Abstract
With the objective to improve the amyloid-β (Aβ) targeting immunotherapy, we investigated the safety and efficacy of an oral vaccine with recombinant adeno-associated virus vector carrying a signal sequence and Aβ1-43 cDNA (rAAV/Aβ) in old non-human primates, 12 African green and 10 cynomolgus monkeys. The enteric-dissolving coated capsules containing rAAV/Aβ were orally administered once or twice, then monkeys’ conditions were carefully observed with complete blood count and laboratory examinations of the sera. General conditions, food intake, water intake, stool conditions, body weight changes, and menstruation cycles were not significantly altered, and laboratory tests and pathological examinations of the systemic organs were unremarkable. Pathological examinations of the brain showed significant reduction of the amyloid plaque burden and intracellular Aβ without inflammatory or hemorrhagic changes in the brain. However, soluble Aβ and some Aβ oligomers were increased in rAAV-treated monkey brains without changes of the neuronal density and vascular amyloidosis. Thus, this vaccine seems to be safe in general, but we must be cautious about the increase of Aβ oligomers after vaccination. This vaccine may be recommended at a very early stage of Alzheimer’s disease when little amyloid is deposited.
INTRODUCTION
Immune-mediated clearance of aggregated amyloid-β (Aβ) is suggested to be one of promising therapeutic strategies for Alzheimer’s disease (AD) [1, 2], because brains of AD patients who had received active immunization with aggregated Aβ1-42 peptide and adjuvant showed successful clearance of amyloid deposits [3], although some patients developed aseptic meningoencephalitis [4]. The meningoencephalitis is supposed to be due to autoimmune mechanisms against Aβ [5, 6] and mediated by T helper type 1 (Th1) T cells [7]. In order to avoid the autoimmune meningoencephalitis, monoclonal antibodies to Aβ and new active immunization strategies targeting Aβ have been applied in patients and animals [8]. A recent result of the phase III clinical trial of a monoclonal antibody solanezmab showed a slight but significant efficacy in mild AD patients, although the primary endpoint was negative [9], suggesting that the Aβ-targeting therapeutic strategy seems to be promising in the early stage or the preclinical stage of AD.
We have developed oral vaccine using recombinant adeno-associated virus vector expressing a secreted form of Aβ1-43 (rAAV/Aβ) and demonstrated significant amelioration of amyloid burden in the brain and improvement of cognitive functions with a significant reduction of Aβ oligomers in amyloid-β protein precursor (APP) transgenic mice (Tg2576) [10, 11]. This strategy has a big advantage in utilizing the gut immune system, because the gut-associated lymphoid tissue is the largest immune system in the body, thus it is efficient even in the elderly, and it induces oral tolerance through induction of Th2 and/or Th3 T cells [12]. Although our previous studies in mice showed its safety, it is necessary to test the vaccine in non-human primates before we test it in humans, because the T cell response to Aβ is suppressed in Tg2576 mice with the B6 background [13]. Since the vaccine is given orally, side effects if any, particularly of the gastrointestinal tract should be carefully observed. Here we show efficient reduction of amyloid burden including intracellular Aβ in aged non-human primates without any significant adverse effects, although it increased soluble Aβ and some Aβ oligomers.
MATERIALS AND METHODS
Mice
To confirm infection and expression of rAAV, we used adult C57BL/6 mice (n = 8).
Monkeys
Nine 16- to 23-year-old captive African green monkeys (Chlorocebus aethiops, mean age = 19 years and 1 month old), three of them originating from their wildlife habitat that looked old, and ten 19- to 24-year-old captive cynomolgus monkeys (Macaca fascicularis, mean age = 22.0 years old) were subjected in this study. In addition, two 5-year-old captive cynomolgus monkeys were used for testing doses of the vaccine. All monkeys were female.
The African green monkeys were kept at the JapanSLC Farm (Japan SLC Inc., Izu, Japan), and cynomolgus monkeys were kept at the Tsukuba Primate Center (present name: Tsukuba Primate Research Center, National Institute of BiomedicalInnovation, Tsukuba, Japan). Each monkey was kept in a 540 mm (W)×600 mm (D)×690 mm (H) cage at SLC and in a 500 mm (W)×800 mm (D)×800 mm (H) cage at Tsukuba, temperature and humidity were controlled at 24±2°C and 55±15%, and at 25±2°C and 55±5% with 12 times/h ventilation, respectively, and 12 h lighting/day from 7:00 to 19:00, and 14 h from 5:00 to 19:00, respectively. They were fed with 100–200 g/day dry foods (Certified Primate Diet 5048, PMI Feeds, USA) at SLC, and 70 g dry foods (Type AS, Oriental Yeast Co., Tokyo) and 100 g apple once a day at Tsukuba, and water was given ad libitum at both facilities.
Production of recombinant adeno-associated virus vector
Recombinant adeno-associated virus (rAAV) vector for expression of Aβ1-43 was described previously [10, 11]. Briefly, Aβ1-43 cDNA was amplified from human APP695 and an APP signal peptide sequence was linked. The PCR-amplified signal peptide sequence and Aβ1-43 was further ligated to the nonfunctional “stuffer” sequences of pBR322 (PvuII-SalI fragment) to achieve the adequate genome size (4465 bp) for the efficient AAV packaging. For a control, rAAV carrying green fluorescent protein (GFP) cDNA was constructed as described [14].
Human embryonic kidney (HEK) 293 cells were cotransfected with pTRUF2/SP-Aβ1-43, plasmid pXX2 and pXX6 as described [15]. Recombinant AAV titers were in the range of 1×1013 to 2×1013 viral genomes per ml.
Oral administration to monkeys
In our previous study, 5×1011 genome of AAV/Aβ was found safe and efficient in mice [10]. When 1×1013 genome of AAV/Aβ was given orally in 5-year-old cynomolgus monkeys, antibodies to Aβ were elevated 1 month after and no serious side effects were observed. Therefore, this dose of vaccine was given orally to aged monkeys.
rAAV/Aβ1-43 in 1-2 ml PBS was mixed with gelatin powder, put in enteric-dissolving coated capsules (No.3 = 0.32 ml volume, Sunsho Pharmaceutical Co., Fujinomiya, Shizuoka, Japan), and 5–7 capsules were given orally into each monkey. At SLC a monkey sitting on a fixation chair swallowed the capsules without anesthesia, when the capsules were given onto the basal area of the monkey’s tongue. After swallowing, a monkey was given some water. At Tsukuba Primate Center, the capsules were given through gastric tube under light anesthesia with 5–10 mg/kg of Ketamine Hydrochloride® (Daiichi-Sankyo Co., Tokyo) followed by intramuscular injection with 0.05 mg/kg of Atropine Sulfate®. Some monkeys received an additional vaccination at the same dose 3 or 6 months after the first vaccination. Before and after vaccination, all animals were examined with regard to their general condition, food and water consumption, stool condition, menstruation, and body weight. The monkeys were sacrificed by deep anesthesia 3, 6, or 12 months after vaccination and autopsy was done.
Ethical feature and regulations
All animal experiments were approved by the ethical committee of the National Institute for Longevity Sciences, the Japan SLC Farm and the Tsukuba Primate Center. Experiments with recombinant DNA were approved by the Recombinant DNA Experiment Safety Committee at each institute, and information of recombinant DNA was properly given to thecontract.
Analysis of anti-Aβ antibodies
Sera were obtained from African green monkeys 1, 2, 3, 6, and 9 months after vaccination and from Cynomolgus monkeys 1, 2, and 3 months after vaccination, diluted in 1000 x, and were applied onto Aβ1-40 or Aβ1-42-coated (5μg/ml) 96-well plates (Nunc, MaxiSorp) after blocking with 5% non-fat milk/TBS-T buffer. The antibody titers were assayed using peroxidase-labeled goat anti-monkey IgG (goat anti-monkey IgG (H/L):HRP, AbD Serotec).
Since Aβ antibodies were not well assayed by the above method, we also applied the method described by others [16] with some modifications. Briefly, sera were diluted 1:100 with dissociation buffer composed of PBS with 1.5% BSA and 0.2 M glycine-acetate, pH 2.5 to a final volume of 500μl and were incubated for 20 min at 23°C. The sera were then pipetted into the sample reservoir of a centrifugal filter device same as Microcon YM-10 (10,000 MW cut-off, Millipore) and centrifuged at 16,000 g for 20 min at 23°C. The sample reservoir was then separated from the flow through, placed inverted into a second tube and centrifuged at 2,000 g for 3 min at 23°C. The collected solution containing Aβ antibodies including those dissociated from Aβ peptide was adjusted to pH 7.0 with 1 M Tris buffer, pH 9.0. The retentate volume was reconstituted to the initial volume of 500μl with ELISA dilution buffer composed of PBS with 1.5% BSA and 0.1% Tween 20. The collected sera were then added to an ELISA plate at several dilutions to determine antibody titers. As a control, the same serum was treated in an identical process except that the sera were diluted into buffer at pH7.0.
Immunohistochemistry
Cryosections were fixed for 15 min with 70% formic acid for Aβ staining in 0.1 M phosphate buffer and rinsed with PBS-Triton before incubation in 0.3% H2O2 in methanol for 30 min. The tissue was stained with mouse monoclonal anti-Aβ antibody (4G8) at room temperature for 2 h. Sections were washed with PBS-Triton before incubation with secondary goat anti-mouse antibody for 2 h. After washes with PBS-Triton, sections were treated with avidin-biotin HRP/DAB.
Examination of amyloid plaque burden
Coronal sections cut through the frontal lobe, parietal lobe, temporal lobe, occipital lobe, and the cerebellum-brainstem were stained with anti-Aβ antibody 4G8. Two sections from each animal were stained. Plaques were counted in a field of 1.105 mm2 by an investigator (TT) in blind on a microscope; the average of counted fields in each specimen was 75 for cynomolgus and 62 for African green monkeys, and the number of plaques was expressed as No. plaques/1.105 mm2. We also applied automated measuring of the Aβ deposits area [10].
In order to count neuronal densities, paraffin sections were stained with an antibody NeuN (Millipore, Billerica, Ma) and counterstained with hematoxylin, and NeuN-positive cells were counted similarly.
Biochemical analysis of the brain
One gram of the frozen brain from the frontal lobe, amygdala, hippocampus, occipital lobe, and cerebellum was homogenized in 10 ml TBS, passed through a 37μm mesh, centrifuged at 100,000 g, and the supernatant was obtained as the soluble fraction. The precipitate was dissolved in 2.1 ml formic acid (FA), then 2.1 ml of 2x TBS was added, and this sample was obtained as the insoluble fraction. The vascular fraction was obtained by washing the mesh with TBS, centrifuged at 100,000 g, the precipitate was dissolved in 1.4 ml FA, and then 1.4 ml of 2x TBS was added. Aβ1-40 and Aβ1-42 in each fraction were measured by ELISA using the assay kit according to the manufacturer’s instruction (IBL, Fujioka, Gunma, Japan).
In addition, we analyzed the TBS-soluble, SDS-soluble and FA-soluble fractions of the parietal brain according to the method previously described [11]. The TBS fraction was also analyzed by westernblotting.
Complete blood count (CBC) and serum chemistry
CBC and serum chemistry in monkeys were examined by an automated analyzer before and after the vaccination at the Tsukuba Primate Center. Examined were white blood corpuscles (WBC), red blood corpuscles (RBC), hemoglobin (HGB), hematocrit (HCT), mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), platelet (PLT), total protein (TP), alubumin (ALB), alubumin/globulin ratio (A/G), blood urea N (BUN), glucose (GLU), glutamate oxaloacetate transaminase (GOT), glutamyltranspeptidase (GPT), alkaline phosphatase (ALP), creatine kinase (CK), lactic acid dehydrogenase (LDH), creatinine (CRE), total cholesterol (TC), free cholesterol (f-CH), triglyceride (TG), C-reactive protein (CRP), P, Ca, Na, K, and Cl.
Finger maze test
Long-term memory in cynomolgus monkeys was examined using a finger maze devise composed of the 4-level maze, which was set on the cage. Each animal was trained from the first level before treatment. When a small piece of apple is put on the maze, an animal slides it with fingers laterally. In order to get an apple piece an animal must slide it each time to an appropriate direction. Otherwise, it is dropped in an error box and a monkey is not rewarded. When an animal was successful in getting 14 or more apple pieces consecutively and this was repeated twice, the animal could go to an upper level. During the training period, all 5 animals in each group completed the second level successfully. Then the vaccine was given, and the study was performed repeatedly at the third level 1 month after vaccination until sacrifice.
Pathological examination
Brain, stomach, duodenum, jejunum, ileum, colon, lung, liver, kidney, and spleen were fixed in para-formaldehyde and embedded in paraffin. Sections were stained with hematoxylin and eosin (H.E.) and examined by a veterinary pathologist (SN) in blind.
Statistical analysis
Statistical analyses were performed by two-way analysis of variance (ANOVA) followed by Student’s t test for two groups comparison. A nonparametric analysis was done by Wilcoxon/Kruskal-Wallis test, which was confirmed by median test with the Yates revision.
RESULTS
Expression of rAAV-transfected genes in vivo
To confirm the infection and expression of rAAV, B6 mice were given orally with rAAV/GFP or rAAV/Aβ, and upper part of small intestine was examined immunohistochemically. As shown, GFP and Aβ protein was expressed mainly in gut epithelial cells and some mononuclear cells in the lamina propria on the 3rd day and even on the 30th day after vaccination (Fig. 1).
African green monkeys
Enrolled African green monkeys in this study are summarized (Supplementary Table 1). Since most of African green monkeys were naturally infected with simian immunodeficiency virus (SIV) and/or simian T cell leukemia virus (STLV), these monkeys might be immunocompromised. Therefore, African green monkeys were mainly used for a safety study.
Twelve monkeys received rAAV/Aβ or rAAV/GFP vaccine once or twice orally. The brains of 4 monkeys (group A) were examined 3 months after a single vaccination; 4 monkeys (group B) were examined 6 months after a single vaccination; 4 monkeys (group C) received an additional vaccine 6 months after and the brains were examined 12 months after the first vaccination. After vaccination, none of the monkeys showed any particular signs such as vomiting, diarrhea, loss of appetite, or loss of body weight, including the monkeys receiving the second vaccine (Supplementary Figure 1). Food consumption and intake of drinking water were also unchanged. CBC and serum chemistry were unremarkable (Supplementary Figure 2).
Senile plaques without cores were common in monkeys (Fig. 2A, B). Senile plaques were very a few in the 3-month-treatment group (group A). Since 3 wild monkeys were included in this group and their real ages were unclear, we did not include those for a statistical analysis. Those in the 6-month-treatment and 12-month-treatment groups (group B and C) showed less numbers of amyloid plaques in each area of the brain in Aβ-vaccinated monkeys without statistical significance because of n = 4 in each (Fig. 2C). However, plaque numbers in total areas were 0.103/1.105 mm2 in the control and 0.054/1.105 mm2 in the Aβ-vaccinated group (p = 0.0033, Fig. 2C). The number of Aβ positive neurons was significantly lower in Aβ-vaccinated monkeys (p = 0.0129) (Fig. 2D). Cerebral amyloid angiopathy (CAA) was seen in an African green monkey, and it was not increased in Aβ-vaccinated monkeys. There were no inflammatory changes and no hemorrhages in all brains.
Serum IgG antibodies to Aβ40 and Aβ42 were measured by ELISA, but pre-immune sera of some monkeys showed a high reactivity to the ELISA kit and it was difficult to see a significant change of IgG antibodies after vaccination. However, when we looked for a correlation of the number of plaques with the IgG antibody titers, we saw that plaque numbers were significantly correlated with anti-Aβ42 but not with anti-Aβ40 titers in the control, and this correlation was not seen in the vaccinated group (Fig. 3A, B).
Pathological studies of visceral organs showed aging changes such as proliferation of mesangium cells in the renal glomeruli, otherwise no specific changes were observed except for some increase of white pulps in the spleen in both control and Aβ-vaccinated groups. We did not perform chemical analysis of the brain of African green monkeys, because some monkeys were infected with simian immune-deficiency virus, simian T-cell leukemia virus, and B virus.
Cynomolgus monkeys
Enrolled cynomolgus monkeys are listed (Supplementary Table 2). Two young monkeys were used for testing the dose of vaccine. All other aged cynomolgus monkeys received the same dose of rAAV/Aβ or rAAV/GFP (n = 5 each) orally, the same dose of the vaccine was given 3 months after, and all monkeys were sacrificed 6 months after the first vaccination.
Activities of the monkeys were not altered in all. Four Aβ-vaccinated monkeys showed a mild appetite loss; it was present before vaccination in two, it appeared transiently after the first vaccination in one, and after the second vaccination in another. Four GFP-vaccinated monkeys also showed a mild appetite loss; it was present before vaccination in three, and it appeared after the first vaccination in one. Four Aβ-vaccinated and three control monkeys showed loose bowel or watery stool occasionally, but it was unrelated to the vaccination. The vaccine did not alter menstruation cycles (Supplementary Figure 3). Body weight was reduced in most of the monkeys (Fig. 4); an Aβ-vaccinated monkey (No. 007) lost weight by 480 gm after the second vaccination, in which appetite was good, activity was not changed, stool was transiently loose, WBC were increased transiently two months after the second vaccination, and blood chemistry was unremarkable. One monkey (No. 9) showed transient elevation of GPT 1 month after the second vaccination. Eight monkeys showed transient hyperglycemia, and two were thought to have diabetes mellitus. CRP was elevated for a long period in one (No. 10) and transiently in one (No. 6). Otherwise, CBC and serum chemistry were all unremarkable (Supplementary Figure 4, Supplementary Table 3). Thus, the body weight loss in cynomolgus monkeys seems to be due to frequent observations.
We also could not confirm elevated antibody titers to Aβ in old cynomolgus monkeys. Sera from some monkeys showed a high reactivity to the ELISA kit as in African green monkeys. We applied a similar method to unmask antibodies from antigens [16], and the result was the same. However, anti-Aβ40 IgG levels were significantly correlated with amyloid plaque numbers (R2 = 0.4927). This correlation was not seen in rAAV/Aβ vaccinated monkeys. Anti-Aβ42 antibody levels were not correlated with plaque numbers in these monkeys.
Amyloid plaques were similar to those seen in African green monkeys except that some monkeys showed vascular amyloid deposits. Amyloid plaque burden was significantly reduced in Aβ-vaccinated monkeys (Fig. 5). Neurons containing intracellular Aβ were significantly reduced in rAAV/Aβ vaccinated monkeys. None of the monkeys showed any inflammatory changes in the brain.
Chemical analyses of the brain did not show any significant changes (Supplementary Figure 5). Therefore, we added an examination of the parietal cortex using the different method described [11]. We found that Aβ42 in the TBS as well as SDS soluble fractions was increased and it tended to be decreased in the FA-soluble fraction in rAAV/Aβ-given monkeys, suggesting that fibrillar Aβ was solubilized by this vaccine (Fig. 6a). The western blotting of the TBS-soluble fraction showed increase of some Aβ oligomers. Quantitation of Aβ dimers was significantly higher in the vaccinated monkey brain (Fig. 6b). However, NeuN-positive neurons were 291.6±11.42/1.105 mm2 (mean±s.e.) for controls and 293.2±11.72/1.105 mm2 for AAV-treated monkeys, and the difference was not significant. Cerebrovascular amyloid was not increased in AAV/Aβ-vaccinated monkeys.
Pathological examination of visceral organs showed similar aging changes as seen in African green monkeys, otherwise all organs examined were unremarkable.
In the finger maze test (Supplementary Figure 6), only one monkey (No. 013) successfully completed the 3rd stage and went up to the 4th stage. Otherwise all other monkeys could not complete the third stage during the examination period (Fig. 7).
DISCUSSION
In this study, cynomolgus monkeys and African green monkeys were immunized by oral administration of rAAV carrying a signal sequence and Aβ1-43 cDNA. In the African green monkey study, 3 wild monkeys were included in group A (3 months observation). Although their ages were estimated to be older than 20 years old from the appearance, the number of senile plaques were too few to see the vaccine effect. This may be due to wrong estimation of their ages or wild monkeys may develop senile plaques at older ages.
In our previous study, AAV/Aβ vaccinated APP transgenic mice showed significant improvement in the Y-maze test, novel object recognition test, Morris water maze test, and conditioned fear learning test [11]. In the present study, we applied a finger maze test in aged cynomolgus monkeys. However, we could not see the effect of vaccine by this test. This test may not be sensitive enough to detect the cognitive dysfunctions at this early stage. Alternatively, since the test was started one month after vaccination, most of the monkeys may have lost memory how to perform the finger maze during this period. Whatever the reason is, most of the monkeys received 5 to 10 pieces of apple in each trial and the pattern was not different between vaccine and control groups, suggesting that adverse effects of this vaccine if any do not seem to affect the performance in treated monkeys in a few months follow-up period.
Non-human primates naturally develop extracellular Aβ deposits in brain with aging [17]. Cynomolgus monkeys from East Asia were adopted and raised for medical use in the Tsukuba Primate Center (present name: Tsukuba Primate Research Center, National Institute of Biomedical Innovation). In the center, brains of 64 cynomolgus monkeys aged from 2 to 35 years old were examined for senile plaques and CAA, and most of the monkeys over 20 years old had senile plaques [18, 19]. As in humans, diffuse plaques were immunoreactive to an Aβ42 end-specific antibody, and plaque cores as well as CAA were mainly immunoreactive to an Aβ40 end-specific antibody [20]. Glycogen synthase kinase 3β, cyclin dependent kinase 5, and tau were elevated, but no phosphorylated tau or PHF were found [21]. In this study, we observed intracellular Aβ-positive neurons. Since Aβ was positive in cytoplasmic granules, it is possible that Aβ is present in cytoplasmic structures such as endosomes or phagosomes. The Aβ-positive neurons were significantly reduced in Aβ-vaccinated monkeys.
Several studies of Aβ immunization in monkeys have been reported. In rhesus monkeys (Macaca mulata), four 15–20 years old rhesus monkeys received aggregated Aβ1-42 with Freund’s complete adjuvant (CFA), followed by immunizations with Aβ1-42 and Freund’s incomplete adjuvant (IFA). Antibodies to Aβ were elevated and plasma Aβ was also elevated. However, since monkeys used were relatively young, the effect on amyloid clearance was inconclusive [22]. Evans et al. immunized 2-3 years old rhesus monkeys by electroporation with AV-1955 composed of DNAs encoding Igκ, 3xAβ1-11, panDR epitope, and 8 T helper epitopes of tetanus toxin, hepatitis B virus, and influenza matrix protein [23]. They observed good IgG response to Aβ, but the treatment effect was not studied. Lemere et al. immunized 16–30 years old Caribbean vervet monkeys (African green monkeys, Chlorocebus aethiops) with a mixture of Aβ1-40 and Aβ1-42 and CFA/IFA [24]. Plasma anti-Aβ antibody levels were elevated, amyloid plaque numbers were significantly reduced, and insoluble Aβx-40 and Aβx-42 but not soluble Aβx-40 and Aβx-42 in the brain were significantly reduced without T and B cell infiltrations in the brain. Trouche et al. immunized young lemur primates (M. murinus) (n = 25) with Aβ1-42 or its derivatives in alum adjuvants [25]. Most of the animals gained weight, but one of adjuvant controls and one of Aβ1-30-treated animals lost weigh with unknown cause. The immunized monkeys showed a high anti-Aβ IgG response, and Aβ1-40 plasma levels correlated well with these antibodies, indicating that the antibodies have a biological effect on Aβ in vivo. Tokita et al. immunized three 15-18-month-old rhesus monkeys with DNA vaccine, and found that significant elevation of Aβ1-42 IgG and reduction of Aβ deposits area [26]. They also described that the reduction of Aβ plaque burden was not confirmed by biochemical analysis, because the deposits were below detection limits biochemically. Interestingly, they describe that Aβ1-11 antibodies were elevated in 2 of 3 monkeys before immunization. Thus, old monkeys have natural antibodies related to Aβ pathology. Such antibodies were reported in humans previously [27], although more papers describe autoantibodies to Aβ as having protective nature [28–30]. In our study, these antibodies were positively correlated to amyloid plaque burden, and this correlation was not seen in Aβ vaccinated monkeys. Thus, it could be possible that oral Aβ vaccine attenuates natural antibodies related to Aβ pathology.
Kofler et al. immunized 10 aged macaques and 8 juvenile pigtailed macaques with aggregated Aβ1-42 mixed with monophosphoryl lipid A as adjuvant [31]. They have found that antibody responses were significantly lower in aged monkeys, and there were no differences in overall amyloid load between immunized and non-immunized animals. However, significant shift in oligomer size in the membrane fraction with an increase in the dimer: pentamer ratio in immunized monkeys. We also confirmed the increase of some Aβ oligomers including dimers. Since soluble Aβ is increased in immunized animals, this could be derived from dissolved amyloid fibrils. Although this study did not show significant reduction of neuronal densities in rAAV/Aβ-vaccinated monkeys, AN-1792 immunized patients showed significant reduction of brain volume [32] and neuronal densities [33]. Brain atrophy was also increased in bapineuzumab-treated patients [34]. Since both treatments showed well reduction of brain amyloid [35, 36], it is highly possible that dissolved amyloid produced Aβ oligomers. Although we did not see a significant reduction of NeuN-positive neurons, we need to see a long-term effect in future. Our result and above reports suggest that immunotherapy should be given prior to amyloid deposition in the brain. When this vaccine is given after amyloid deposition, it might be necessary to administer antibodies that trap toxic Aβ oligomers.
The mechanism how oral rAAV/Aβ works is speculative. In this monkey study amyloid burden including intracellular Aβ was significantly reduced without showing high antibody titers. It could be possible that high antibody titers may not be necessary. Indeed, oral immunization of tg2576 mice with rAAV/Aβ1-21 carrying shorter Aβ cDNA showed much lower antibody titers, yet the effect was equal as rAAV/Aβ1-43 [10]. Alternatively, the effect could be obtained by T cell mediated mechanisms [37]. As shown by Butovsky et al., Th2 biased immunizations with glatiramer acetate showed beneficial effects in AD model mice [38]. If this is the case, our oral vaccine has an advantage compared to other active immunization, because the gut immune system is shifted to Th2 immune responses [39]. Recently, APP/PS1 transgenic mice orally given with AAV/Aβ1-43 showed clearance of plaque amyloid in association with increased antibodies to Aβ and increased autophagy markers [40]. Thus, augmented autophagy is another possible mechanism for Aβ clearance due to this vaccine.
The AAV vector is safe [41], and it is widely tested in humans for gene therapy of congenital enzyme defects and supplementation of neurotrophic factors in AD without showing any side effects. Previously we showed that the viral DNA was detected only in the gut after oral vaccination [10]. Since gut epithelial cells are renewed in a few days, majority of the transfected gene is deleted quickly. However, some of the gene seems to be retained episomally in the epithelial stem cells of the gut for a while.
In conclusion, our oral rAAV/Aβ vaccine seems to be easy, safe, effective, relatively inexpensive, and long lasting. This vaccine should be given prior to amyloid deposition in the brain. In the later stage, administration of antibodies that trap toxic Aβ oligomers may be required after this vaccine.
Footnotes
ACKNOWLEDGMENTS
This study was partially supported by the grant from the National Institute of Biomedical Innovation (NIBIO) and Comprehensive Research on Aging and Health, the Ministry of Health, Labour and Welfare. The authors are grateful to Mr. Takayuki Fukazawa at Sunsho Pharmaceutical Co. for their providing us enteric dissolving capsules, and Dr. Xiao Xiao at University of Pittsburgh for providing us plasmids for making AAV vectors. The authors are grateful to Dr. Keiji Terao at the Tsukuba Primate Center, Mr. Kitazawa, Mr. Nagata, and Mr. Tamura at Japan SLC Inc., Dr. Miwako Kato and Dr. Khoji Kitaguchi at National Institute for Longevity Sciences, and Ms. Yukako Hasegawa and Professor Heii Arai at Juntendo University for their kind help in this study.
