological, molecular, and functional changes in response to stimuli. This condition can lead to neuronal damage and is an early event in AD progression. Astroglial reactivity was assessed by measuring the optical density of the glial marker GFAP (Fig. 10). The optical density of GFAP did not differ significantly between groups in the DG (Fig. 10c) or the neocortex (Fig. 10d).

Fig. 10.

Astroglial activation is unaltered in the dentate gyrus (DG) and neocortex (NCtx) of 8-month-old 5xFAD female mice treated with AP205 VLPs and pE3Aβ VLPs. Representative images (A, B) show GFAP staining (green) in the hippocampal DG and NCtx, respectively. The scale bar was 200 µm. Measurements of GFAP optical density were done in the DG (C) and NCtx (D). Data are shown as mean values and SD (n = 11–13). Statistical analysis was one-way ANOVA followed by Holm-Sidak’s post hoc test

Microglia are immune cells in the brain that phagocytize Aβ and apoptotic cells, thereby promoting their clearance from the brain parenchyma. To evaluate whether the reduced plaque area observed in 5xFAD mice treated with vaccine candidates was associated with decreased neuroinflammation, microglial reactivity was assessed in the hippocampal DG and neocortex.

The density of Iba-1-positive cells in the hippocampal DG was significantly different between groups (F₂,₂₇ = 6.89, P = 0.004, Fig. 11a and Fig. 11c). Holm-Sidak’s analysis showed a significant reduction in Iba-1-positive cell density in the AP205 VLP and pE₃Aβ VLP groups compared to the PBS group (P < 0.002 and P < 0.018, respectively).

Fig. 11.

Microglial reactivity is decreased in the hippocampal DG and NCtx of 8-month-old female 5xFAD mice treated with AP205 VLP and pE₃Aβ VLP. Representative images of Iba-1 staining (green) in the DG (a) and NCtx (b). Scale bar = 200 µm. Quantification of Iba-1-positive cell density in the DG (c) and NCtx (d). Data are presented as mean ± S.D., n = 11–13. Statistical analysis: one-way ANOVA followed by Holm-Sidak’s post-hoc test. *P < 0.05 and **P < 0.01 vs. PBS

Similarly, in the neocortex, the density of Iba-1-positive cells was significantly different between groups (F₂,₃₁ = 4.28, P = 0.023, Fig. 11b and Fig. 11d). Post-hoc analysis revealed a significant decrease in the density of Iba-1-positive cells in the AP205 VLP and pE₃Aβ VLP groups compared to the PBS group (P = 0.028 and P = 0.021, respectively).

Double staining of microglia and Aβ plaques showed that microglial cells were primarily located near Aβ plaques in all groups (Fig. 12a and Fig. 12b). However, the area taken up by Iba-1-positive cells in close proximity to Aβ plaques did not differ significantly between groups (Fig. 12c and Fig. 12d), and neither did their density (Fig. 12e and Fig. 12f).

Fig. 12.

Microglial co-localization near Aβ plaques in 8-month-old female 5xFAD mice treated with AP205 VLP and pE₃Aβ VLP. Representative images of Iba-1 (red) and 6E10 (green) staining in the DG (a) and NCtx (b). Scale bar = 200 µm. Quantification of area populated by Iba-1 and 6E10-positive cells in the DG (c) and NCtx (d), as well as the density of the co-localized cells in the DG (e) and NCtx (f). Data are presented as mean ± S.D., n = 11–13. Statistical analysis: one-way ANOVA followed by Holm-Sidak’s post hoc test

Discussion

In this study, we utilized the bacteriophage AP205 VLP platform to develop candidate compounds for an active immunization approach targeting the pE₃ Aβ3–9 epitope (pEFRHDS). Mice were administered either AP205 VLPs or pE₃Aβ VLPs once every 2 weeks (a total of four doses), following the protocol described by Illouz et al. (Illouz et al. 2021). We observed a significant reduction in Aβ plaque area compared to the adjuvanted PBS control group. However, this effect appeared to be largely due to the AP205 VLP carrier itself, as the addition of the pE₃ Aβ3–9 epitope did not result in a statistically significant further decrease in Aβ plaque area. To our knowledge, this is the first study to clearly demonstrate the beneficial effect of a VLP carrier alone in reducing Aβ plaque load.

In previous studies, the effect of VLP carriers without an epitope has either not been observed or was not investigated. For example, Wiessner et al. (Wiessner et al. 2011) evaluated the effects of phage Qβ carrier VLPs but found no changes in Aβ plaque area. Similarly, Fu et al. (Fu et al. 2017) reported that immunization with the norovirus P particle carrier alone did not reduce Aβ plaque area. In contrast, in other studies, the impact of the carrier alone remains unknown. For instance, Bach et al. (Bach et al. 2009) demonstrated that Aβ-displaying retroviral VLPs reduced both the number and area of amyloid plaques, but the study did not include a control group with the carrier VLP alone. Similarly, Zagorski et al. (Zagorski et al. 2023) showed the efficacy of their vaccine compared to PBS, but they did not report the effects of their non-VLP carrier, multi-TEP.

It is worth noting that we did not include wild-type mice as controls in our study. This decision was based on the well-documented cognitive impairments observed in 5xFAD mice and the absence of Aβ accumulation in the brains of healthy mice. The female 5xFAD mice used in our experiments exhibited Aβ plaques in the neocortex and hippocampal dentate gyrus (DG). However, all 5xFAD groups performed similarly in the behavioral tests conducted. One possible explanation is that the 6-month time point may have been too early to detect significant cognitive and locomotor impairments, despite studies showing that working memory deficits occur in 5xFAD mice at this age (Habashi et al. 2022; Shukla et al. 2013). Even though there was a clear preference of mice to stay in the NE quadrant in the probe trial at 8 months, it can be explained by the fact that this was the release quadrant in this trial, and multiple animals (especially those from the AP205 VLP group) chose to stay in it rather than engage in locating the platform. Therefore, it seems that the animals in all of the groups did not learn the location of the platform, indicating similar impairments in spatial memory. Although at 8 months the pE₃Aβ VLP-treated group showed higher percentage of alternations than at 6 months, this group had lower (although insignificantly) alternation percentage at 6 months compared to other study groups.

Our immunohistochemical analysis revealed that active immunization did not reduce the number of Aβ or pE₃Aβ plaques in the hippocampal DG. However, pE₃Aβ VLP treatment significantly reduced Aβ (but not pE₃Aβ) plaque numbers in the neocortex. Additionally, pE₃Aβ VLP administration resulted in smaller Aβ plaque areas in both brain regions but had no significant effect on the area of pE₃Aβ plaques. While astroglial response remained unaltered by pE₃Aβ VLP treatment, microglial response was significantly reduced in both the hippocampal DG and neocortex of 5xFAD mice treated with either pE₃Aβ VLPs or AP205 VLPs. This finding aligns with our observation that pE₃Aβ VLP administration decreased Aβ plaque areas, with most microglial cells located around plaques in all experimental groups.

An important limitation of the study is that only female 5xFAD mice were used in this study, only female 5xFAD mice were used in this study, therefore we were not able to examine the sex-related differences in behavior and biochemical parameters. Since there are multiple reports on sex differences in 5xFAD mice (Zagorski et al. 2023; Todorovic et al. 2020; Poon et al. 2023), we cannot assume that the results of our study would be similar if it was done in male 5xFAD mice.

In conclusion, our study demonstrates a modest effect of AP205 VLPs and pE₃Aβ VLPs in reducing Aβ load and microgliosis in the brains of 8-month-old female 5xFAD mice. Given the positive effects observed with the carrier VLPs, further exploration of AP205 VLPs displaying other Aβ or tau epitopes as potential vaccine candidates against Alzheimer’s disease is warranted.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

V.P.: formal analysis, investigation, visualization, writing – original draft, writing – review & editing. I.L.: conceptualization, formal analysis, investigation, methodology, visualization, writing – original draft, writing – review & editing. D.S.: investigation, writing – review & editing. S.A.: investigation, writing – review & editing. B.L.R.: Formal analysis, writing – review & editing. J.U.: Formal analysis, writing – review & editing. B.J.: project administration, supervision, writing – review & editing. K.T.: conceptualization, formal analysis, funding acquisition, methodology, project administration, supervision, validation, writing – review & editing. The authors declare that all data were generated in-house and that no paper mill was used.

Funding information

This study was funded by the State Research Programme (project “BioMedPharm” nr. VPP-EM-BIOMEDICĪNA-2022/1–0001).

Data availability

All source data for this work (or generated in this study) are available upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.