|Home | About | Journals | Submit | Contact Us | Français|
Active amyloid-β (Aβ) immunotherapy is under investigation to prevent or treat early Alzheimer's disease (AD). In 2002, a Phase II clinical trial (AN1792) was halted due to meningoencephalitis in ~6% of the AD patients, possibly caused by a T-cell-mediated immunological response. Thus, generating a vaccine that safely generates high anti-Aβ antibody levels in the elderly is required. In this study, MER5101, a novel conjugate of Aβ1-15 peptide (a B-cell epitope fragment) conjugated to an immunogenic carrier protein, diphtheria toxoid (DT), and formulated in a nanoparticular emulsion-based adjuvant, was administered to 10 mo-old APPswe/PS1ΔE9 transgenic (Tg) and wild-type (Wt) mice. High anti-Aβ antibody levels were observed in both vaccinated APPswe/PS1ΔE9 Tg and Wt mice. Antibody isotypes were mainly IgG1 and IgG2b, suggesting a Th2-biased response. Re-stimulation of splenocytes with the Aβ1-15:DT conjugate resulted in a strong proliferative response, whereas proliferation was absent after re-stimulation with Aβ1-15 or Aβ1-40/42 peptides, indicating a cellular immune response against DT while avoiding an Aβ-specific T cell response. Moreover, significant reductions in cerebral Aβ plaque burden, accompanied by attenuated microglial activation and increased synaptic density, were observed in MER5101 vaccinated APPswe/PS1ΔE9 Tg mice compared to Tg adjuvant controls. Lastly, MER5101 immunized APPswe/PS1ΔE9 Tg mice showed improvement of cognitive deficits in both Contextual Fear Conditioning (CFC) and the Morris Water Maze (MWM). Our novel, highly immunogenic Aβ conjugate vaccine, MER5101, shows promise for improving Aβ vaccine safety and efficacy and therefore, may be useful for preventing and/or treating early AD.
Alzheimer's disease (AD) is the most common form of dementia. AD is characterized pathologically by the aggregation and deposition of cerebral amyloid beta (Aβ), neuritic plaques, gliosis, neuron loss, and neurofibrillary tangles (Selkoe, 2001). Aβ is thought to play a critical role in AD pathogenesis (Golde, 2003), suggesting that therapies to inhibit its production, enhance its degradation or improve its clearance from the brain would benefit AD patients. Schenk first reported Aβ lowering by active Aβ1-42 immunization in PDAPP transgenic (Tg) mice (Schenk et al., 1999). Thereafter, many Aβ immunization studies have reported reduced cerebral Aβ levels and/or improved cognition in mice (Bard et al., 2000; Janus et al., 2000; Lemere et al., 2000; Morgan et al., 2000; Weiner et al., 2000; Bard et al., 2003), non-human primates (Gandy et al., 2004; Lemere et al., 2004), and to some extent, humans (Nicoll et al., 2003; Bayer et al., 2005). The AN1792 human clinical trial was halted in 2002 due to meningoencephalitis in 6% of immunized AD patients (Orgogozo et al., 2003; Ferrer et al., 2004; Gilman et al., 2005), possibly due to T cell recognition of the self-antigen, Aβ1-42, in combination with a strong, Th1-based adjuvant, QS21 (Cribbs et al., 2003). Although the dosing ended prematurely, only 19.7% of immunized patients developed an antibody response (Gilman et al. 2005)
Whereas B-cell epitope resides within Aβ1-15 in humans (Geylis et al., 2005), monkeys (Lemere et al., 2004) and mice (Lemere et al., 2000; McLaurin et al., 2002; Agadjanyan et al., 2005), T-cell epitopes reside within Aβ16-42 (Cribbs et al., 2003; Monsonego et al., 2003). Immunization with Aβ N-terminal derivatives generated antibodies, lowered Aβ (Bard et al., 2003; Maier et al., 2006; Seabrook et al., 2006; Seabrook et al., 2007), and protected cognition (Sigurdsson et al., 2004; Maier et al., 2006) in AD-like mouse models. A Th2-biased adjuvant may help safely induce effective immune responses in the absence of Th1-mediated events (Lemere et al. 2000; Lemere et al. 2002; Maier et al. 2005; (Asuni et al., 2006; Ghochikyan et al., 2006).
In this study, we used a novel Aβ B-cell epitope vaccine, MER5101, composed of Aβ1-15 conjugated via 7 amino acids to a carrier protein, diphtheria toxoid (DT), and formulated in a Th2-biased adjuvant, MAS-1. MAS-1 adjuvant is a water-in-oil nano-particle emulsion that differs from IFA and uses metabolizable components that are better tolerated than the mineral oil-based IFA. MAS-1 enhances antibody responses to protein antigens up to 50-fold compared to alum and more than 10-fold compared to oil-in-water adjuvants (unpublished results). Self-antigen conjugate vaccines adjuvanted with MAS-1 targeting gastrin and gonadotrophin-releasing hormone have been tested clinically in approximately 1,500 cancer patients, including gastrointestinal adenocarcinoma and prostate cancer(Simms et al., 2000; Smith et al., 2000; Brett et al., 2002; Gilliam et al., 2004; Ajani et al., 2006). These patients included elderly and those further immunocompromised by late stage cancer and chemotherapy; nevertheless, immune response rates reached 80% across all groups. Here, we assessed the efficacy of MER5101 in AD-like Tg mice.
Immunization treatment was performed in 10 mo-old APPswe/PS1ΔE9 Tg mice on a mixed C57BL/6 × DBA2 (B6D2F1) background and B6D2F1 Wt mice (Taconic Farms, Germantown, NY). APPswe/PS1ΔE9 Tg mice harbor the Swedish APP (K594N/M595L) and PS1dE9 (deletion of exon 9) human transgenes under a mouse prion protein promotor (Jankowsky Human. Mol. Genetics 2004). Tg breeders were originally obtained from The Jackson Laboratory (Bar Harbor, ME) and bred in our colony by crossing a APPswe/PS1ΔE9 mice on a C57BL6 background with DBA/2 Wt mice. APPswe/PS1ΔE9 mice begin to exhibit cognitive deficits starting at 6 mo, which are exacerbated with age (Puolivali, J. Neurobio Dis 2002; Park, JH. J. Neurosci. 2006 Gimbel. DA. J. Neurosci 2010). All animal use was approved by the Harvard Standing Committee for Animal Use and was in compliance with all state and federal regulations.
The MER5101 vaccine is composed of multiple copies of Aβ1-15 peptide conjugated to diphtheria toxoid (DT) using a seven amino acid spacer to enhance presentation of self epitopes and immunogenicity, and formulated in Mercia's Th2-biased adjuvant, MAS-1 (Mercia Pharma, Inc., New York, NY), a stable water-in-oil (W/O) emulsion. Conjugate antigen was prepared by linking Aβ1-15 peptide to DT with the heterobifunctional crosslinker ε-maleimidocaproic acid N-hydroxysuccinimide ester, yielding conjugate with a molar substitution ratio of 18.6 peptides per DT by Amino Acid Analysis of conjugate. MER5101 vaccine formulations were prepared on each injection day by combining aqueous phase containing antigen with MAS-1 in a 30:70 water-to-oil ratio by weight, vortexing to prepare a pre-emulsion, then emulsifying by repeated passage between two syringes connected with a plastic stopcock. The same volume of sterilized PBS solution was administered to the MAS-1 adjuvant (vehicle) controls. Vaccinated APPswe/PS1ΔE9 Tg (n=5; 2F, 3M) and Wt mice (n=6; 2F, 4M) received five subcutaneous (sc) injections of 100 μg Aβ1-15:DT conjugate in an injection volume of 0.1 mL MER5101 per dose. Vehicle control Tg (n=6; 2F, 4M) and Wt (n=5; 2F, 3M) mice received five s.c. injections of 0.1 mL/dose of MER5101 placebo, comprising PBS (no antigen) emulsified with MAS-1. The first two injections were administered bi-weekly, followed by 3 additional injections given four weeks apart. Mice underwent cognitive testing at 14 months of age, followed by euthanasia.
Plasma was collected from the tail vein as previously described (Maier et al., 2005). Ten days after the final immunization, mice were killed by CO2 inhalation and transcardially perfused with 20 ml PBS. The brain was removed and divided sagittally. One hemibrain was fixed for 2 h in 10% buffered formalin and embedded in paraffin as previously described (Leverone et al., 2003; Maier et al., 2005); the other hemibrain was snap frozen in liquid nitrogen and stored at -80°C for biochemical analysis. TBS soluble and 5M guanidine–soluble (i.e. TBS insoluble) brain homogenates were prepared as previously reported (Weiner et al., 2000).
Low pH dissociation treatment for plasma samples was conducted as previously described (Li et al., 2007) with some modification. Briefly, plasma was diluted 1:100 with dissociation buffer (1.5% BSA and 0.2 M glycine-acetate in PBS) at pH 3.5, and incubated at room temperature (RT) for 20 min. The plasma was then transferred into the sample reservoir of a centrifugal filter device, YM-30 (30,000 MW cut-off; Millipore) and centrifuged at 8,000 × g for 20 min at RT. The solution remaining in the sample reservoir was collected and adjusted to pH 7.0 with 1M Tris buffer, pH 9.0. Anti-Aβ antibodies in pre- and post-dissociated plasma samples were measured by ELISA as previously described (Spooner et al., 2002). ELISAs for antibody isotypes and epitope mapping were performed as previously reported (Lemere et al., 2002). The standards for the IgG1 and IgG2b isotype ELISAs were purchased from Southern Biotechnology Associates Inc., while the standards for the IgG2a and IgM ELISAs were purchased from Invitrogen (Carlsbad, CA). The following overlapping Aβ fragment peptides were used for Aβ epitope-mapping by ELISA: Aβ1-15, Aβ1-7, Aβ3-9, Aβ3-13, Aβ7-12, Aβ11-25, Aβ26-42, Aβ1-40 (Biopolymer Laboratory, UCLA, Los Angeles, CA).
Levels of Aβx-40 and Aβx-42 in brain homogenates (TBS soluble and guanidine soluble) and in plasma were measured by chemiluminescent BetaMarkTM x-40 and x-42 ELISA kits (Covance), respectively. ELISA for Aβ1-total level in brain homogenates and plasma was performed as described previously (Peng et al., 2010).
Conditioned media from Chinese hamster ovary (CHO) cells stably transfected to express mutant human APP (cell line 7PA2, kind gift from Dr. Dennis Selkoe, Boston, MA) or non-transfected CHO cells was centrifuged to remove cellular debris. The conditioned media was then incubated with plasma (1:50) from immunized (adjusted to 1 mg/mL of anti-Aβ antibody) or control mice. The monoclonal anti-Aβ antibody 6E10 (1:100) (Covance Laboratories, Dedham, MA, USA) served as a positive control. A standard immunoprecipitation procedure using Protein G beads (Pierce, Rockford, IL) was performed, with the products being electrophoresed on 12% Bis-Tris gels (Invitrogen) before being transferred to 0.2μm nitrocellulose membranes. The anti-Aβ polyclonal antibody, R1282 (gift Dr. Dennis Selkoe, Boston, MA), was used to probe the blots, and was visualized using a Li-Cor scanner (LI-COR Biosciences, Lincoln, NE).
All cell culture reagents were obtained from Invitrogen (Carlsbad, CA, USA). Spleens of mice were pooled within each treatment group. Splenocytes were isolated, cultured and re-stimulated as previously described (Maier et al., 2005). Aβ1-15:DT, Aβ1-15 and Aβ1-40/42 (3:1 ratio) peptides were added to splenocytes at final concentrations of 0, 0.5, 5 or 50 μg/ml in triplicate wells. To measure proliferation, 1 μCi of [3H]-thymidine was added to cells at 72 h. Eighteen hours later, the cells were harvested and thymidine incorporation was measured using a 1450 Microbeta liquid scintillation counter, (PerkinElmer, Waltham, MA). The stimulation index (SI) was calculated using the following formula: counts per minute (CPM) of well with antigen/CPM with no antigen. A SI index greater than 3 indicates a proliferative cellular immune response of the splenocytes to the peptide.
Splenocyte cell culture supernatants were collected at 72 h, just prior to the addition of 3H-Thymidine and stored at -80°C for cytokine assays. The Mouse Pro-Inflammatory TH1/TH2 9–plex from MesoScale Discovery (Gaitherburg, MD) was used, following the manufacturer's protocol, to measure cytokine release. In brief, cell culture supernatants from the following re-stimulation conditions were loaded undiluted in duplicate: 50μg/mL Aβ1-40/42; 50μg/mL Aβ15:DT; media alone (negative control) and 5μg/mL Concavalin A (positive control). The Multi-Spot ELISA plate was pre-coated with antibodies specific for the following cytokines: IFN-Y, IL-1β, IL-10, IL-12 total, IL-2, IL-4, IL-5, KC and TNF-α and detected with the appropriate SULFO-TAG detection antibodies. Light emitted upon electrochemical stimulation was read using a SECTOR Imager 2400A.
IHC was performed on 10 micron paraffin sections of human AD brain and mouse brain as reported previously (Lemere et al., 2000) using Vector Elite ABC kits (Vector Laboratories, Burlingame, CA). The following antibodies were used for neuropathological analysis: R1282 (a general Aβ polyclonal antibody, 1:1000; gift D. Selkoe, Brigham and Women's Hospital, Boston, MA), anti-Aβ 6E10 (raised against Aβ1-16 peptide, 1:1000; Covance, Dedham, MA), anti-Iba-1 (a marker for both resting and activated microglia, 1:1000; Wako Chemicals, Richmond, VA), anti-CD45 (a marker for activated microglia, 1:5000; Serotec, Raleigh, NC), anti-GFAP (a marker for astrocytes, 1:500; Dako, Carpinteria, CA), anti-synaptophysin (SYP; a pre-synaptic protein marker, 1:200; Sigma-Aldrich, St. Louis, MO), anti-post-synaptic density protein-95 (PSD-95; a post-synaptic protein marker, 1:200; EMD Millipore, Billerica, MA), anti-mouse CD5 (a marker for T-lymphocytes, 1:200; BD Pharmingen/Biosciences, San Jose, CA). One-percent aqueous Thioflavin S (Sigma Aldrich, St. Louis, MO) was used to visualize fibrillar amyloid in plaques and blood vessels. Hemosiderin staining using 2% ferrocyanide (Sigma) in 2% hydrochloric acid was used to detect microhemorrhages. Quantification of total R1282 IR and Thioflavin S (Thio S) staining was conducted using BIOQUANT image analysis (Nashville, TN). The threshold of detection was held constant during the entire analysis. The percent area occupied by R1282 immunoreactivity (IR) and Thio S staining was quantified for three equidistant sagittal planes, 300 microns apart per mouse in the following regions: hippocampus (HC), cortex (CTX), and cerebellum (CB). Thioflavin S-positive vascular amyloid in CB was evaluated semiquantitatively for 3 equidistant sagittal planes 300 μm apart per brain region per mouse using the following criteria: ‘0’ = no Thio S positive blood vessels; ‘1’ = 1–10 Thio S positive blood vessels; ‘2’ = 11–20 Thio S positive blood vessels, and ‘3’ = 21 or more Thio S positive blood vessels.
The mouse-monoclonal anti-APP antibody, 22C11 (1:1000; EMD Millipore Corp, Billerica, MA), was applied to paraffin sections followed by Alexa Fluor 488 goat anti-rabbit IgG secondary antibody (1:100; Invitrogen, Calsbad, CA) for immunofluorescent labeling of dystrophic neurites within neuritic plaques.
Cognitive efficacy of MER5101 was evaluated using Contextual Fear Conditioning (CFC) and Morris water maze (MWM) testing. The operator was blinded to treatment group and genotype throughout all behavioral testing. The CFC test was performed over two days. On day 1, the mouse was placed in a chamber with visible cues on the sidewall for 3 min and then given a 30s sound signal followed by a 0.7 mA electric stimulation on the floor of the chamber. Fifteen seconds after the electric shock, the mouse was moved back to its home cage. On day 2, the mouse was placed in the same chamber with the same environment but without the shock for 3 min. During this period, mouse freezing behavior was checked every 10s and the number of freezings recorded.
MWM test was conducted as previously described (Peng et al. 2010) Spatial learning was evaluated by the escape latency for each mouse to reach and climb onto the platform in 5 daily hidden platform trials. Spatial memory was evaluated by the analysis of the dwelling time in each quadrant of the pool and the analysis of an annulus-crossing index (ACI) in two probe trials, which were performed 2 h and 24 h after the last hidden platform trial on day 5. The ACI represents the number of crosses over the platform site in the target quadrant (quadrant with escape platform) adjusted for crosses over the platform sites in alternative quadrants. A visible cue test was performed to confirm the visual abilities of each mouse.
Prism 4 software (GraphPad Inc., San Diego, CA) was used to analyze the data. All data were expressed as mean ± the standard error of the mean (SEM). Kruskal-Wallis nonparametric one-way analysis of variance (ANOVA) analysis was used with the treatment as a between-subject factor, and training days as a within-subject factor for CFC and the hidden platform trial of MWM. The Mann-Whitney U (MWU) non-parametric analysis was used to compare ACI and percent dwelling time values from the probe trials in MWM, percent area of immunoreactivy in IHC and histology, and Aβ levels by ELISA. A value of p≤0.05 was considered significant for all statistical tests.
Five s.c. immunizations were administered to 10 mo-old mice on days 0, 14, 42, 84 and 112. Plasma was collected at one day prior to the first immunization (baseline) and on days 28, 56, 96 and 121. Anti-Aβ antibody levels in mouse plasma samples were measured using ELISA. Anti-Aβ antibodies were detected in both APPswe/PS1ΔE9 Tg and Wt mice immunized with Aβ1-15:DT in MAS-1 adjuvant by day 28 (Figure 1A). Peak antibody levels reached 978.6±242.3 μg/mL and 1893.8±95.6 μg/mL in APPswe/PS1ΔE9 Tg and Wt mice, respectively, by day 56 and declined thereafter. Vehicle-treated (MAS-1 alone) mice did not produce detectable anti-Aβ antibodies. Low pH dissociation of antibodies from their antigens in plasma (Li et al., 2007) revealed that anti-β antibody levels in immunized Tg mice were maintained at high levels (1070.0±416.8 μg/mL and 1008.7±205.2 μg/mL at days 96 and 121, respectively), whereas the antibody levels in immunized Wt mice were unchanged by low pH dissociation (Figure 1B).
Plasma antibodies from vaccinated Tg and Wt mice bound Aβ plaques in human AD cortical brain sections (Figure 1C and E) while plaques were not labeled by plasma from vehicle controls (Figure 1D and F). To further confirm the plasma antibody specificity for recognizing Aβ, conditioned media from 7PA2 cells, known to produce naturally-secreted Aβ monomers and oligomers but not aggregates (Walsh et al. 2005), was examined by immunoprecipitation (IP). Plasma anti-Aβ antibodies from vaccinated mice bound Aβ monomers and oligomers (Figure 1H).
Ig isotyping of anti-Aβ antibodies was performed on the final bleed sample (day 121) using isotype-specific ELISAs (Fig. 2A). IgG1, a non-inflammatory Th2 Ig, was the most predominant Ig isotype in vaccinated APPswe/PS1ΔE9 Tg and Wt mice, followed by IgG2b (Th2), which was about three-times higher than the level of the pro-inflammatory IgG2a isotype (Th1). Low levels of IgM were detected in all four groups of animals.
To further confirm the isotype specificity, plasma from immunized mice was used for immunostaining of human AD brain sections in combination with IgG-specific biotinylated secondary antibodies (including IgG1, IgG2b and IgG2a). Consistent with the ELISA results, plaque staining with immunized mouse plasma was strongest using the IgG1 secondary antibody (Figure 2B), with slightly less intense staining with the IgG2b secondary antibody (Figure 2C), and little or no detectable staining with the IgG2a secondary antibody (Figure 2D).
Pre-incubation of the immunized mouse plasma with short Aβ peptides prior to ELISA indicated that the majority of antibodies recognized a B cell epitope within the Aβ1-15 region, but did not recognize epitopes within the Aβ mid-region or C-terminus (Figure 2E).
To characterize the cellular immune response, splenocytes were isolated and pooled within each group of immunized animals, and re-stimulated with various concentrations of Aβ1-15:DT conjugate, unconjugated Aβ1-15 peptide or full-length Aβ1-40/42 (3:1 ratio). Proliferation (i.e., stimulation index, SI) was highest for splenocytes from immunized Wt mice re-stimulated with the Aβ1-15:DT conjugate, whereas re-stimulation with the same conjugate induced moderate proliferation in splenocytes from immunized APPswe/PS1ΔE9 Tg and Wt mice (Figure 3A). The proliferative response was enhanced with increasing amounts of stimulating peptide, indicating a dose-dependent cellular immune response. No proliferation was observed after re-stimulation with either Aβ1-15 peptide (Figure 3B) or full-length Aβ1-40/42 peptide (Figure 3C) in any mouse group, which is consistent with DT providing carrier function for the antigen construct and the maintenance of self tolerance to Aβ in the Aβ1-15:DT immunized Tg mice.
Cytokine analysis of cell supernatants from the Wt mice demonstrated elevated levels of IFNγ, IL-10 and IL-2, and IL-4, IL-5 and KC to lesser levels, in Aβ1-15:DT compared to Aβ40/42 restimulated cells from s.c. immunized mice (Table 1). Compared to the Wt mouse cells, and in accord with lower proliferation levels seen in the Tg versus Wt mouse cells, supernatants from cells taken from the Tg mice showed smaller cytokine effects in response to stimulation with Aβ1-15:DT, including elevations in IFNγ and KC, and reduced production of IL-12. The Wt mouse cell data could indicate the induction of Treg cells; FACS analysis of the in vitro stimulated splenocytes indicated that MER5101 specifically induced T cells with the CD4+/CD25+/Foxp3 regulatory phenotype responsive to Aβ15:DT, but not Aβ1-15 epitope or intact Aβ1-40/42 (Table 2). Treg induction could have therapeutic implications for vaccine safety, as IL-10 and IFNγ production by Treg cells has been shown to relate to therapeutic anti-inflammatory activity in NOD mice immunized s.c. with Insulin B chain 9-23 in IFA (Fousteri G, 2010).
Plaque deposition in APPswe/PS1ΔE9 Tg mice begins in cortex at ~5-6 months of age, after which, the plaque burden increases and extends to hippocampus (HC) and cerebellum with aging(Garcia-Alloza et al., 2006) MER5101 vaccination significantly reduced total Aβ immunoreactivity (IR) detected by R1282, a general Aβ polyclonal antibody, in hippocampus (HC) by 38.3% (p=0.0089), cortex (CTX) by 29.5% (p=0.041) and in cerebellum (CB) by 42.1% (p=0.0087) in APPswe/PS1ΔE9 Tg mice vs. Tg vehicle controls (Figure 4A-G). Aβ fibrillar plaques (Thioflavin S-positive) were also markedly diminished in MER5101 immunized mice by 45.4% (p=0.0079), 52.1% (p=0.0069) and 55.2% (p=0.028) in HC, CTX and CB, respectively (Figure 4H-N).
Vascular amyloid is most prominent in the leptomeningeal blood vessels in the cerebellum in APPswe/PS1ΔE9 Tg mice. Therefore, we semiquantitatively scored the amount of vascular amyloid in MER5101 vaccinated and vehicle control Tg mice. We observed no significant differences (p=0.16) in Thio S-positive vascular amyloid between MER5101 vaccinated and vehicle Tg mice (data not shown).
Biochemical analysis of cerebral Aβ levels by ELISA showed that TBS-soluble and insoluble (guanidine-HCL-soluble) levels of Aβx-40 were slightly but not significantly decreased by 13.6% (p=0.069) and 10.1% (p=0.33), respectively, in vaccinated Tg mice compared to Tg vehicle controls (Figure 4O). Soluble Aβx-42 was modestly reduced by 14.6% (p=0.089), while insoluble Aβx-42 was significantly decreased by 29.0% (p=0.041, Figure 4P) in immunized Tg mice vs. Tg vehicle controls. In addition, soluble and insoluble levels of Aβ1-total were lowered by 45.0% (p=0.028) and 42.8% (p=0.026), respectively (Figure 4Q) in the vaccinated Tg mice. Plasma levels of Aβx-40 (Figure 4O), Aβx-42 (Figure 4P) and Aβ1-total (Figure 4Q) were unchanged by MER5101 vaccination.
Plaque-associated microgliosis and astrogliosis have been reported previously in human AD brains (Itagaki et al., 1989; Mattiace et al., 1990) and in AD-like animal models (Frautschy et al., 1998; Gordon et al., 2002). In this study, immunoreactivities for Aβ (Figure 5A, B), Iba-1 (Figure 5C, D), a marker for both resting and activated microglia, and CD45 (Figure 5E, F), a marker for activated microglia, were reduced in the hippocampus of MER5101 immunized APPswe/PS1ΔE9 mice compared to Tg vehicle controls. Changes in GFAP-positive astrocytes were less obvious between Tg vehicle controls and MER5101 vaccinated Tg mice (Figure 5G, H).
Synaptic immunostaining was performed using a pre-synaptic marker, synaptophysin (SYP) and a post-synaptic marker, post-synaptic density-95 (PSD-95). APPswe/PS1ΔE9 Tg mice showed reduced synaptic density in hippocampus compared to age-and-gender-matched Wt control mice (Figure 6). MER5101 vaccinated Tg mice showed significantly higher densities of both SYP (Figure 6B-D) and PSD-95 (Figure 6F-H) in hippocampal CA1 and CA3 regions compared to Tg vehicle controls (Figure 6A and E), suggesting that hippocampal synaptic density, thought to be reduced in APPswe/PS1ΔE9 mice due to Aβ toxicity (Blanchard et al., 2003; Trinchese et al., 2004), was restored by MER5101 immunization. Neuritic plaques labeled with an APP monoclonal antibody, 22C11, were reduced in hippocampus in immunized Tg mice (Figure 6J) compared to Tg vehicle controls (Figure 6I). In addition, neither CD5-positive T cells nor hemosiderin-positive microhemorrhages (Pfeifer et al., 2002) were observed in the brain parenchyma in any of the mice (data not shown).
Cognitive efficacy of MER5101 in APPswe/PS1ΔE9 Tg mice was determined by CFC and MWM. In the CFC task (Figure 7A), Wt mice treated with either MER5101 (solid circles) or vehicle (open circles) showed similar freezing frequency (7.8±1.2 and 8.2±1.1, respectively). Vehicle-treated APPswe/PS1ΔE9 mice (open triangles) exhibited a modest reduction in the number of freezings (6.2±1.2; p=0.16), implying a potentially weak recall for the stimulation by environmental conditions. However, MER5101 vaccination led to a significant increases in freezing frequency (9.8±2.6; p=0.023) in APPswe/PS1ΔE9 Tg mice (solid triangles).
Interestingly, we have found that Aβ deposition in hippocampus was inversely correlated (Linear Regression p=0.0084, r2=0.68, strong trend) with the number of freezings in the CFC in the MER5101 vaccinated Tg mice, but not in the vehicle control Tg mice. Although there were no overall significant correlations between anti-Aβ antibody levels, Aβ deposition in hippocampus or CFC freezing, within the vaccinated Tg group, the two mice with the highest number of freezings, also had the highest anti-Aβ antibody levels. In addition, the hippocampal Aβ plaque burden for each of these 2 mice (ranked lowest and third lowest in the vaccinated Tg mice) was lower than that of any of the vehicle control Tg mice. Similarly, within the vehicle treated Tg group, the mouse who had the highest number of freezings, also had the lowest hippocampal Aβ level. Thus, while there are no significant correlations, there are some trends for correlations between CFC task, plaque level and antibody levels.
In the MWM task, a 5-day hidden platform trials followed by 2 probe trials which were conducted 2h and 24h, respectively, after the last hidden platform trial on day 5 were performed to evaluate both spatial learning acquisition and spatial memory in APPswe/PS1ΔE9 Tg and Wt mice. Overall, vehicle-treated APPswe/PS1ΔE9 Tg mice showed slower learning acquisition during days 3-5 compared to vehicle-treated Wt mice, however, MER5101 vaccinated APPswe/PS1ΔE9 Tg mice displayed improved learning acquisition during days 2-5 versus adjuvant-treated Tg controls, comparable to the performance of age-matched Wt mice (Figure. 7B).
The results of the first probe trial (2h post the last hidden platform trial on day 5) indicated that all groups of mice retained short-term memory for the platform location and preferentially dwelled in the target quadrant (Figure 7C, D). In the second probe trial (24h post the last hidden platform trial on day 5), immunized and vehicle-treated Wt mice (36.8±4.8% and 32.9±4.2%, respectively) as well as MER5101 vaccinated APPswe/PS1ΔE9 Tg mice (34.7±7.9%), still showed preferential memory for the target quadrant and spent comparable time on platform exploration; however, vehicle-treated APPswe/PS1ΔE9 Tg mice (24.1±6.1%) showed no preference for target quadrant (Figure 7E). Consistently, the vehicle-treated APPswe/PS1ΔE9 Tg mice showed a negative ACI, indicating that they continued to search for the platform in a all quadrants of the pool, whereas, the MER5101 vaccinated APPswe/PS1ΔE9 Tg mice showed a positive ACI, similar to the ACI of Wt mice, suggesting that 4 months of immunization with Aβ1-15:DT/MAS-1 improved spatial memory deficits in APPswe/PS1ΔE9 Tg mice. The 2 groups of APPswe/PS1ΔE9 Tg mice (vehicle and vaccinated) were not significantly different from each other in the second probe trial, possibly due to small groups sizes and high variability, however, a strong trend for improved spatial memory was observed in the vaccinated Tg mice.
The objective of this study was to test a novel Aβ B-cell epitope immunogen conjugate (Aβ1-15:DT) with a Th2-biased adjuvant (MAS-1), MER5101, in Wt and APPswe/PS1ΔE9 Tg mice for its immunopotency and its subsequent effects on Aβ-related pathologies and cognition. Previously, we reported that Aβ1-15 alone was not an effective primary immunogen (Leverone et al., 2003), but priming by injection of full-length Aβ followed by intranasal boosting with either Aβ1-15 peptide or dendrimeric Aβ1-15 (16 copies of Aβ1-15 on a lysine tree; dAβ1-15) formulated in a mucosal adjuvant, E. coli LT(R192G), resulted in high anti-Aβ antibody levels and reduced cerebral amyloid plaque burden in J20 APP Tg mice (Leverone et al., 2003; Seabrook et al., 2006). Intranasal delivery of dAβ1-15/LT(R192G) without a priming injection also induced robust anti-Aβ levels and lowered cerebral Aβ levels and plaques in these mice while avoiding an Aβ-specific cellular immune response (Seabrook et al., 2007). Intranasal administration of 2×Aβ1-15 (tandem repeat of Aβ1-15 linked by 2 lysine residues) had similar effects and improved cognition in the MWM (Maier et al., 2006).
In the present study, we coupled Aβ1-15 to the immunogenic carrier protein, diphtheria toxoid (DT), via a seven amino acid spacer, to provide T cell help for Ig production and immunological memory. DT has long been approved for use in childhood and adult vaccines and is available as a GMP compliant component (Audibert et al., 1982; Wenger et al., 1991; Lambert et al., 1998). The adjuvant utilized in this study, MAS-1, is a 30:70 water-in-oil (w/o) nanoparticular emulsion-based delivery system with a median globule size of 300 nm. MAS-1-based self-antigen conjugated vaccines have been reported to be safe and well tolerated, uniquely producing robust Th2 type neutralizing antibody responses while avoiding Th1 cell mediated cytotoxicity or breaking immune self tolerance of the native antigen [Gilliam et al. 2004; Simms et al. 2000].
We first tested MER5101 in young DBA2 (6 wk-old) and aged C57BL/6 (12 mo-old) mice. After 4 immunizations, significant Aβ antibody levels were observed in both mouse strains (unpublished data). Importantly, the vaccine induced predominantly Th2 antibody isotypes against the N-terminal region of Aβ1-42. We then tested MER5101 vaccination in APPswe/PS1ΔE9 Tg mice and found that it was effective in stimulating a strong humoral immune response with Th2-biased antibody isotypes in the absence an Aβ-specific cellular immune response, T cell infiltration, or microhemorrhage in the brain.
Notably, plasma antibody levels decreased sharply after Day 56 in immunized APPswe/PS1ΔE9 Tg mice, which may have been due to tolerance of the immunogen or by the antibodies binding with circulating Aβ in plasma. The reduction of antibody levels in plasma of vaccinated APPswe/PS1ΔE9 Tg mice at Days 96 and 121 was restored after low-pH dissociation treatment (Li et al., 2007), whereas the antibody levels in immunized Wt mice or adjuvant control mice were not affected, implying that peripheral anti-Aβ antibody levels were maintained and suggesting that a portion of them were bound to Aβ in plasma.
The potent immune response induced by MER5101 effectively lowered cerebral AD pathologies in APPswe/PS1ΔE9 Tg mice. High cerebral levels of Aβ42 have been reported to be associated with plaque load and cognitive status in AD patients (Mann et al., 1996; Cataldo et al., 2004; Cosentino et al., 2010) and AD-like Tg mice (Radde et al., 2006; Murphy et al., 2007). Aβx-42 is more hydrophobic and aggregates more readily than Aβx-40. The ratio of Aβ42/40 is increased in AD patients, and soluble Aβ42 oligomers are considered extremely toxic to neurons, thereby contributing to cognitive decline in AD (Lambert et al., 1998; Fukuda et al., 1999). In this study, Aβ plaques were attenuated in immunized APPswe/PS1ΔE9 Tg mice, which correlated with reduced insoluble Aβ extracted from brain, similar to the results of our previous studies (Maier et al., 2006; Seabrook et al., 2007). MER5101 lowered mainly insoluble Aβx-42 and both soluble/insoluble Aβ1-total but had no effect on vascular amyloid. Fibrillar Aβ plaques have been shown to correlate with microglial activation in many AD-like Tg mouse models (Frautschy et al., 1998; Gordon et al., 2002; Kitazawa et al., 2005). Here, MER5101 immunized Tg mice had fewer Iba-1- and CD45-immunopositive microglia in HC compared to Tg vehicle controls, perhaps due to the lower plaque burden. MER5101 also lowered Aβ in the TBS-soluble brain fraction, which presumably contains Aβ oligomers with toxicities that are thought to affect neuron viability and lead to cognition decline (Shankar et al., 2008; Tomic et al., 2009). MER5101 vaccinated APPswe/PS1ΔE9 Tg mice had higher synaptic density and fewer neuritic plaques, which together, may have contributed to their superior cognitive performance compared to the vehicle-treated Tg mice.
The exact mechanism by which MER5101 vaccination mediates cerebral Aβ clearance is not clear yet, however, different mechanisms of Aβ removal by antibodies have been proposed. These include (1) Fc receptor (FcR) dependent (Bard et al., 2000; Bard et al., 2003; Nicoll et al., 2003; Ferrer et al., 2004) or independent (Bacskai et al., 2002; Das et al., 2003; Tamura et al., 2005) microglia-mediated phagocytosis, (2) antibody-mediated Aβ disaggregation and neutralization of Aβ toxicity (Solomon et al., 1997; Frenkel et al., 2000; Bacskai et al., 2001; Oddo et al., 2006), (3) a shift in efflux of Aβ from brain to periphery known as “the peripheral sink hypothesis” (DeMattos et al., 2001; Sigurdsson et al., 2001; Sigurdsson et al., 2002; Lemere et al., 2003), and/or (4) intracerebral sequestration of monomeric Aβ (Yamada et al., 2009). In our prior immunization studies (Leverone et al., 2003; Maier et al., 2006), we observed a simultaneous lowering of cerebral soluble Aβ and an increase in plasma Aβ levels in vaccinated mice that generated high anti-Aβ antibody levels, suggesting a potential antibody-driven shift of soluble Aβ from the brain to the periphery and leading to enhanced catabolism of Aβ, consistent with the “peripheral sink hypothesis” (DeMattos et al., 2001). In the present study, while cerebral soluble Aβ levels were reduced in brain homogenates of MER5101 vaccinated APPswe/PS1ΔE9 Tg mice, there was no change in plasma Aβ levels before or after low-pH dissociation of the antigen/antibody complexes compared to Tg vehicle controls. Plasma Aβ was undetectable post-dissociation due to dilution in the dissociation buffer. Nevertheless, the dissociation treatment resulted in an overt unmasking of anti-Aβ antibody levels in plasma samples collected from MER5101 vaccinated Tg mice at the end of treatment, suggesting that some portion of plasma Aβ may have been bound by antibodies prior to dissociation by low pH. Thus, the “peripheral sink hypothesis” may be one of the mechanisms for Aβ clearance in this study, although other aforementioned mechanisms may have contributed, too.
Clearance of fibrillar Aβ may require antibodies to across the blood brain barrier (BBB), bind with cerebral Aβ, and recruit activated microglia for phagocytosis through an FcR-dependent (Bard et al., 2000; Bard et al., 2003; Nicoll et al., 2003; Ferrer et al., 2004) or -independent manner (Bacskai et al., 2002; Das et al., 2003; Tamura et al., 2005). In this study, the levels of insoluble Aβ (guanidine-soluble fraction) in brain homogenates as well as fibrillar Aβ plaques (Thio S labeled) were significantly lower in APPswe/PS1ΔE9 Tg mice after MER5101 vaccination, suggesting that in addition to the peripheral sink, the anti-Aβ antibodies may have directly interacted with compacted Aβ assemblies in brain and induced phagocytosis by microglia. Although microglial activation was not enhanced but attenuated in immunized Tg mouse brains at the end of this study, it is possible that FcR-mediated Aβ clearance may have occurred early and had subsided by the end of the study due to the removal of plaques. Interestingly, activated microglia surrounded the remaining plaques, suggesting an ongoing phagocytotic process
Taken together, our data demonstrate that MER5101, a novel conjugate vaccine composed of the B-cell epitope, Aβ1-15, and an immunological carrier protein, DT, in a Th2-type adjuvant, MAS-1, is sufficient for eliciting a robust humoral immune response, resulting in predominantly Th2-biased immunoglobulins without Aβ-specific T cell reactivity in APPswe/PS1ΔE9 Tg mice, and is effective in lowering Aβ-correlated pathologies and improving cognitive deficits. Prior results using the MER platform in roughly 1,500 patients for vaccines targeting self hormones gastrin-17 (G17) and gonadotrophin-releasing hormone (GnRH) in the elderly and immunocompromised individuals elicited robust antibody responses specific to the self hormones. Systemic toxicities were not reported, nor were toxicities in organs producing the targeted hormones or bearing receptors for these hormones, including gastrointestinal organs in patients immunized against gastrin 17 and hypothalamus/pituitary tissues in patients immunized against gonadotrophin-releasing hormone, suggesting that Th1 type cell-mediated toxicities were avoided. These observations with vaccines comprising self peptide-to-DT conjugates formulated with MAS-1 adjuvant, along with the responses to MER5101 described here, suggest that MER5101 may be safe and effective in generating antibody titers in the elderly and lowering cerebral Aβ deposition prior to cognitive changes or in early stages of AD. We believe that our novel Aβ immunogen conjugate together with the Th2-type adjuvant, MAS-1, is a promising candidate for a second-generation vaccine leading to an effective immunotherapy for AD while avoiding a potentially deleterious pro-inflammatory cellular immune response.
Funding provided by NIH RO1 AG20159 (and ARRA Supplement) to CAL. Peter Blackburn is the President and Founder and Stephen Grimes is a Principal Scientist at Mercia Pharma, Inc.
Conflict of Interest: All other authors declare no competing interests.