Induction of a Th2 immune response by co-administration of recombinant adenovirus vectors encoding amyloid β-protein and GM-CSF

Vaccine 23 (2005) 2977–2986

Induction of a Th2 immune response by co-administration of recombinant adenovirus vectors encoding amyloid ␤-protein and GM-CSF Hong-Duck Kima , Yunpeng Caoa,d , Fan-Kun Kongc , Kent R. Van Kampenc , Terry L. Lewisa , Zhendong Maa , De-chu C. Tangb,c , Ken-Ichiro Fukuchia,∗ a

b

Department of Genetics, Schools of Medicine and Dentistry, University of Alabama at Birmingham, KHGB 640B, 720 20th Street South, Birmingham, AL 35294-0024, USA Department of Dermatology, School of Medicine, University of Alabama at Birmingham, Birmingham, AL 35294, USA c Vaxin Inc., Birmingham, AL 35294, USA d Department of Neurology, First Affiliated Hospital, China Medical University, Shenyang 110001, PR China Received 3 June 2004; received in revised form 10 December 2004; accepted 13 December 2004 Available online 22 January 2005

Abstract Lines of experimental evidence indicate that induction of humoral immune responses in transgenic mouse models of Alzheimer disease (AD) by repeated injection of synthetic amyloid ␤-protein (A␤) is effective in prevention and clearance of deposits of A␤ aggregates in the brain of the mice. We have tested a non-injection modality whereby replication-defective adenovirus vectors encoding A␤ or the 99-amino acid carboxyl terminal fragment of A␤ precursor were intranasally administered to mice to elicit immune responses against A␤. When mice were immunized only with the adenovirus vectors, immune responses against A␤ were negligible. By co-immunization with an adenovirus vector encoding granulocyte-macrophage colony stimulating factor (GM-CSF), the adenovirus vector encoding A␤ effectively elicited an immune response against A␤. Immunoglobulin isotyping demonstrated a predominant IgG1 and IgG2b response, suggesting a Th2 anti-inflammatory type. Thus, adjuvantation is essential for induction of an immune response against A␤ by adenovirus-mediated nasal vaccination. © 2005 Elsevier Ltd. All rights reserved. Keywords: Alzheimer disease; Vaccine; Granulocyte-macrophage colony stimulating factor

1. Introduction Alzheimer disease (AD) is the most common neurodegenerative disease in the elderly and characterized by amyloid plaques, neurofibrillary tangles, and neuronal loss in the brain. To date, no satisfactory treatments are available for AD. The main constituent of amyloid plaques is amyloid ␤-protein (A␤). Increasing evidence supports the notion that A␤ and amyloid ␤-protein precursor (APP) play important roles in the pathogenesis of AD. Mutations in three different genes (APP, presenilin 1 and 2) implicated in the etiology of familial AD have been shown to increase A␤ production, particularly, a longer form of A␤ consisting of 42 amino acids ∗

Corresponding author. Tel.: +1 205 975 8715; fax: +1 205 975 5519. E-mail address: [email protected] (K.-I. Fukuchi).

0264-410X/$ – see front matter © 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.vaccine.2004.12.015

(A␤1-42). A␤1-42 is considered to be highly amyloidogenic and oligomeric forms of A␤ are neurotoxic [1,2]. Overexpression of the mutant forms of APP in transgenic mice led to AD-like pathologies including amyloid plaques in brain. Immunization of these AD model mice with synthetic A␤ by repeated needle injection prevented or reduced A␤ deposits [3] and ameliorated their memory and learning deficits [4,5]. Clinical trial of A␤ immunization, however, was halted due to brain inflammation presumably induced by toxic A␤ and/or T-cell-mediated immune response [6,7]. Thus, safer immunization modalities have to be exploited for AD treatment. Genetic immunization is an approach for eliciting immune responses against specific proteins by expressing genes encoding the proteins in animal’s own cells and can be a safer modality for AD treatment. Genetic immunization may simplify the vaccination protocol because the difficult steps of

2978

H.-D. Kim et al. / Vaccine 23 (2005) 2977–2986

protein purification and combination with adjuvant, both routinely required for vaccine development, are eliminated. It has been demonstrated that genetic vaccines may be more effective than contemporary clinically licensed vaccines in animal models [8]. Indeed, we previously demonstrated that nasal administration of adenovirus vector encoding the tetanus toxin C fragment induces a protective immune response against tetanus in mice [9]. In the present study, we have tested a noninjection vaccination modality whereby defective adenovirus vectors encoding A␤ or the 99-amino acid carboxyl terminal fragment (C99) of A␤ precursor (APP) are intranasally inoculated into mice to elicit an immune response against A␤. Endogenous expression of the antigens may play important roles not only in the relatively greater efficacy of genetic vaccines but also in induction of a different type of immune response compared with conventional vaccines. Because cytokines play a critical regulatory role in the development of these immune responses induced by genetic immunization and granulocyte-macrophage colony stimulating factor (GMCSF) is one of the most potent cytokines that augment such immune responses [10,11], we have tested the potential of GM-CSF as a genetic adjuvant in induction of an immune response against A␤. We have performed histopathological and immunohistochemical analyses of the immunized mice to determine the safety of the modality, also.

2. Materials and methods 2.1. Construction of expression vectors In order to optimize expression of A␤, four expression vectors (Table 1) were constructed. These vectors differ in presence and absence of a Kozak consensus sequence [12] and in DNA sequences of codon usage [13]. In choosing more frequently used codons for A␤amino acid sequences, putative hairpin formation were avoided using GENETYX software for DNA analysis (Software Development Co. Ltd., Tokyo, Japan). All the cDNA constructs were placed under the control of the CMV enhancer/␤-actin promoter in pCAGGS [14] or pCAGGSnc [15]. (i) pCA-A␤ for

expression of methionine + A␤1-42: to obtain the initiation ATG codon + A␤1-42 sequence + stop TAG codon, PCR was performed using: ALA12 primer 5 -CCG CTC GAG GTC GCG ATG GAT GCA GAA TTC CGA CAT GAC TCA-3 , and ALA11 primer 5 -CCG CTC GAG GGG GGT CTA CGC TAT GAC AAC ACC GCC CAC CAT-3 , and pCA-S␤C as a template [16], which contain the native cDNA sequence of C99. A 162-bp PCR product was digested with XhoI and cloned into the XhoI cloning site of pCAGGS [14] to generate pCA-A␤ that contains the native translation start site of APP. (ii) pCA-SA␤ for expression of the signal sequence of APP + A␤1-42: the 162-bp PCR product above was digested with EcoRI and XhoI restriction enzymes and cloned into the EcoRI and XhoI sites of pKS+ -S␤C [16] from which the EcoRI/XhoI fragment was previously removed, to create pKS+ -SA␤. The NotI/XhoI fragment containing the signal sequence + A␤1-42 was isolated from pKS+ -SA␤ and the NotI site was ligated onto the NotI site of pCAGGSnc [15] that was previously digested with restriction enzymes, NotI and ClaI. The other ends were ligated after blunt end formation by Klenow fill-in reaction to create pCA-SA␤ that has the native translation start site of APP. (iii) pCA-fSA␤ for expression of the signal sequence of APP + A␤1-42 using Kozak sequence and more frequently used codons for human and mouse: using four overlapping oligonucleotides, a double stranded DNA for an EcoRI site + Kozak consensus sequence + signal sequence of APP + A␤1-42 sequence + stop TAA codon + EcoRI site was synthesized. Two oligonucleotides, 5 -TTG AAT TCG CCA CCA TGC TGC CCG GCC TGG CCC TGC TGC TGC TGG CCG CCT GGA CCG CCA G-3 and 5 -GTG GTG CAC CTC GTA GCC GCT GTC GTG TCT GAA CTC GGC GTC GGC TCT GGC GGT CCA GGC G-3 , with 15 overlapping nucleotides (underlined) were annealed and the double stranded DNA was synthesized using Taq DNA polymerase. The other set of oligonucleotides, 5 -GAG GTG CAC CAC CAG AAG CTG GTG TTC TTC GCC GAG GAC GTG GGC AGC AAC AAG GGC GCC ATC AT-3 and 5 -TTG AAT TCT TAG GCG ATC ACC ACG CCG CCC ACC ATC AGG CCG ATG ATG GCG CCC TTG-3 , were annealed and the double stranded DNA was similarly produced. After Alw44I

Table 1 Expression vectors and their DNA constructs for expression of A␤1-42 Expression vectors

Translation start sites and coding sequences

pCA-A␤ pCA-SA␤ pCA-fSA␤ pCA-fKSA␤

5 -XhoI-Nativea -Metb -A␤1-42c -Stopd -XhoI-3 5 -BamHI-Nativea -Signale -Metb -A␤1-42c -Stopd -XhoI-3 5 -EcoRI-Kozakf -fg Signale -fg A␤1-42c -Stopd -EcoRI-3 5 -EcoRI-Kozakf -fg KSignalh -fg A␤1-42c -Stopd -EcoRI-3

a b c d e f g h

Native translation start signal of APP, GTCGCG. Methionine codon, ATG. cDNA for the A␤ peptide consisting of 42 amino acids, DAEFRHDSGYEVHHQKLVFFAEDVGSNKGAIIGLMVGGVVIA. A stop codon. cDNA for the signal peptide of APP, MLPGLALLLLAAWTARA. A Kozak consensus sequence, GCCACC. More frequently used codons in humans and mouse were used for cDNA. cDNA for the signal peptide of immunoglobulin Kappa light chain, MSVPTQVLGLLLLWLTDARC.

H.-D. Kim et al. / Vaccine 23 (2005) 2977–2986

digestion, both of the double stranded DNA were ligated and cloned into the EcoRI site of pCAGGS to create pCA-fSA␤. (iv) pCA-fKSA␤ for expression of the signal sequence of immunoglobulin Kappa light chain + A␤1-42 using Kozak sequence and more frequently used codons for human and mouse. Using the following four overlapping oligonucleotides, pCA-fKSA␤ was constructed as in pCA-fSA␤. The first two sets are 5 -TTG AAT TCG CCA CCA TGA GCG TGC CCA CCC AGG TGC TGG GCC TGC TGC TGC TGT GGC TGA CCG ACG CCA G-3 and 5 -GTG GTG CAC CTC GTA GCC GCT GTC GTG TCT GAA CTC GGC GTC GCA TCT GGC GTC GGT CAG CC-3 The second two sets are the same as the second sets in pCA-fSA␤. (v) pCA-fSC99 was constructed using six different overlapping oligonucleotides. The first pair at the 5 end was 5 -TTG AAT TCG CCA CCA TGC TGC CCG GCC TGG CCC TGC TGC TGC TGG CCG CCT GGA CCG CCA G-3 and 5 -GTG GTG CAC CTC GTA GCC GCT GTC GTG TCT GAA CTC GGC GTC GGC TCT GGC GGT CCA GGC G-3 . The next at the middle fragment was 5 -GAG GTG CAC CAC CAG AAG CTG GTG TTC TTC GCC GAG GAC GTG GGC AGC AAC AAG GGC GCC ATC AT-3 and 5 -GGG TGA TCA CGA TCA CGG TGG CGA TCA CCA CGC CGC CCA CCA TCA GGC CGA TGA TGG CGC CCT TG-3 . The third was 5 -TCG TGA TCA CCC TGG TGA TGC TGA AGA AGA AGC AGT ACA CCA GCA TCC ACC ACG GCG TGG TGG AGG TGG ACG CCG CCG TGA CCC CCG AGG AGA GA-3 and 5 -TTG AAT TCT TAG TTC TGC ATC TGC TCG AAG AAC TTG TAG GTG GGG TTC TCG TAG CCG TTC TGC TGC ATC TTG CTC AGG TGT CTC TCC TCG GGG GT-3 . After making double stranded DNA using TaqDNA polymerase, the first and the second pairs were ligated using Alw44I site (bold) and the second and the third pairs were ligated using BclI site (bold). The resulting cDNA was cloned into the EcoRI site of pCAGGS to create pCA-fSC99. The authenticity of the five expression vectors was confirmed by DNA sequencing of isolated clones. 2.2. Transfection and comparison of secreted A␤ by immunoprecipitation and ␤-galactosidase activity The methods for transfection and ␤-galactosidase activity assay have been described previously [17]. In brief, 2 ␮g of pCA␤, a ␤-galactosidase reporter expression vector [17], and 2 ␮g of each expression vector described above were co-transfected into COS cells that had been inoculated at concentrations of 1.5 × 105 cells into a 60 mm-dish 1 day before transfection using LifpofectamineTM (Invitrogen, Carlsbad, CA) according to the manufacturer’s protocol. Three independent transfections for each expression vector were performed. Sixty hours after transfection, the cells and media were harvested for ␤-galactosidase activity assay and immunoprecipitation of secreted A␤, respectively. ␤-Galactosidase activity of each cell sample was determined by fluorometric assay at 350 nm excitation and 450 nm

2979

emission using cell lysate containing 100 ␮g protein and 0.6 mM MUG as a substrate. For immunoprecipitation of A␤, the medium was centrifuged at 12,000 × g for 1 h and the supernatant was brought to 10 mM Tris (pH 8.0), 1% Nonidet P-40, 0.5% cholic acid, 0.1% SDS, 5 mM EDTA, 2 ␮g/ml leupeptin, and 0.1 ␮g/ml pepstatin. After preabsorption with 50 ␮l of protein A-agarose (protein A-agarose was previously loaded with rabbit antibody against mouse antibody), the supernatant was incubated with 6E10 antibody (Signet Pathology Inc., Dedham, MA) against A␤1-17 and protein A-agarose (pretreated as described above) for 16 h. Precipitates were washed three times with Tris-buffered saline containing 1% Nonidet P-40, 5 mM EDTA, 2 ␮g/ml leupeptin and 2 mM phenylmethylsulfonyl fluoride and then boilded in 20 ␮l of two times Laemmli buffer for 5 min before being subjected to a 16.5% Tris–Tricine SDS–polyacrylamide gel electrophoresis (PAGE). Five micrograms of synthetic A␤1-42 was added to each PAGE as a standard. Western blotting was performed using 6E10 antibody and an enhanced chemiluminescence system (Amersham, Arlington Heights, IL) as described previously [18]. The relative concentration of the protein (A␤) was determined by densitometric scanning and normalized by ␤-galactosidase activity. The relative concentration of the protein (A␤) was determined by densitometric scanning of the membrane using a Fluor-S MultiImager and PDQuest software (Bio-Rad, Hercules, CA). The ratio of a densitometric reading of A␤ protein to that of 5 ng synthetic A␤ was used as Western blot data (Fig. 1b). For each sample, the ratio was normalized by ␤-galactosidase activity (Fig. 1b). By this method, the minimum detectable A␤ was ∼1 ng. The data were analysed using analysis of variance (ANOVA). 2.3. Construction and preparation of E1/E3-defective adenovirus vectors E1/E3-defective adenoviral vectors, AdfKSA␤, AdC99, and AdLacZ were prepared using HEK293 cells and AdEasy® Basic Kit (American Type Culture Collection, Manassas, VA). The DNA fragments containing the CMV enhancer/␤-actin promoter + cDNA (for antigens) + ␤-globin poly A signal were isolated from the expression vectors (pCA-fKSA␤, pCA-fSC99, and pCA␤) by SalI and HindIII digestion and cloned into the XhoI (compatible with SalI) and HindIII sites of pShuttle plasmid (AdEasy® Basic Kit) to produce the adenovirus vectors. pCA␤ encodes cDNA for ␤-galactosidase under the control of the CMV enhancer/␤actin promoter. After homologous recombination between the shuttle plasmids and pAdEasy-1 in Escherichia coli BJ5183 cells, E1/E3-defective adenoviral vectors encoding A␤, C99, and ␤-galactosidase were prepared in HEK293 cells. After several rounds of passage in HEK293 cells, the adenovirus vectors were purified using the BD Adeno-X virus purification kits (BD Biosciences, Palo Alto, CA) and the titers of the purified adenovirus vectors were determined using the Adeno-X rapid titer kits (BD Bioscience) according

2980

H.-D. Kim et al. / Vaccine 23 (2005) 2977–2986

Fig. 1. Optimization of secreted A␤ by Western blot analysis and ␤-galactosidase assay. COS cells were transfected with pCA␤ carrying the LacZ gene and one of the following four expression vectors: pCA-A␤, pCA-SA␤, pCA-fSA␤, and pCA-fKS␤. Three days after the transfection, immunoprecipitated A␤ was subjected to PAGE followed by Western blotting. Three independent transfections were performed and one of the Western blot results is shown (a). The means of the ratios of Western signals from the samples to that of 5 ng A␤ are plotted as open bar graphs in (b). The means of the ratios normalized by ␤-galactosidase are shown as closed bar graphs in (b). Error bars are M.S.E. * P < 0.01 for pCA-fKS␤ compared with any other vectors.

to the manufacturer’s protocols. The produced adenovirus vectors for expression of A␤, C99, and ␤-galactosidase were designated as AdfKSA␤, AdC99, and AdLacZ, respectively. Expression of the antigens in transduced HEK293 cells was determined by Western blotting using the enhanced chemiluminescence system and 994B antiserum (raised against the 39 amino acid C-terminal residues of APP) [18] for C99 and 6E10 antibody (Signet Pathology Inc., Dedham, MA) for A␤ as described above. An adenovirus vector encoding murine GM-CSF (AdGMCSF) [19] was prepared by ultracentrifugation over a cesium chloride gradient as previously described [9]. 2.4. Experimental animals Seven-month-old Tg13592 mice and their non-transgenic littermates were used. Establishment, propagation, and maintenance of the transgenic mouse line, Tg13592, were previously described [18,20]. All of the Tg13592 mice used in this study had been back-crossed to C57BL/6 mice more than 10 generations. Tg13592 mice overexpress the signalpeptide + C99. Four to six Tg13592 mice were used for each adenovirus vector treatment with and without AdGM-CSF and for phosphate-buffered saline (PBS) treatment. Nontransgenic littermate mice were used for AdfKSA␤ treatment with (n = 6) and without AdGM-CSF (n = 6) and for PBS treatment (n = 5). To study the safety of adenovirus vaccina-

tion, 6-month-old C57BL/6J mice were used for AdfKSA␤ treatment with AdGM-CSF (n = 6) and for PBS treatment (n = 4). To examine expression of human A␤ in the nasal tissues of mice vaccinated with AdfKSA␤, 2-month-old C57BL/6J mice were treated with AdfKSA␤ (n = 3) and with PBS (n = 2) as controls. All animal protocols used for this study were prospectively reviewed and approved by the Institutional Animal Care and Use Committee of the University of Alabama at Birmingham. 2.5. Immunization protocol Intranasal inoculation was carried out by pipetting 20 ␮l of adenovirus (1 × 108 plaque forming unit, PFU) into one of the nostrils of an anesthetized mouse with or without AdGMCSF (1 × 108 PFU) followed by a booster every 3 weeks for 12 weeks or 18 weeks. Blood was collected through the tail 0 (preimmune), 6, 12, and 18 weeks after the initial immunization. To examine expression of human A␤ in nasal tissues, C57BL/6J mice were sacrificed 5 days after intranasal inoculation of AdfKSA␤. 2.6. ELISA for determination of serum titers After incubating blood at 4 ◦ C for 3 h, the blood was centrifuged for 10 min at 10,000 × g. The serum was collected and frozen at −80 ◦ C. Anti-A␤ immunoglobulin (IgM,

H.-D. Kim et al. / Vaccine 23 (2005) 2977–2986

IgG, IgG1, IgG2a, and IgG2b) titers were determined by enzyme-linked immunosorbent assay (ELISA) as described by Spooner et al. [21] with modification. Microtiter wells were coated with A␤ 1–42 (5 ␮g/ml) in 50 mM carbonate buffer pH 9.6 overnight at 4 ◦ C and rinsed three times with washing buffer [PBS containing 0.05% Tween-20]. Microtiter wells were treated with blocking buffer (5.0% goat serum, 1% BSA and 0.05% Tween-20 in PBS) for 2 h at room temperature. The serum samples were diluted with PBS and added to the microtiter wells. After incubation for 2 h at room temperature, the plates were washed five times with the washing buffer and incubated for 1 h with an appropriate horseradish peroxidase (HRP)-conjugated detection antibody. The detection antibodies were diluted at 1:2000 for anti-mouse IgG and IgM and 1:1000 for anti-mouse IgG1, IgG2a, and IgG2b in the blocking buffer. The detection antibodies were purchased from Zymed (South San Francisco, CA). After washing the plates with the washing buffer, plates were incubated with TMB (Kirkegaard & Perry Laboratories Inc., Gaithersburg, MD) for 15 min and the reaction was stopped with the addition of 1N H2 SO4 . Optic densities at 450 nm were determined using a Microplate Reader (Labsystems, Finland). Anti-A␤ antibody titers in the mouse sera were determined using serial dilutions of 6E10 (monoclonal anti-A␤ antibody) as the standard. Therefore, the concentrations (␮g/ml) of the serum titers presented here reflect the concentrations of 6E10 antibody, which produce the same ELISA readings, and may not accurately represent the absolute amounts. In the present study, the linear regression was found between 0.1 and 24 ng/ml of IgG. The minimum detectable anti-A␤ titer was 0.1 ng/ml, which was determined by adding two standard deviations to the mean absorbance obtained when the zero standard was assayed 32 times. Comparison of treatment groups was performed by ANOVA using the SigmaStat software (SPSS Science, Chicago, IL). P < 0.05 was considered statistically significant. 2.7. Histopathological and immunohistochemical analyses Mice were sacrificed by intraperitoneal injection of sodium pentobarbital. The heads were fixed in 10% formaldehyde:90% alcohol and decalcified. Six-micrometer sections were prepared for hematoxylin–eosin staining and immunohistochemistry. To detect human A␤ expression in the nasal lesions and brain, the avidin–biotin immunoperoxidase method was employed using Vectastain ABC kit (Vector, Burlingame, CA) and 6E10 antibody that specifically reacts with human A␤. Endogenous peroxidase was eliminated by treatment with 3% H2 O2 for 30 min. After washing with distilled water, the sections were blocked with 10% horse serum in 0.1 M Tris–saline (TBS) (pH 7.4) for 60 min at room temperature and incubated with 6E10 antibody (1 ␮g IgG/ml) in 0.1 M TBS containing 10% horse serum for 16 h at 4 ◦ C. The sections were rinsed

2981

in 0.1 M TBS containing 1% serum and incubated with biotinylated secondary antibody for 60 min at room temperature. After washing, the sections were incubated with Vectastain ABC reagent for 60 min at room temperature. Peroxidase activity was detected by treatment with 3,3 diaminobenzidine. The sections were counterstained with hematoxylin. The head sections covered the ethmoid turbinates, the nasalpharyngeal duct, the base of the nasal septum and various glandular structures within the upper respiratory tract. Level II sections were taken through the level of the incisive papilla anterior to the first palatial ridge. Level III sections were taken through the middle of the second molar tooth and include sections of the rostral portions of the olfactory lobe of the brain. The eyes, glandular structures supporting the eyes, bones, and muscles of the skull, and jaws were included in various sections. Maxillary sinuses, venous sinuses, septal glands, and the sustentacular cells supporting the epithelium of the respiratory tract were also observed microscopically.

3. Results 3.1. Optimization of secreted A␤ It has been demonstrated that the use of a Kozak consensus sequence and more frequently utilized codons for human and mouse can increase translation of mRNA in mammalian cells and transgenic mice [12,13]. Therefore, we have constructed two expression vectors, pCA-fSA␤ and pCA-fKSA␤ (Table 1), using the Kozak sequence and more frequently used codons for human and mouse [22]. In addition, we have tested two different signal sequences, the native signal sequence of APP in pCA-SA␤ and pCA-fSA␤ as well as that of immunoglobulin Kappa light chain in pCA-fKSA␤ (Table 1). These vectors and pCA-A␤ (the native A␤ DNA sequence without a signal sequence) were co-transfected with the LacZ gene into COS cells. The media from the cells were subjected to immunoprecipitation using 6E10 antibody against A␤ followed by Western blotting and densitometric scanning. A representative Western blot from three independent transfection experiments is shown in Fig. 1a. Levels of secreted A␤ from these expression vectors were compared with 5 ng of synthetic A␤ (Fig. 1b). The data were normalized by ␤-galactosidase activities (Fig. 1b). Approximately, 10 ng/5 ml medium of A␤ was detected in COS cells transfected with pCA-fKSA␤ while COS cells tranfected with other vectors produced 3–5 ng/5 ml medium of A␤. Thus, consistently high levels of A␤ secretion were obtained from pCA-fKSA␤ (P < 0.01). There were no significant differences among the three other vectors. Accordingly, we have decided to use pCA-fKSA␤ to produce an adenovirus vector. The Kozak sequence and frequently used codons were also used to construct an expression vector for C99 of APP.

2982

H.-D. Kim et al. / Vaccine 23 (2005) 2977–2986

Fig. 2. HEK293 cells express A␤ and C99 after transduction with AdfKSA␤ and AdC99, respectively. Cell lysates from HEK293 cells transduced with AdfKSA␤ and AdC99 were subjected to PAGE followed by Western blotting using 6E10 antibody against A␤ and 994B antiserum against the C-terminus of C99. High expression of A␤ and C99 are seen as 4 kDa (a) and 14 kDa (b) protein (indicated by the arrows).

Fig. 3. Expression of human A␤ in respiratory epithelium in mice immunized with AdfKSA␤. Five days after intranasal inoculation of AdfKSA␤, mice were sacrificed. Serial cross-sections of nasal airway were subjected to immunostaining with 6E10 antibody (a) or with normal mouse serum (b). Human A␤ was detected by 6E10 antibody (a) (arrows) but not by normal mouse serum (b). O: olfactory epithelium; R: respiratory epithelium. Scale bars 20 ␮m (a and b).

3.2. Confirmation of A␤ and C99 expression in HEK293 cells Adenovirus vectors, AdfKSA␤ and AdC99, were prepared in HEK293 cells. To confirm expression of A␤ and C99 from the vectors, HEK293 cells were transduced and the cell lysates (20 ␮g of protein for each sample) were subjected to Western blotting using 6E10 and 994B antibodies against A␤ and C99, respectively. A␤ and C99 were observed as 4 kDa and 14 kDa protein, respectively (Fig. 2). 3.3. Confirmation of A␤ expression in nasal epithelium after intranasal inoculation of AdfKSA␤ To study expression of human A␤ in vaccinated mice, C57BL/6J mice were sacrificed 5 days after intranasal inoculation of AdfKSA␤ (1 × 108 PFU/mouse) or PBS. The sections of nasal lesions and brains from the mice were subjected to immunohistochemical analysis using 6E10 antibody specific to human A␤. Specific A␤-immunoreactivity was exclusively found in nasal epithelium. Particularly, respiratory epithelial cells were specifically stained with 6E10 antibody in the AdfKSA␤-treated airway (Fig. 3a and b). The mice treated with PBS did not show specific immunoreactivity to 6E10 antibody.

3.4. Adjuvantation is essential for induction of an immune response to A␤ by AdfKSA␤ Anti-A␤ antibody titers in the mouse sera were determined by ELISA at each of four times points (weeks 0, 6, 12, and 18) using serial dilutions of 6E10 antibody as the standard (Table 2 and Fig. 4). After the second vaccination with AdfKSA␤ and AdGM-CSF, anti-A␤ antibodies (>0.5 ␮g/ml serum) were detected in two of five Tg13592 mice and in three out of six non-transgenic littermates. After the fourth vaccination with AdfKSA␤ and AdGM-CSF, all of the mice had more than 1.3 ␮g/ml serum titers irrespective of whether the mice were Tg13592 or non-transgenic littermates. These titers were markedly boosted by the following immunization up to week 18 (>18 ␮g/ml). There is no significant difference in anti-A␤ serum titers between Tg13592 mice and their non-transgenic littermates immunized with AdfKSA␤ and AdGM-CSF (P > 0.3 for IgG). Contrary to AdfKSA␤, only one out of five Tg13592 mice immunized with AdC99 and AdGM-CSF showed a significant serum titer (0.9 ␮g/ml) 6 weeks after the initial vaccination and the titer increased to 12 ␮g/ml following booster vaccination. The other mice in this group did not develop

H.-D. Kim et al. / Vaccine 23 (2005) 2977–2986

2983

Table 2 Characterization of anti-A␤ antibodies induced by adenovirus vaccines Mouse genotype

Immunogensa and adjuvant

Mice with A␤ antibody

IgGb

IgG1b

IgG2bb

Non-Tg Non-Tg Tg13592 Tg13592 Tg13592

PBS fKSA␤ + GM-CSF PBS fKSA␤ + GM-CSF C99 + GM-CSF

0/5 6/6 0/5 5/5 1/5

0 21.03 ± 0.60c 0 22.48 ± 0.65c 11.93d

0 19.82 ± 0.42c 0 19.42 ± 0.56c 11.01d

0 21.43 ± 1.08c 0 21.69 ± 0.62c 13.50d

a b c d

Immunogens and an adjuvant were produced from AdfKSA␤, AdC99, and AdGM-CSF. Anti-A␤-antibody titers by ELISA 18 weeks after the initial vaccination. Means ± S.E.M. The means of two measurements from one Tg13592 mouse that had a significant immune response.

significant antibodies against A␤ up to 18 weeks even with booster vaccination (Table 2). No anti-A␤ antibody was detected in serum from mice treated with PBS. Anti-A␤ antibody titers from Tg13592 mice vaccinated with AdGM-CSF alone (four mice) or with AdLacZ and AdGM-CSF (five mice) were very modest or negligible throughout the experiments. Anti-A␤ antibody titers fromTg13592 and their littermates (five to six mice for each group) immunized only with AdfKSA␤ or AdC99 were negligible up to 12 weeks even after the booster immunization (P < 0.0001 compared with the mice immunized with AdfKSA␤ and AdGM-CSF at week 12) (Fig. 4). Therefore, we terminated the experiments of these groups at week 12. Thus, GM-CSF is essential for induction of a significant immune response to A␤ by adenovirus vectors. 3.5. Induced anti-A␤ antibodies are predominantly IgG1 and IgG2b: type 2 T-helper cell immune response To quantify the immunoglobulin isotypes of the antiA␤ antibodies in the mice immunized with AdfKSA␤ and AdGM-CSF or AdC99 and AdGM-CSF, ELISA was performed using isotype specific antibodies (IgM, IgG1, IgG2a, and IgG2b) and 6E10 antibody as a standard (Table 2 and Fig. 3b and c). The induced anti-A␤ antibodies are predominantly of the IgG1 and IgG2a isotypes. Compared with IgG1 and IgG2b, increases in IgG2a were modest (<5 ␮g/ml at 18 weeks) and those in IgM was negligible. 3.6. Safety of nasal vaccination by adenovirus-vectored A␤-vaccines To investigate possible side effects of nasal vaccination by adenovirus-vectored A␤-vaccines, histopathological examination was performed on tissue sections from the upper respiratory tract and brain of six C57BL/6J mice immunized with AdfKSA␤ and AdGM-CSF. Four C57BL/6J mice treated with PBS were used as controls. The mice were sacrificed at 10 months of age 12 days after the sixth immunization. Compared with control mice, no significant abnormalities associated with adenovirus vaccination were observed in the respiratory epithelium, the nasal turbinates, the maxillary sinuses, the eyes, the olfactory tissue, or any other structures

examined within the sections of the nasal tissues. Two of the animals immunized with AdfKSA␤ and AdGM-CSF had cystic dilation of the lacrimal glands, unilaterally while one of the animals had a benign adenoma of one of the glandular structures associated with the eye. Based on the small number of animals examined, these later findings are biologically significant, but not necessarily treatment related. No inflammatory response related to immunization was observed in the mice. No amyloid deposition was found in the brain by immunohistochemistry.

4. Discussion A Kozak consensus sequence and more frequently used codons for human and mouse are known to increase translation of mRNA in mammalian cells and transgenic mice [12,13]. In this limited study, we did not observe such an effect on secreted A␤ by changing the cDNA sequence of A␤. The use of the immunoglobulin Kappa light chain signal sequence, however, produced a significantly greater level of secreted A␤ in COS cells than the native signal sequence of APP. The reasons for these discrepancies are not clear. It is possible that the cDNA of A␤ is too short (177 bp in the coding sequence) to produce any significant difference in translation. Because we did not determine the levels of cytosolic A␤, it is also possible that the Kozak sequence and more frequently used codons increased translation but the signal sequence of APP is less effective in leading A␤ to secretory pathways than that of immunoglobulin when these signal sequences are fused to A␤. Considering efficient secretion of highly diverse immunoglobulin, the latter explanation is quite plausible. As we previously demonstrated [9], intranasal administration of adenovirus vectors encoding the tetanus toxin C fragment elicits a protective immune response against tetanus in mice. In the present study, however, intranasal administration of AdfKSA␤ alone was ineffective in eliciting an immune response against the self-antigen, A␤. The hurdle of A␤’s low immunogenicity was overcome by co-administration of an additional adenovirus vector that expresses murine GM-CSF as an adjuvant. The present results indicate that GM-CSF can enhance the induction of humoral immune response against a self-antigen. GM-

2984

H.-D. Kim et al. / Vaccine 23 (2005) 2977–2986

Fig. 4. Induction of anti-A␤ antibodies after vaccination with AdfKSA␤ and AdC99 combined with or without AdGM-CSF as an adjuvant. Serum titers for anti-A␤ antibodies were determined by ELISA using anti IgG (a), IgG1 (b), and IgG2b (c) at the indicated time points: pre-serum (week 0), 6 weeks, 12weeks, and 18 weeks after the initial vaccination. Inoculations without AdGM-CSF were done up to 12 weeks. Tg13592 and their littermates (Non-Tg) were used. For the AdfKSA␤ and AdGM-CSF treatment groups, the means of serum titers from six Non-Tg mice and five Tg13592 mice are shown. For the AdC99 and AdGM-CSF treatment group, the data from only one Tg13592 mouse that had a significant immune response are shown. The error bars are S.E.M. * P < 0.0001 for AdfKSA␤ with AdGM-CSF compared with AdfKSA␤ without AdGM-CSF.

CSF has previously been shown to be a potent adjuvant for inducing both humoral and cellular responses even against self-antigen [23–27]. GM-CSF modulates immune responses by acting upon the antigen presentation process and amplifies the immune responses without changing the nature of immune responses from Th1 to Th2 type or vise versa [28–31]. The current study suggests that GM-CSF is effective in recruiting A␤-presenting cells which may have led to the marked production of anti-A␤ antibodies in mice. The predominant antibody isotypes elicited by intranasal administration of AdfKSA␤ in mice are IgG2b and IgG1, indicating a Th2 response. Our results are consistent with those of other investigators reporting production of predominant IgG1 and IgG2b by synthetic A␤1-40/42 regardless of the administration routes (intranasal or intraperitoneal) and adjuvants (Freund or heat-labile enterotoxin) [32–35]. Most of the antibodies produced by immunization with synthetic

A␤1-40/42 recognize epitopes with A␤1-15 [34–36]. A␤115 is thought to induce a Th2 anti-inflammatory type response in mice [33]. These data suggest that immunization of mice with A␤ predominantly induces a Th2 response irrespective of immunization routes, vehicles, and adjuvants. It has been reported that AD model mice (Tg2576) that constitutively overexpress APP show a decreased antibody response when immunized with synthetic A␤ compared with non-transgenic controls [36]. The high level of A␤ in the blood of the AD transgenic mice is considered to lead to hyporesponsiveness of A␤-reactive T or B cells [36]. To mimic this hyporesponsiveness, we used Tg13592 mice that constitutively overexpress C99 in almost all tissues under the control of the ␤-actin promoter and that have a high level of plasma A␤ [20]. Contrary to the previous report, we did not observe significant differences in the levels of anti-A␤ antibody titers between Tg13592 and their littermates after

H.-D. Kim et al. / Vaccine 23 (2005) 2977–2986

intranasal administration of the adenovirus vectors. GM-CSF may be effective in overcoming this hyporesponsiveness in the transgenic mice. All of the Tg13592 mice vaccinated with AdfKSA␤ and AdGM-CSF produced anti-A␤ antibodies but only one out of five Tg13592 mice vaccinated with AdC99 and AdGM-CSF did (Table 2). C99 is the only precursor of A␤ and A␤ is produced by ␥-secretase cleavage from C99. We have tested AdC99 in anticipation of even greater secretion of A␤ compared with AdfKSA␤, because C99 has the membrane spanning domain of APP to become a substrate of ␥-secretase. As demonstrated by Western blotting (Fig. 1), secreted A␤ in 5 ml culture media from COS cells transfected with pCAfKSA␤ was readily detectable but that from pCA-fSC99 was indiscernible by Western blotting (data not shown). Thus, it is most likely that A␤ production was too small to effectively elicit an immune response in the mice vaccinated with AdC99. Microscopic examination of tissues taken from the mice vaccinated with AdfKSA␤ and AdGM-CSF did not reveal any significant changes that could be identified as treatment related. Particularly, no abnormal changes were noted in the respiratory epithelium, the ethmoid turbinates, or in the sinuses and no inflammatory response associated with the adenovirus vaccination was found. In the vaccinated group, two biologically significant changes were noted: cystic dilation of portions of the lacrimal glands in two animals and in one, a benign adenoma of the lacrimal gland. Without examination of a larger sample and without knowledge of the incidence of age-related changes for this particular strain of mouse, one cannot state whether or not these changes are treatment related. We are currently testing this modality in transgenic AD mouse models. The preliminary data indicate that nasal immunization with AdfKSA␤ and AdGM-CSF is effective in reducing A␤ load in the brain. Intranasal immunization with adenovirus vectors has potential advantages over conventional needle injection with synthetic peptides. The procedure eliminates pain, purification and modification of protein, and adverse side effects associated with synthetic peptide injection and so may reduce medical costs.

Acknowledgements We thank T.K. Hinds and K. Kamino for 994B antibody and Adam Maxwell and Timothy Gunn for technical support. This research is supported in part by the National Institute of Health and Alzheimer’s Association.

References [1] Lambert MP, Viola KL, Chromy BA, et al. Vaccination with soluble Abeta oligomers generates toxicity-neutralizing antibodies. J Neurochem 2001;79:595–605.

2985

[2] Walsh DM, Klyubin I, Fadeeva JV, et al. Naturally secreted oligomers of amyloid beta protein potently inhibit hippocampal longterm potentiation in vivo. Nature 2002;416:535–9. [3] Schenk D, Barbour R, Dunn W, et al. Immunization with amyloidbeta attenuates Alzheimer-disease-like pathology in the PDAPP mouse. Nature 1999;400:173–7. [4] Morgan D, Diamond DM, Gottschall PE, et al. A beta peptide vaccination prevents memory loss in an animal model of Alzheimer’s disease. Nature 2000;408:982–5. [5] Janus C, Pearson J, McLaurin J, et al. A beta peptide immunization reduces behavioral impairment and plaques in a model of Alzheimer’s disease. Nature 2000;408:979–82. [6] Check E. Nerve inflammation halts trial for Alzheimer’s drug. Nature 2002;415:462. [7] Orgogonzo JM, Gilman S, Dartigues JF, et al. Subacute meningoencephalitis in a subset of patients with AD after Abeta42 immunization. Neurology 2003;61:46–54. [8] McDonnell WM, Askari FK. DNA vaccines. N Engl J Med 1996;334:42–5. [9] Shi Z, Zeng M, Yang G, et al. Protection against tetanus by needlefree inoculation of adenovirus-vectored nasal and epicutaneous vaccines. J Virol 2001;75(23):11474–82. [10] Dranoff G. GM-CSF-secreting melanoma vaccines. Oncogene 2003;22(20):3188–92. [11] Martinez-Moczygemba M, Huston DP. Biology of common beta receptor-signaling cytokines: IL-3, IL-5, and GM-CSF. J Allergy Clin Immunol 2003;112(4):653–65. [12] Kozak M. Determinants of translational fidelity and efficiency in vertebrate mRNAs. Biochimie 1994;76:815–21. [13] Okabe M, Ikawa M, Kominami K, Nakanishi T, Nishimune Y. ‘Green mice’ as a source of ubiquitous green cells. FEBS Lett 1997;407:313–9. [14] Niwa H, Yamamura K, Miyazaki J. Efficient selection for highexpression transfectants with a novel eukaryotic vector. Gene 1991;108:193–200. [15] Fukuchi K, Kunkel DD, Schwartzkroin PA, et al. Overexpression of a C-terminal portion of the ␤-amyloid precursor protein in mouse brains by transplantation of transformed neuronal cells. Exp Neurol 1994;127:253–64. [16] Fukuchi K, Kamino K, Deeb SS, Smith AC, Dang T, Martin GM. Overexpression of amyloid precursor protein alters its normal processing and is associated with neurotoxicity. Biochem Biophys Res Commun 1992;182:165–73. [17] Fukuchi K, Hearn MG, Deeb SS, et al. Activity assays of nine heterogeneous promoters in neural and other cultured cells. In Vitro Cell Dev Biol Anim 1994;30A:300–5. [18] Fukuchi K, Pham D, Hart M, Li L, Lindsey JR. Amyloid-beta deposition in skeletal muscle of transgenic mice: possible model of inclusion body myopathy. Am J Pathol 1998;153:1687–93. [19] Lee CT, Wu S, Ciernik IF, et al. Genetic immunotherapy of established tumors with adenovirus-murine granulocyte-macrophage colony-stimulating factor. Hum Gene Ther 1997;8:187–93. [20] Fukuchi K, Ho L, Younkin SG, et al. High levels of ␤-amyloid protein in peripheral blood do not cause cerebral ß-amyloidosis in transgenic mice. Am J Pathol 1996;149:219–27. [21] Spooner ET, Desai RV, Mori C, Leverone JF, Lemere CA. The generation and characterization of potentially therapeutic A␤ antibodies in mice: differences according to strain and immunization protocol. Vaccine 2002;21:290–7. [22] Nakamura Y, Wada KN, Doi H, Kanaya S, Gojobori T, Ikemura T. Codon usage tabulated from the international DNA sequence databases. Nucleic Acids Res 1996;24:214–5. [23] Kwak LW, Young HA, Pennington RW, Weeks SD. Vaccination with syngeneic, lymphoma-derived immunoglobulin idiotype combined with granulocyte/macrophage colony-stimulating factor primes mice for a protective T-cell response. Proc Natl Acad Sci USA 1996;93(20):10972–7.

2986

H.-D. Kim et al. / Vaccine 23 (2005) 2977–2986

[24] Kinoshita Y, Kono T, Yasumoto R, et al. Antitumor effect on murine renal cell carcinoma by autologous tumor vaccines genetically modified with granulocyte-macrophage colony-stimulating factor and interleukin-6 cells. J Immunother 2001;24(3):205–11. [25] Aarts WM, Schlom J, Hodge JW. Vector-based vaccine/cytokine combination therapy to enhance induction of immune responses to a self-antigen and antitumor activity. Cancer Res 2002;62(20): 5770–7. [26] Disis ML, Shiota FM, McNeel DG, Knutson KL. Soluble cytokines can act as effective adjuvants in plasmid DNA vaccines targeting self tumor antigens. Immunobiology 2003;207(3):179–86. [27] Nemunaitis J, Sterman D, Jablons D, et al. Granulocyte-macrophage colony-stimulating factor gene-modified autologous tumor vaccines in non-small-cell lung cancer. J Natl Cancer Inst 2004;96(4):326–31. [28] Wang J, Snider DP, Hewlett BR, et al. Transgenic expression of granulocyte-macrophage colony-stimulating factor induces the differentiation and activation of a novel dendritic cell population in the lung. Blood 2000;95:2337–45. [29] Morrissey PJ, Bressler L, Park LS, Alpert A, Gillis S. Granulocytemacrophage colony-stimulating factor augments the primary antibody response by enhancing the function of antigen-presenting cells. J Immunol 1987;139:1113–9. [30] Tazi A, Bouchonnet F, Grandsaigne M, Boumsell L, Hance AJ, Soler P. Evidence that granulocyte macrophage-colony-stimulating factor regulates the distribution and differentiated state of dendritic cells/Langerhans cells in human lung and lung cancers. J Clin Invest 1993;91:566–76.

[31] Stampfli R, Wiley RE, Neigh GS, et al. GM-CSF transgene expression in the airway allows aerosolized ovalbumin to induce allergic sensitization in mice. J Clin Invest 1998;102:1704–14. [32] Leverone JF, Spooner ET, Lehman HK, Clements JD, Lemere CA. Abeta1-15 is less immunogenic than Abeta1-40/42 for intranasal immunization of wild-type mice but may be effective for “boosting”. Vaccine 2003;21(17–18):2197–206. [33] McLaurin J, Cecal R, Kierstead ME, et al. Therapeutically effective antibodies against amyloid-beta peptide target amyloid-beta residues 4–10 and inhibit cytotoxicity and fibrillogenesis. Nat Med 2002;8(11):1263–9; Spooner ET, Desai RV, Mori C, Leverone JF, Lemere CA. The generation and characterization of potentially therapeutic Abeta antibodies in mice: differences according to strain and immunization protocol. Vaccine 2002;21(3–4):290–7. [34] Town T, Tan J, Sansone N, Obregon D, Klein T, Mullan M. Characterization of murine immunoglobulin G antibodies against human amyloid-beta1-42. Neurosci Lett 2001;307(2):101–4. [35] Lemere CA, Spooner ET, Leverone JF, Mori C, Clements JD. Intranasal immunotherapy for the treatment of Alzheimer’s disease: Escherichia coli LT and LT(R192G) as mucosal adjuvants. Neurobiol Aging 2002;23(6):991–1000. [36] Monsonego A, Maron R, Zota V, Selkoe DJ, Weiner HL. Immune hyporesponsiveness to amyloid beta-peptide in amyloid precursor protein transgenic mice: implications for the pathogenesis and treatment of Alzheimer’s disease. Proc Natl Acad Sci USA 2001;98(18):10273–8.