|Home | About | Journals | Submit | Contact Us | Français|
Objective: To demonstrate that in APPswe/PS1DE9 transgenic mice, gene gun mediated Aβ42 gene vaccination elicits a high titer of anti-Aβ42 antibodies causal of a significant reduction of Aβ42 deposition in brain.
Methods: Gene gun immunization is conducted with transgenic mice using the Aβ42 gene in a bacterial plasmid with the pSP72-E3L-Aβ42 construct. Enzyme-linked immunoabsorbent assays (ELISA) and Western blots are used to monitor anti-Aβ42 antibody levels in serum and Aβ42 levels in brain tissues. Enzyme-linked immunospot (ELISPOT) assays are used for detection of peripheral blood T cells to release γ-interferon. Immunofluorescence detection of Aβ42 plaques and quantification of amyloid burden of brain tissue were measured and sections were analyzed with Image J (NIH) software.
Results: Gene gun vaccination with the Aβ42 gene resulted in high titers of anti-Aβ42 antibody production of the Th2-type. Levels of Aβ42 in treated transgenic mouse brain were reduced by 60-77.5%. The Mann-Whitney U-test P = 0.0286.
Interpretation: We have developed a gene gun mediated Aβ42 gene vaccination method that is efficient to break host Aβ42 tolerance without using adjuvant and induces a Th2 immune response. Aβ42 gene vaccination significantly reduces the Aβ42 burden of the brain in treated APPswe/PS1DE9 transgenic mice with no overlap between treated and control mice.
Alzheimer’s disease (AD) is the most common cause of dementia and currently there is no cure or effective treatment to prevent its progression. The cause of AD has been associated with the accumulation, aggregation, and deposition of amyloid beta peptides (Ah) in cerebral cortex, hippocampus, and other subcortical structures [1,2]. Aβ is derived proteolytically from a glycosylated membrane protein, beta-amyloid precursor protein [3,4] (APP). APP is ubiquitously expressed but at higher levels in the central nervous system [5,6]. The derived Ah in brain is 39 to 43 amino acids in length but the major processed species are Aβ40 and Aβ42. The aggregated form of Aβ42 has been identified to play a major role in the initiation of neuropathology of AD [7-9].
In recent years, reduction in the level of Aβ in brain has become a major therapeutic goal in prevention and treatment of AD. In transgenic mice carrying mutations causal of AD, reduction of the amyloid burden in brain and cognitive improvement have been achieved by active immunization of these mice with Aβ42 peptide or by infusion of anti-Ah antibodies [10-14]. A clinical trial of patients with Alzheimer’s disease was conducted with active Aβ42 peptide immunization but the trial was halted due to the complications associated with a cell-mediated cytotoxic T cell mediated meningoencephalitis in 6% of immunized AD patients [15,16]. Neuropsychological testing of immunized patients showed some slowing of cognitive loss in a subset of patients with significant anti-Ah antibody production who subsequently had Ah plaque reduction and clearance on postmortem examination [17-21].
In the present paper, we demonstrate that in APPswe/PS1ΔE9 transgenic mice, gene gun mediated Aβ42 gene vaccination efficiently elicits a high titer of anti-Aβ42 antibodies and the Aβ42 deposition in the brains of these mice is significantly reduced without a significant adverse cellular immune response. We believe gene vaccination can be an approach to reduce Aβ levels in the brain of patients with Alzheimer’s disease.
APP-PS1 co-injected transgenic mice (7-10weeks) carrying the human APP-Swedish and PS1DE9 mutations were purchased from Jackson Laboratory (Bar Harbor, ME) (Stock Number 004462). These animals develop numerous amyloid plaque deposits in the cerebral cortex and hippocampus by 6 months of age. The wild type female Balb/c mice (age 5-7week) were obtained from Harlan Inc. (Indianapolis, IN). The use of animals for this study was approved by the UT Southwestern Medical Center Animal Care and Use Committee.
The open reading frame of the human Aβ42 and Aβ16 genes were chemically synthesized with optimal codons for mammalian cell expression and cloned into the genetic immunization vector as described previously . The Aβ42 and Aβ16 genes were fused with the leader sequence of adenovirus E3 gene . A lysosome targeting sequence was fused downstream of the Aβ42 and Aβ16 sequences . For comparison, we also replaced the E3 leader with the human alpha-1 antitrypsin leader sequence, or insert between the leader and Aβ42 or Aβ16 , a B-cell targeted CTA1-DD gene  (ADP-ribosylation domain of cholera toxin with an Ig-binding fragment D of Staphylococcus aureus protein A) or the rat cartilage oligomerization matrix protein (COMP) gene . The artificial SP72 and CMV promoters were also compared [22,26].
Immunizations of mice with plasmid DNA encoding the Aβ42 gene were performed on mouse ear skin using the Helios gene gun (Bio-Rad, Hercules, CA) as described . Briefly, DNA-coated gold particles were prepared by binding DNA to the gold beads (1-2Am, Ferro Electronic Material Systems, South Plainfield, NJ) in 1Ag DNA/1mg gold beads in the presence of spermidine and calcium chloride. The DNA-gold was attached to the insides of the tubing and cut into 1.3cm long as a bullet. The DNA-coated gold particles were bombarded to both sides of the mouse ears using the gene gun with a helium gas of 400psi. The APPswe-PS1ΔE9 mice (N = 4) were immunized with the Aβ42 gene starting from the age of 3months with first three immunizations at 2-week intervals and then once per month for a total of 11 immunizations. The control group (N = 4) received either a non-Ah peptide vector (2) or Aβ16 peptide vector (2) with the same immunization schedule. The blood was drawn from the tail vein and the serum sample was used to monitor the humoral immune response.
Enzyme-linked immunoabsorbent assay (ELISA) and Western blot were used to monitor the humoral immune responses . In brief, mouse blood was collected from tail vein and serum was used to detect Aβ peptide by ELISA with 96-microwell plate coated with GST-Ah proteins. For Western blot, the GST-Ah proteins in bacteria extract were separated by SDS-PAGE, blotted onto nitrocellulose, incubated with the sera at 1:2000 dilutions. Antibodies against Aβ were detected using peroxidase-conjugated affinity-purified rabbit antiserum against mouse Ig. To determine the specific isotypes, sera from mice were diluted 1:200 and tested with ELISA method as described above. To detect mouse IgG1, IgG2a, we used anti-mouse Ig-subclass-specific rabbit antibody (Pierce, Rockford, IL), followed by incubation with HRP-conjugated Donkey anti-rabbit IgG.
The cell-mediated immune response was evaluated by enzyme-linked immunospot assays (ELISPOT assay) for detection of peripheral blood T cell to release interferon-gamma during in vitro re-stimulation with Aβ peptide . Briefly, 96-well polyvinylidine diflouride (PVDF) plates (Millipore, Bedford, MA) were coated with antibody to IFNs gamma. Cells were cultured at 2 × 105 per well in 0.2ml of medium for re-stimulation with Aβ peptides. After 36h of incubation at 37 °C, the plates were washed, incubated with biotinylated anti-mouse IFN-gamma, and further with Streptavidin-AP conjugate. After three washes, spots were developed with one-step NBT/BCIP reagent. Spots were counted using a stereomicroscope.
The frontal lobe of brain was homogenized in 10volumes of guanidine-tris buffer (5.0M guanidine HCl/50mM Tris-HCl, pH 8.0). The homogenates were mixed for 3 to 4h at RT and stored at 20°C until measured [28,29]. For ELISA assay, 96-well plates (Nunc) were coated with polyclonal antibody (50μl, 1:400 dilutions) that specifically recognize the C-terminal of Aβ42 (Sigma) at 4°C overnight. The plate was further blocked by adding 100μl blocking buffer (PBS with 1% milk, 0.05% Tween 20) for 2h in RT. After washing with washing buffer (PBS containing 0.05% Tween 20), the brain extracts in guanidine buffer (4Ag protein perμl) or the plasma were diluted in 1:100 in blocking buffer and added to the well. The plate was incubated overnight in 4°C. After washing, 50μl of monoclonal anti-Aβ1-17 antibody (Sigma, 1:1000 in blocking buffer) was added to the well to detect the bound Aβ42. After 2-h incubation in RT, 50μl of peroxidase-labeled anti-mouse immunoglobulin antibody (dilution of 1:1000) was added to each well. Following 2-h incubation and washing, 50μl of substrate solution (TMB, Calbiochem, La Jolla, CA) was added and terminated after 20min RT incubation with 50μl 0.5M HCl. The different concentrations (1-2000ng/ml) of synthetic Aβ42 peptide were used to establish the standard curve. The absorbance of the plates was read at 450nm with a spectrophotometer.
Portions of the brain from APPswe/PS1ΔE9 transgenic mice were dissected and put into a tube filled with PBS and the tube was snap frozen in liquid nitrogen. Part of the brain was also fixed with paraformaldehyde and embedded in paraffin. The frozen tissues were subsequently cut into 8-Amsections with a cryotome, collected on glass slides, dried and fixed for 4h in 4% paraformaldehyde. After washing with PBS, the section was penetrated with 0.2% Triton X-100 for 10min and further washed and blocked with blocking buffer (1% BSA in PBS with 0.05% Tween 20) for 30min. After incubation with rabbit polyclonal anti-serum  (A1916, Sigma, St. Louis, MO) against Aβ42(5Ag/ml, 2h at RT), the sections were washed three times for 5min each in PBS, again treated with blocking buffer (5min, RT) before reacting with the secondary antibody (Alexa488-labeled goat anti-rabbit IgG, (10Ag/ml, 1h, RT) (Invitrogen Co.; Carlsbad, CA). Finally, the preparations were washed three times in PBS and observed by confocal fluorescence microscopy. The amyloid plaque burden was analyzed with Image J (NIH) software. Five representative sections of each mouse brain were imaged and the areas and densities of the plaques were measured by the software. Sections from paraffin embedded tissue (8Am) were deparaffinized and immunolabeled as above. In addition, formic acid antigen retrieval was performed on sections of frontal cortex from one control and one treated mouse and stained as above.
Because the assumptions of the two independent samples t-test were violated (unequal group variances and non-normal distributions), quantitative measurement of Aβ42 burden in frozen and paraffin-embedded sections for control and treatment groups were compared using the non-parametric Mann-Whitney U-test. Analyses were performed using SAS and statistical testing was conducted using p < 0.05.
The Aβ42 or Aβ16 genes were cloned into a genetic immunization plasmid vector under the control of an synthetic mammalian-cell specific promoter named SP72 and fused upstream with a leader signal of adenovirus E3 gene (E3L) (pSP72-E3L-Aβ42 or Aβ16 ) or alpha1 antitrypsin leader signal (ATL) (pSP72-ATL-Aβ42). SP72 promoter was replaced with cytomegalovirus-immediate early enhancer/promoter (CMV) to create pCMV-ATL-Aβ42, or insert COMP or CTA1-DD gene between ATL and Aβ42 gene to create (pSP72-ATL-COM-Aβ42, SP72-ATL-CTA1-Aβ42). All the constructs have the same downstream endosome targeting sequence fused in frame with Aβ42 or Aβ16 . The constructs were sequenced to confirm the insert and the correct open reading frame. We immunized initially the wild type Balb/c mice with these plasmid constructs by gene gun to compare the efficiency to incite the Aβ42 immune response. After two gene gun immunizations to the mice ears (each immunization injecting 4Ag DNA in 4 bullets in both sides of the ears) at 2-week intervals, tail vein blood was taken 2weeks after the second immunization and ELISA was performed to measure the anti-Ah antibodies in sera. We identified that the pSP72-E3L-Ah42 construct showed the best antibody response. The fusion of the adjuvant sequence COM, or CTA1-DD did not enhance the humoral immune responses against Aβ peptide. SP72 promoter is slightly better than CMV and the E3 leader is better than the AAT leader sequence (Fig. 1A). The result prompted us to apply the SP72-E3L-Aβ42 plasmid for gene immunization in APPswe/PS1ΔE9 transgenic mice without using any other immune adjuvant. The encoded protein of E3 leader, human Aβ42 and endosomal targeting sequence are shown in Fig. 1B and the finally applied plasmid vector construct is shown in Fig. 1C.
On the basis of data obtained in wild-type Balb/c mice, we used the pSP72-E3L-Aβ42 construct for gene immunization in APP/PS1 co-injected double transgenic mice APPswe/PS1ΔE9. These mice begin to develop amyloid plaques by the age of 6months. Four treated mice received the human Aβ42 gene vaccine at the age of 3months and four control mice received DNA plasmid without the Aβ peptide gene (2) or with the Aβ16 gene (2). The humoral response was detectable by ELISA 2weeks after the third immunization and reached a peak level after six immunizations. The antibody titer against Aβ1-16 was estimated at 1:10,000 and 1:4000 for Aβ17-28 and Aβ29-42 peptides in all four treated Tg mice. There are no detectable specific anti-Ah antibodies in the sera of vector only control Tg mice. Results of ELISA and Western blots (Fig. 2A and B) showed similar conclusions. Fig. 2A shows the antibody concentrations of the sera of Aβ42 treated mice obtained 2weeks after the last immunization with about 1.2Ag/ml specific anti-Aβ1-16 antibodies and 350ng/ml of anti-Aβ17-28 and Aβ29-42.
We also analyzed isotypes of anti-Aβ42 antibodies in the sera of the Aβ42 treated Tg mice. All treated Tg mice generated IgG1 antibodies, whereas the level of IgG2a was in an undetectable level (Fig. 2C). The production of IgG1 antibodies is an indirect measure of the relative contribution of Th2-type cytokines, whereas IgG2a antibodies reflect the contribution of Th1 cytokines to the immune response. Thus, our data indicate that gene gun mediated Aβ42 gene vaccine gun induces a highly Th2-polarized response. Consistent with the Th2 response in antibody isotyping, these Aβ42 treated mice also showed insignificant cellular immune responses as tested by ELISPOT stimulated with Aβ1-42 and Aβ9-18 synthetic peptide although a slightly higher number of interferon positive T cells was observed (Fig. 2D).
After 11 Aβ42 gene vaccinations, the Aβ ELISA analysis was performed to measure the brain and blood Aβ42 levels in 15-month-old Tg mice. The Aβ of forebrain tissue was extracted with 5M guanidine-tris buffer and subjected to sandwich ELISA to measure the specific Aβ42. The median level of Aβ42 in the cerebral cortex of the four control mice at 15months was 4500ng/g wet tissue. In contrast, Aβ42 gene vaccinated animals had 66% less Aβ42 at 15months (1500 ng/g) than the control vector immunized group (p < 0.05) (Fig. 3A). The heparinized plasma was directly subjected to Aβ42 measurement without further extraction and showed a 33% increase of plasma Aβ42 in the vaccinated mice (160ng/ml) than the four control mice (120ng/ml), though the differences were not statistically different (Fig. 3B).
To evaluate the Aβ burden in vaccinated Tg mouse brain, we performed fluorescence immunolabeling for Aβ42 on sections from both snap frozen tissue and paraformaldehyde fixed and paraffin embedded tissue. All eight mice (treated N = 4, and control N = 4) were subjected to immunolabeling analysis for Aβ42 with sigma antibody A1916. Preparations from all four control mice showed numerous large volume plaques in cortical and hippocampal regions. In contrast, while plaques remained in preparations from all four vaccinated mice, they were smaller and fewer in number. Representative photomicroimages of Aβ42 labelled sections from control and treated mice are shown in frozen sections of frontal cortex (Fig. 4) and in paraformaldehyde fixed paraffin-embedded sections from hippocampus (Fig. 5). To quantitatively measure the brain Aβ42 burden in these mice, 10 representative high-resolution pictures were taken with confocal microscopy from cortical and hippocampal regions of each mouse (5 in frozen sections and 5 in paraffin-embedded sections). The fluorescence plaque areas of each microimage were traced by Image J (NIH) software and the plaque areas and mean plaque intensity were registered by the software. The plaque intensity was defined as 0 in black and 255 in white and the fluorescence intensity was registered as a gray signal in between. The fluorescence plaque areas were registered in units of pixels, a small dot defined by the software and computer, in a given resolution. The density (total amount) was given by the measured plaque area in pixels multiplied by the average plaque intensity. With this method, we quantitatively evaluated the plaque burden in brain and calculated the reduction of Aβ42 in the treated mice compared to the untreated control mice. The measured values are given in Tables Tables11 and and2.2. Table 1 shows the quantitative result of frozen sections and Table 2 of paraffin-embedded sections. With frozen sectioning methods, the median plaque area of four control mice with five imaged areas being measured in each mouse was 177,100 pixels and 31,500 pixels in the four treated mice. The plaque area was reduced by 82.2% in the treated compared to the control mice. Similarly, the median density of four control mice (25.3 million pixels) was also significantly reduced in the treated mice (5.7 million pixels) with a Mann-Whitney P = 0.0286. There was a 77.5% reduction in plaque burden in treated mice as calculated by plaque density (total amount) (Table 1). We used the same method to measure the Aβ42 burden reduction in the vaccinated mice brains in paraffin-embedded sections. The median plaque volume in four control mice was 201,600 pixels and 69,100 pixels in four vaccinated mice with a 65.7% reduction in plaque area in the treated mice. The median plaque density (total amount) of the control mice was 35.3 million pixels and 14.2 million pixels in the treated mice. The reduction of the plaque burden in the treated mice was 60% as calculated by the plaque density( total amount) with a Mann-Whitney P = 0.0286 (Table 2). In conclusion, with Image J software quantitation of Aβ42 burden in a total of 80 representative confocal imaged areas (40 in control and 40 in treated mice), the reduction of Aβ42 burden was 60-77.5% in the gene vaccinated group of mice compared to the control group of mice. Both tissue processing methods give the same conclusion and there is no overlap of the data between the control and treated groups both for frozen and paraffin-embedded brain for Aβ42 levels. Sections from one control and one treated mouse which included formic antigen retrieval methodology were studied with Image J software. Formic acid antigen retrieval is commonly employed in the diagnostic evaluation of human AD brain tissue for Aβ levels. The median plaque density (total amount) of the vector only control mouse (C1) frontal cortex was 36.2 million pixels and the median plaque density (total amount) of the treated mouse (T2) frontal cortex was 21.6 million pixels, giving a 41% reduction in Aβ42 immunolabelling. This additional evaluation is included to assure quantitation of all potential amyloid antigenicity and demonstration of significant reduction in Aβ42 immunolabelling in the treated mouse brain. In addition, the immune response induced by gene vaccination to Aβ42 does not produce any obvious signs of damage to the neurons in Aβ42 gene-immunized animals. Histological examination of brain revealed no signs of immune-mediated complications (data not shown). The weight gain is similar in both groups with an average 40g body weight at 15months of age.
APPswe/PS1ΔE9 double transgenic mice develop Aβ42 plaques beginning at the age of 6months, much earlier than APPswe/PS1(A246E) (9months) and PDAPP (APPswe) Tg mice [30-32] (18months). We demonstrate, even in this highly Aβ expressing strain, significant attenuation of the Aβ42 plaque formation in cortical and hippocampal brain regions in all four mice with Aβ42 gene vaccination. To our knowledge, our findings represent the first report of gene vaccination that significantly reduced the Aβ42 plaque burden in a co-injected double transgenic animal AD model, although we and others have reported the Aβ42 gene vaccine can elicit an immune response to Aβ42[22,33,34].
Gene vaccination using a plasmid vector has many advantages over peptide or protein vaccination including simple, low-cost preparation and purification [35-37]. Furthermore with gene vaccination, the immune response of the host can be easily manipulated to bias towards a Th1 or Th2 type reaction. It has been found that muscle injection of gene vaccines generally induces a Th1 predominant reaction. In contrast, gene gun-based gene vaccines can induce predominantly Th2 responses with certain antigens [38,39]. In our study, gene gun-administered Aβ42 gene vaccination induces a Th2 response without a detectable Th1 type reaction and the cellular response to Aβ42 is insignificant. This point may be a key advantage for gene vaccination, as it could avoid cell-mediated encephalitis caused by Aβ42 peptide immunization. Recent reports on clinical trials with gene vaccines are encouraging that this method can be effective in humans .
In the present study, we show that the construct SP72-E3-Aβ42 elicits an excellent immune response. The SP72 promoter was rationally designed and it is a part of our effort to create a better immunization vector to improve the gene vaccination method. The E3 leader may help to target Aβ42 peptide to the endoplasmic reticulum to bind MHC molecules for presentation. The Aβ16 containing vector served as a positive control as it did not reduce Aβ42 plaque density compared to vector only vaccinated mice. Antibody against the entire Aβ42 peptide may be required for subsequent Aβ42 targeting and reduction from brain.
In summary, we have developed a gene gun-mediated Aβ42 gene vaccination method that is efficient to break host Aβ42 tolerance without using adjuvant and induces a Th2 immune response that minimizes producing a T cell mediated meningoencephalitis. In the present studies, we demonstrate further that Aβ42 gene vaccination can significantly reduce the Aβ42 burden of the brain in all treated APPswe/PS1ΔE9 transgenic mice with no overlap between treated and control mice. Collectively, the results of this study support the position that Aβ42 gene immunization should now be studied in a non-human primate to see if is safe and effective and then consider it in a clinical trial of patients with Alzheimer’s disease .
This study was supported by P30AG12300 Alzheimer’s Disease Center grant from the National Institutes of Health, National Institute on Aging, Bethesda, MD; U01 AG16976 the National Alzheimer’s Coordinating Center grant, Seattle, WA; and The Rudman Foundation, Dallas, Texas; The Luttrell foundation, Belmont, California and unrestricted funds to SAJ in the Center for Biomedical Inventions. The Winspear Family Special Center for Research on the Neuropathology of Alzheimer Disease, Dallas, provided support for this research. The University of Texas Southwestern Pathology and Immunohistochemistry Laboratory, Dallas, performed the histological processing and immunohistochemical stains. We thank Dr. Weiping Li for help with confocal microscopy. We thank Dr. Thomas Kodadek and Dr. Ramon Diaz-Arrastia for their helpful suggestions.