Journal of Neuroimmunology 236 (2011) 27–38

Contents lists available at ScienceDirect

Journal of Neuroimmunology

j ourna l homepage: www.e lsev ie r.com/ locate / jneuro im

A mimotope peptide of Aβ42 fibril-specific antibodies with Aβ42 fibrillationinhibitory activity induces anti-Aβ42 conformer antibody response by a displayedform on an M13 phage in mice

Koichi Tanaka, Masaaki Nishimura, Yuya Yamaguchi, Shuhei Hashiguchi, Sho Takiguchi, Makoto Yamaguchi,Haruna Tahara, Takuma Gotanda, Risa Abe, Yuji Ito, Kazuhisa Sugimura ⁎Department of Chemistry, Biotechnology and Chemical Engineering, Graduate School of Science and Engineering, Kagoshima University, Kagoshima 890-0065, Japan

Abbreviations: AD, Alzheimer's disease; Aβ, amyloid β pcontaining 40 amino acids; Aβ42, amyloid β protein conrecombinant human prion protein; scFv, single-chain vphosphatase; HRP, horseradish peroxidase; IgG, immunhexafluoro-2-propanol; DMSO, dimethyl sulfoxide; PBS, psodiumdodecyl sulfate; ThT, thioflavinT;RT, roomtemperatELISA, enzyme-linked immunosorbent assay; SPR, surface punit; HPLC, high-performance liquid chromatography; TEMcopy; WST-8 2-(2-methoxy-4-nitrophenyl)-3-(4-nitrophetetrazolium, monosodium salt; FBS, fetal bovine serumlipopolysaccharide; MyD88, myeloid differentiation factosubcutaneous.⁎ Corresponding author at: 1-21-40 Korimoto, Kagoshim

285 8345; fax: +81 99 258 4706.E-mail address: kazu@be.kagoshima-u.ac.jp (K. Sugi

0165-5728/$ – see front matter © 2011 Elsevier B.V. Aldoi:10.1016/j.jneuroim.2011.04.010

a b s t r a c t

a r t i c l e i n f o

Article history:Received 5 October 2010Received in revised form 22 April 2011Accepted 24 April 2011

Keywords:Amyloid betaMimotope peptideHuman antibodyPeptide mimicsPhage display libraryVaccine

In Alzheimer's disease (AD), amyloid-β (Aβ) peptides accumulate in the brain in different forms, includingfibrils and oligomers. Recently, we established three distinct conformation-dependent human single-chain Fv(scFv) antibodies, including B6 scFv, which bound to Aβ42 fibril but not to soluble-form Aβ, inhibiting Aβ42fibril formation.In this study, we determined the mimotopes of these antibodies and found a common mimotope sequence,B6-C15, using the Ph.D.-C7C phage library. The B6-C15 showed weak homology to the C-terminus of Aβ42containing GXXXG dimerization motifs. We synthesized the peptide of B6-C15 fused with biotinylated TAT atthe N-terminus (TAT-B6-C15) and characterized its biochemical features on an Aβ42-fibrillation reaction invitro. We demonstrated that, first, TAT-B6-C15 inhibited Aβ42 fibril formation; secondly, TAT-B6-C15 boundto prefibril Aβ42 oligomers but not to monomers, trimers, tetramers, fibrils, or ultrasonicated fragments;thirdly, TAT-B6-C15 inhibited Aβ42-induced cytotoxicity against human SH-SY5Y neuroblastoma cells; and,fourthly, whenmice were administered B6-C15-phages dissolved in phosphate-buffered saline, the anti-Aβ42conformer IgG antibody response was induced.These results suggested that the B6-C15 peptide might provide unique opportunities to analyze the Aβ42fibrillation pathway and develop a vaccine vehicle for Alzheimer's disease.

rotein; Aβ40, amyloid β proteintaining 42 amino acids; rhPrP,ariable fragment; AP, alkalineoglobulin G; HFIP, 1,1,1,3,3,3-hosphate-buffered saline; SDS,ure;BSA,bovine serumalbumin;lasmon resonance; RU, response, transmission electron micros-nyl)-5-(2,4-disulfophenyl)-2 H-; TLR, Toll-like receptor; LPS,r 88; i.p., intraperitoneal; s.c.,

a 890-0065, Japan. Tel.:+81 99

mura).

l rights reserved.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

An antibody defines mimotopes that represent epitopes of aconformational structure. We previously reported a monoclonalantibody-assisted molecular molding design demonstrating that thepeptide phage library clones selected by a conformational epitope-recognizing and inhibitory monoclonal antibody might display

moieties that mimic a receptor/ligand-like three-dimensional struc-ture (Fukumoto et al., 1998).

We recently established human scFvs that bind to the Aβ42 fibrilbut not to the soluble form (Yoshihara et al., 2008). According to ourstrategy of molecular molding design, it is possible to isolate a B-cellmimotope but not a T-cell epitope peptide from a peptide-displayingphage library employing anti-Aβ42 fibril-specific scFvs as a mold.

In a transgenic mouse model of Alzheimer's disease (AD), both activeand passive therapeutic immunizations prevented or resolved severalprogressive AD-like pathologies in conjunctionwith the disappearance ofAβ42plaques in thebrain (Schenket al., 1999;Bardet al. 2000; Janus et al.,2000; Morgan et al., 2000; Bacskai et al., 2001; DeMattos et al., 2001). Inhumans, it has been reported that passive immunization using ahumanized anti-Aβ42 antibody showed positive cognitive effects atlower doses in clinical trials, although the disappearance of the Aβ42plaque showed no significant relation to the improvement of symptoms(Klyubin et al., 2008; Salloway et al., 2009). These results suggest thatAβ42-immunotherapy cannot reverse damage to neurons once the taupathology has proceeded. However, several studies showed that directinfusion of anti-Aβ42 IgG improved the cognitive functions of APP-transgenic mice (Dodart et al., 2002; Kotilinek et al., 2002; Wilcock et al.,2004;Billings et al., 2005;Hartmanet al., 2005; Lee et al., 2006;Oddoet al.,

28 K. Tanaka et al. / Journal of Neuroimmunology 236 (2011) 27–38

2006; Wilcock et al., 2006). Therefore, vaccine development is importantfor the prevention of the accumulation of abnormal Aβ42 conformersfrom their very early generation (Brody and Holtzman, 2008; Holtzman,2008). As long as the Aβ42 peptide is utilized as a vaccination antigen,T-cell immunity, including autoimmunity, cannot be fully avoided. B-cellmimotopepeptides permit thedevelopment of antigens that induceAβ42conformer-specific antibody responses but not an Aβ42-specific T-cellresponse (Fukumoto et al., 1998; Schneeberger et al., 2009).

In this study, we attempted to isolate and characterize this kind ofnovel synthetic peptide.We found a commonmimotope, B6-C15, sharedwith three Aβ42 fibril-specific scFvs. This peptide shows very weaksequence homology to Aβ24-30, which corresponds to the loop regionbetween the β1 and β2 strands at the C-terminus of Aβ42. We preparedthe synthetic peptide of B6-C15 flanked with the biotinylated TATpeptide (TAT-B6-C15) and performed the biochemical characterizationof this peptide. We showed that TAT-B6-C15 inhibits Aβ42 fibrillationand exhibits binding activity to prefibril Aβ42 oligomers but not tofreshly prepared monomers, trimers, tetramers, fibrils, or its sonicatedfragments. Furthermore, TAT-B6-C15 showednobindingactivity tootheraggregated proteins such as Aβ40 and Prion. We also showed TAT-B6-C15 inhibited Aβ42-induced cytotoxicity to SH-SY5Y cells. The TATpeptide did not show these inhibitory binding activities.

Recently, we demonstrated thatM13 phages remarkably induce anti-phage IgG production in primary response by a simple intraperitonealadministration of an M13 phage dissolved in PBS, whose response istotally dependent on MyD88 signaling (Hashiguchi et al., 2010).Therefore, we examined the immune response against the B6-C15phage in mice. We showed that the B6-C15 phage induces an anti-Aβ42IgG response even though a phage displayed only 5 copies of the B6-C15motifwhenmicewere administered i.p. or s.c.with aphosphate-bufferedsolution of the B6-C15-phage. Indeed, T-cell immunity directed to Aβ42was not observed.

Thus, B6-C15 may be useful to develop biochemical and immuno-logical therapeutic reagents for Alzheimer's disease.

2. Materials and methods

2.1. Reagents

Synthetic Aβ42peptides, biotinylatedTAT-B6-C15, a serinederivative(TAT-B6-S15) and TAT were purchased from the Peptide Institute, Inc.(Osaka). Anti-Aβfibril scFv (B6)was established as described (Yoshiharaet al., 2008). The peptide-displaying phage libraries (Ph.D.-12 and Ph.D.-C7C)were obtained fromNew England Biolabs (Beverly, MA). ThioflavinT and mouse anti-Aβ mAb (BAM-10) were from Sigma (St. Louis, MO).Mouse anti-Prion Protein mAb (SAF32) was purchased from SPI-Bio(Massy). Mouse anti-M13 mAb and anti-E tag mAb were from GEHealthcare (Piscataway, NJ). Alkaline-phosphatase (AP)- or horseradishperoxidase (HRP)-conjugated streptavidin was purchased from VectorLaboratories (Burlingame, CA), and AP- or HRP-conjugated goat anti-mouse IgGwas from Jackson ImmunoResearch (WestGrove, PA). Biotin-conjugated anti-M13 mAb and anti-mouse IgG were prepared asdescribed (Chyung and Selkoe, 2003) using Sulfo-NHS-biotin labelingkits (Pierce, Rockford, IL). Streptavidin colloidal gold (diameter ofcolloidal gold, 15 nm) was purchased from EY Laboratories (San Mateo,CA). Polystyrene 96-well microplates were obtained from Nunc(Roskilde), and 5–20% gradient gels were from the Atto Corporation(Tokyo). The PVDF and nitrocellulose filter were from MilliporeCorporation (Bedford, MA), and the luminol reagent (ECL) was fromGE Healthcare (Piscataway, NJ).

2.2. Mice

All animal experiments were performed in accordance with theagreement of the Genetic Modification Safety Committee and theguidelines of the Animal Care and Use Committee of Kagoshima

University. BALB/c (female, 6-wk old) and C57BL/6 mice (female, 6-wkold) were purchased from CLEA Japan, Inc. (Tokyo). All mice werehoused under specific pathogen-free conditions.

2.3. Preparation of Aβ conformers

Synthetic Aβ42 and Aβ40 peptides (purity 90–95% bymass spectrum,Peptide Institute Inc., Osaka)were solubilized in 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP, Wako, Tokyo) at a concentration of 1 mg/ml (222 μM)and separated into aliquots in microcentrifuge tubes. HFIP was removedunder vacuum in a Speed Vac, and the peptide filmwas stored at−80 °Casdescribed(Stineet al., 2003). Immediatelyprior touse, theHFIP-treatedAβwas dissolved in dimethyl sulfoxide (DMSO,Wako, Tokyo) to 4 mg/ml(888 μM) and diluted with a 20 mM phosphate buffer (pH 7.4) at aconcentration of 40 μM. Soluble Aβ was prepared immediately afterphosphate buffer dilution as described above. Fibrillar Aβ42 or Aβ40 wasprepared by incubating each Aβ at 37 C for 48 h or 192 h, respectively.Aβ42 fibril fragments were prepared by a sonication procedure asdescribed (Jan et al., 2008). The Aβ42 fibrils (40 μM) were diluted with a20 mMphosphate buffer at the concentration of 10 μMandmechanicallyfragmented into smallerfibrillar structures by ultrasonication on ice usingaVC-50 (BayerMedical, Tokyo) equippedwith a5-mmdiametermicrotip(20×5 s pulses; amplitude, 40%; output, 6W). Prefibril oligomers Aβ42and Aβ40 were periodically sampled at 1.5 h and 72 h after the onset offibrillation assay as Thioflavin T negative conformers, respectively.

2.4. Biopanning

Biopanning was performed as described (Gejima et al., 2002;Hashiguchi et al., 2003). Unrelated control scFv and anti-Aβ fibrilscFvs (B6, B7, D1 or F10) were separately immobilized on polystyrene96-well microplates (1 μg/100 μl/well). The Ph.D.-12 or Ph.D.-C7Cphage library (1.5×1011 plaque-forming unit [pfu]) was preabsorbedon a plate immobilized by unrelated control scFv to remove anyphages that were not specifically bound to anti-Aβ fibril scFv. Thephages were incubated at room temperature (RT) for 1 h withimmobilized anti-Aβ fibril scFv, and the binding phages were elutedwith 0.1 M glycine-HCl (pH 2.2). The eluate was immediatelyneutralized with 1.0 M Tris–HCl (pH 9.1) and amplified by infectingER2738. To select a more specific phage clone, the amplified eluatewas incubated with anti-fibril Aβ scFv containing 0.25% BSA for thesecond and third rounds of affinity selection. Phage clones wereisolated after the third round of affinity selection.

2.5. Elisa

ELISA was performed as described (Gejima et al., 2002; Hashiguchiet al., 2003). To determine the binding specificity of the peptide phageclone to B6 scFv, microplates were coated with B6 scFv or control scFv(40 ng/40 μl/well). Phage clones (40 μl) were added to the wells,followed by incubation with biotinylated anti-M13 mAb at a dilutionof 1:1000 and detection by AP-conjugated streptavidin at a dilution of1:1000. To determine the binding activity of synthetic peptide (TAT-B6-C15 or TAT-B6-S15) to B6 scFv, neutravidin (200 ng/40 μl/well:83 nM) was adsorbed on a microplate and TAT-B6-C15 or TAT-B6-S15(48 pg/40 ml/well: 300 nM) was added. After that, B6 scFv or controlscFv was added to the wells followed by incubation of anti-E tag mAb(1:1000) or BAM-10 (1:1000) and detection by AP-conjugated goatanti-mouse IgG at a dilution of 1:1000. Todetermine the binding activityof B6 scFv to Aβ conformers, each Aβ conformer (50 ng/40 μl/well) wasadsorbed on 96-well microplates and blockedwith 0.5% gelatin/PBS. B6scFv, control scFv, or BAM-10 (200 ng/40 μl/well) was added to thewells, followed by incubation of anti-E tag mAb at a dilution of 1:2000and detection by AP-conjugated goat anti-mouse IgG at a dilution of1:2000. To determine the anti-Aβ conformer antibody in sera, each Aβconformer (50 ng/40 μl/well) was coated on 96-well microplates and

29K. Tanaka et al. / Journal of Neuroimmunology 236 (2011) 27–38

blockedwith0.5% gelatin/PBS. Seradiluted 1:50were added to thewellsand detected by AP-conjugated goat anti-mouse IgG at a dilution of1:1000. To determine the anti-phage antibody or anti-B6-C15 antibodyin sera, a wild-type M13 phage (50 ng/40 μl/well) or TAT-B6-C15(50 ng/40 μl/well) was coated on 96-well microplates and blockedwith0.5% gelatin/PBS. Sera diluted 1:2000 were added to the wells anddetected by AP-conjugated goat anti-mouse IgG at a dilution of 1:1000.All ELISA assays were finally added 50 μl of a substrate containing2 mg/ml of p-nitrophenyl phosphate and 10% diethanolamine. Absor-bance was read at 405 nm by the microplate reader (NJ-2300, Nunc,Tokyo).

2.6. DNA sequencing

The inserted DNA was sequenced with a primer (5′-CCCTCATAGT-TAGCGTAACG-3′) by use of an ABI Genetic Analyzer PRISM 3100 (PEBiosystems, Foster City, CA) as described (Fukumoto et al., 1998).

2.7. Transmission electron microscopy (TEM)

Two μl of fibril Aβ42 was applied to a 0.6% polyvinyl formal-coatedsingle-hole copper grid. After 20 min, the solution was excluded with afilter paper, and the grid was washed with water. The grid was blockedwith 0.5% gelatin/PBS for 10 min, floated onto a droplet of anti-Aβantibody (BAM-10, 1:100) or B6-phage solution (1012 virion/ml) withPBS, and incubated for 60 min at RT. The BAM-10-treated grid waswashed with 0.1% Tween20/water, floated onto a droplet biotinylatedanti-mouse IgG solution, and incubated for 30 min at room temperature.The grid was washed with 0.1% Tween20/water and floated onto adroplet gold-labeled streptavidin solution. Then, itwaswashedwith 0.1%Tween20/water and stained with 0.1% phosphotungstic acid (Wako,Tokyo) for 90 s. The solution was excluded with a filter paper, and thegrid was visualized using an H-7100 transmission electron microscopeoperating at 75 kV (HITACHI, Tokyo).

2.8. Thioflavin T (ThT) assay

Amyloid fibril formation was followed by a fluorometric ThT assay(LeVine, 1993). For the fibril inhibition assay, an aliquot of the Aβ42stock solution was diluted to 40 μM in a 20 mM phosphate buffer inthe absence or presence of peptides (20 μM or 5 to 20 μM) andincubated at 37 C for 96 h. Aβ42was periodically removed and dilutedto a final concentration of 2 μM in a 20 mM phosphate buffer, pH 7.4,containing 10 μM ThT. After 2 min of incubation at room temperature,the fluorescence of the ThT dye was measured at an emissionwavelength of 482 nm using the excitation wavelength of 450 nm(Wallac 1420 ARVOsx, Wellesley, MA).

2.9. Immunoblotting analysis

Non-denaturedAβ42conformersweresubjected toSDS-PAGE(5–20%or 15% gels) and blotted to a PVDF filter using a semidry electroblotter(Sartorius, Tokyo). After blocking was carried out with 0.5% gelatin/PBS,the conformers were detected using a combination of BAM-10 (1:1000)with HRP-conjugated goat anti-mouse IgG (1:1000) using ECL on animage analyzer LAS-1000 (Fuji Film, Tokyo). The proteinmarker usedwasa broad-range prestained protein marker (New England Biolabs, Beverly,MA).

2.10. Dot blotting

Aβ conformers were generated several times during incubation afterthe onset of the fibrillation assay at 40 μM (181 μg/ml). Recombinanthuman prion protein (rhPrP) conformers were prepared as described(Luhrs et al., 2006). Briefly, 1 μl of the generated Aβ40, Aβ42(181 ng/1 μl), or rhPrP conformers (200 ng/1 μl) was hand-spotted

onto a nitrocellulose filter. After being blocked and washed, the filterwas probed with biotinylated peptides (200 nM: 0.9 μg/ml), scFvs(0.2 μg/ml), BAM-10 (0.2 μg/ml), or SAF32 (0.2 μg/ml). The filter wasthen incubated with the appropriate HRP-conjugated streptavidin(1:1000), anti-E tag antibody (1:1000), and/or HRP-conjugated anti-mouse antibody (1:1000). Blots were detected with ECL and LAS-1000.

2.11. Size-exclusion chromatography

The Aβ conformers generated during 1.5 h of incubation after theonset of the Aβ42 fibrillation assay (20 μg/110 μl: 40 μM) were run onSuperdex 75 prep-grade columns (GE Healthcare, Piscataway, NJ) andelutedwith a 20 mMphosphatebuffer. Shimadzu LC20high-performanceliquid chromatography (HPLC, Shimadzu, Kyoto) was used to collect1-minute fractions, which were immediately used for analyses.

2.12. Surface plasmon resonance (SPR) analysis

SPR analysis was performed as described (Yoshinaga et al., 2008).The binding activities of the TAT-B6-C15 peptide for Aβ42 weredetermined using the BIAcore X100 biosensor system (Biacore,Uppsala). The TAT-B6-C15 peptide (7 μg/70 μl) was injected andcoupled with a CM5 sensor chip via the activated carboxyl groups onthe surface of the chip. The BIAcore amine coupling kit was usedaccording to the manufacturer's instructions. After the unoccupiedgroups were blocked with 1.0 mol/l ethanolamine-HCl (pH 8.5), eachAβ conformer solution was injected onto the sensor chip at a flow rateof 30 μl/min at 25 C by using a PBS running buffer. We thenmonitoredboth the association and dissociation reactions. The sensorgramswererepresented as response units (RU) obtained by examining non-specific binding to the reference cell.

2.13. Cytotoxicity assay (WST assay)

A cytotoxicity assay was performed as described (Kayed et al.,2003). Human SH-SY5Y neuroblastoma cells were maintained inDMEM with 10% heat-inactivated FBS, penicillin (100 U/ml), andstreptomycin (100 μg/ml) in 5% CO2 at 37 °C. Cells were plated in 96-well plates at 1×104/100 μl/well in duplicate and differentiated inserum-free DMEMmediumwith N2 supplement (Invitrogen, CA) and10 μMall-trans retinoic acid for 72 h in a humidified incubator (e.g., at37 C, 5% CO2). Culture media were aspirated, and 75 μl of Ham's F 12serum-free media was added. Then, 25 μl of Aβ42 or control sampleswere added at a final concentration of 10 μMof Aβ42 and 5 μMof TAT-B6-C15 or 5 μM of TAT. Plates were incubated for 18 h at 37 C.Afterwards, 10 μl of Cell Counting Kit-8 (Dojindo, Kumamoto) wasadded in each well and incubated for 4 h. Living cells were measuredas the absorbance at 450 nm of reduced WST-8 (2-(2-methoxy-4-nitrophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium,monosodium salt).

2.14. Immunization

M13 phages were partially purified from the culture supernatant bytwo polyethylene glycol (PEG) precipitations, and then PEG-purifiedphages were further purified with CsCl using an ultracentrifuge asdescribed (Smith and Scott, 1993). One μg of the M13 phagescorresponds to 3.7×1010 virion (Li et al., 2010). LPS contamination wasestimated using ToxinSensor Chromagenic LAL Endotoxin assay kit(GenScript, Piscataway, NJ). Mice were administered i.p. or s.c. withvarying doses of phages dissolved in PBS (100 μl/mouse). Sera werecollected 10 days after the final immunization. Bloodwas taken from thelateral tail vein, and sera were stored at−20 C until use.

30 K. Tanaka et al. / Journal of Neuroimmunology 236 (2011) 27–38

2.15. T-cell proliferation assay

A T-cell proliferation assay was performed as described (Fukumotoet al., 1998). Mice were administered s.c. with a B6-C15 phage(30 μg/200 μl) with an alum (2 mg/200 μl) solution (200 μl/mouse) atdays 0, 14, 28, and 42. Five weeks after the final immunization,splenocytes (1.5x105/100 μl/well) were cultured in triplicate using a 96-wellflat-bottomplate (Iwaki Glass, Tokyo) for 72 h. Cellswere stimulatedwith a B6-C15 phage (ranging from0.02 to 5 μg/100 μl/well), aWT phage(0.02 to 5 μg/100 μl/well), Aβ42 (0.08 to 6 μg/100 μl/well), or TAT-B6-C15(0.08 to 20 μg/100 μl/well). Cells were pulsed with [3H] thymidine(0.5 μCi, Amersham, St. Louis, MO) for the last 18 h. They were harvestedusing an LM-101 Labo Mash (Labo Science Co.) harvester, and the [3H]thymidine uptake (c.p.m.) was counted on a MicroBeta 1450 Triluxscintillation counter (Wallac, Turku, Finland).

3. Results

3.1. Mimotopes of fibril Aβ-specific scFvs

We have established three human scFvs that bind to the fibril Aβ42but not to soluble forms of Aβ42 (Yoshihara et al., 2008). To determinethe binding motifs of these scFvs, random peptide-displaying phagelibraries, Ph.D.-12 and Ph.D.-C7C, were used according to the molecularmolding method as described previously (Fig. 1a, Ref. Fukumoto et al.,1998). We randomly selected 8 to 15 phage clones and analyzed theirmotif gene sequences. Results are summarized in Fig. 1b.

In our previous report (Yoshihara et al., 2008), a linear bindingmotif of B6 scFv selected from Ph.D.-12 showed weak homology toAβ31-36 at the β2 strands of the C-terminus of Aβ42 (Fig. 1c). In thecase of Ph.D.-C7C, all eight clones selected by B6 scFv displayed onlythe B6-C15 motif (Fig. 1b-d). B6-C15 showed weak homology to theC-terminus (23–31) of Aβ42, which includes the first GXXXGdimerizing motif of an amyloid precursor protein (APP) andcorresponds to the loop region between the β1 and β2 strands atthe C-terminus of Aβ42 (Fig. 1c, Ref. Harmeier et al., 2009; Luhrs et al.,2005; Munter et al., 2007). Interestingly, this mimotope was alsodetected from two other scFv clones, B7 (B7-C15) and D1 (D1-C5).When compared with other motifs derived from Ph.D.-C7C, G atposition 2 and PWM at position 5–7 are highly conserved residues inthe B6-C15 mimotope (Fig. 1b), suggesting their particular impor-tance in reference to the antibody specificity to Aβ42 fibril.

3.2. Transmission electron microscopy (TEM) of Aβ42 fibril bound withB6 scFv-phages

As an antibody-displaying phage is also visible in TEM, weobserved the binding complex of an Aβ42 fibril bound with B6 scFv-phages by TEM. When the Aβ42 fibril was stained with BAM-10(murine anti-N-terminus of the Aβ42 antibody) in conjunction withbiotinylated anti-mouse IgG and streptavidin colloidal gold, a numberof antibody-binding stains were visible along the side of the Aβ42fibril (Fig. 2a). In contrast, as shown in Fig. 2b, B6 scFv-phages do notbind alongside the Aβ42 fibril, but, rather, they bind to ends or gapsthat are scarcely observed in a fibril. To characterize the biochemicalfeature of the B6 scFv-binding motif on the Aβ42-fibrillation reaction,we chemically synthesized and characterized B6-C15.

3.3. Chemical synthesis of B6-C15 mimotope

As B6-C15 is a disulfide-bonded mimotope peptide, the syntheticB6-C15 motif needs to be flanked with short peptides at both ends.MOE analyses suggested that a spacer with longer length was neededto avoid steric interference on the binding of a biotinylated peptide tostreptavidin (Supplementary Fig. 1). Therefore, we prepared thesynthetic peptides by conjugating the biotinylated 11mer amino-acid

peptide, TAT at the N-terminus, which might act as a spacer betweenbiotin and B6-C15 moiety, and also confer hydrophilicity as well asprovide permeability of blood–brain barrier (Fig. 3a). Disulfidebonding was confirmed by the mass spectrum. To examine theintegrity of the B6-C15motif, we tested the binding activity of B6 scFvto TAT-B6-C15 immobilized on an ELISA plate (Fig. 3b). B6 scFv clearlybound to TAT-conjugated B6-C15. To test the role of the circular formof the motif, B6-S15, which was a serine derivative in place ofcysteines, was also synthesized. As shown in Fig. 3b, B6 scFv exhibitedbinding activity to TAT-B6-S15, while it showed no binding activity tobiotinylated TAT. Unrelated scFv showed binding activity to neitherbiotinylated TAT-B6-C15, TAT-B6-S15, nor TAT. These results indicat-ed that the disulfide bonding was not essential for B6 scFv-bindingspecificity. For further characterization of the B6-mimotope, TAT-B6-C15, but not its serine derivative, was studied because it exhibitedrelative stability in blood at a half-life of 18 h.

3.4. TAT-B6-C15 inhibits the fibrillation of Aβ42

As B6-C15 was a mimotope shared with three anti-Aβ42 fibril scFvs,we examined the influence of the presence of TAT-B6-C15 on Aβ42fibrillation. As shown in Fig. 4a, TAT-B6-C15 showed a marked inhibitoryeffect on Aβ42 fibril formation. An unrelated peptide, Cryj2, did not showan inhibitory effect. TAT did not affect Aβ42 fibrillation, which wasconfirmed by TEM (Supplementary Fig. 2). TAT-B6-C15 inhibited Aβ42fibril formation in adose-dependentmanner (Fig.4b). Todirectly visualizethedegreeofAβ42fibrillation, anAβ42 solutionwasperiodically sampledalong the time course of Aβ42 fibrillation, and an immunoblotting assaywas performed using an anti-pan Aβ42murine antibody, BAM-10, whichrecognizes the N-terminal epitope irrespectively of any conformers ofAβ42 (Fig. 4c). At 0 h, freshly prepared Aβ42 peptides were resolved intothree bands: themonomer, trimer, and tetramer of Aβ42 because of veryquick assembly reaction. Dimers were not visible, which was differentobservation from human in vivo results (Klyubin et al., 2008; Shankaret al., 2008). Six hr later in the absence of peptides, Aβ42 oligomers with30–200 kDa appeared, while bands of monomers, trimers, and tetramersdisappeared. Twenty-four hr later, themajority of Aβ42 conformers couldnot enter the separation gel stacking in a top well. Aβ42 oligomers werehardly observed at 96 h, since Aβ42 totally converted to Aβ42 fibrils.Similar patterns were observed in the presence of TAT or an unrelatedpeptide, Cryj2. In contrast, when Aβ42 fibrillation proceeded to in thepresence of TAT-B6-C15, Aβ42 oligomerswith a broad range ofmolecularweights were visible at 6 h and retained this pattern even 96 h later.

3.5. TAT-B6-C15 binds to prefibril Aβ42 conformers but not to Aβ42monomers or Aβ42 fibrils

To determinewhich Aβ42 conformerwas targetedwith TAT-B6-C15,we performed dot hybridization, in which various Aβ conformersperiodically sampled after the onset of Aβ42 fibrillation were hand-spotted and incubated with biotinylated TAT-B6-C15, biotinylated TAT,BAM-10 IgG, or B6 scFv. As shown in Fig. 5a, BAM-10 detected everysample tested at various incubation times, indicating the existence of anequal amount of Aβ42 at each spot. When TAT-B6-C15was reacted witha replicated membrane filter, it showed binding activity to the Aβconformers generated 1.5 and 3 h after the onset of Aβ fibril formation,whereas it did not bind to freshly prepared Aβ42 containing amonomer,trimer, or tetramer. Then, TAT-B6-C15 gradually lost its binding activityto Aβ conformers generated later than 3 h (see at 48 h, Fig. 5a). Thisfinding was unexpected because TAT-B6-C15 was a B cell mimotopepeptide of the anti-Aβ42fibril scFv, B6. TAT showednobinding activity toany spot.

The lack of binding of TAT-B6-C15 to fibrils appeared to be due to thelow concentration of fibril ends, which decreased along with fibrilelongation. Therefore, to examinewhether TAT-B6-C15 binds to the fibrilend, we performed dot hybridization using the same amount of

b

c

FrequencyClone numberPeptide library ScFv Sequence

aAß42 fibril

Specific phage motif Mold (B6 scFv)

Aß42 fibril mimotope peptide

anti Aß42 fibril (B6)

Abs

orba

nce

@ 4

05nm

d

0.0

0.2

0.4

0.6

0.8

1.0

1.2

B6-C15

B6-C2

B6-C5

B6-C6

B6-C7

B6-C8

B6-C9

B6-C10

B6 scFvcontrol scFvgelatin

Fig. 1.Mimotopes of Aβ42 fibril-specific scFvs. a, Molecular molding method: peptide-phage clones selected by an inhibitory monoclonal antibody display moieties that mimic theirbinding structures and exhibit agonistic or antagonistic activity (Fukumoto et al., 1998). b, Mimotope sequences of anti-fibril Aβ42 scFvs (B6, B7, and D1) were identified with arandom peptide-displaying phage library (Ph.D.-12 or Ph.D.-C7C). Frequency: The number of mimotope-positive clones/total clones tested. c, Sequence homology of the B6-C15, B6-L1, or B6-L10 peptide motif to the Aβ42 sequence. B6-C15 was identified as a common epitope of B6, B7, and D1 scFv. Analysis was performed using Clustal W ver. 3.1. The markindicated below the epitope sequence, “*,” “:,” or “.,” indicates identical, conserved, or semi-conserved residues, respectively. d, Biopanning of peptide phages by B6 scFv. Eight B6scFv-specific clones were selected by biopanning from the Ph.D.-C7C library (1.5×1011 pfu). Phages were reacted to B6 scFv or unrelated scFv-coated ELISA plates and detected withbiotinylated anti-M13 IgG and AP-conjugated streptavidin.

31K. Tanaka et al. / Journal of Neuroimmunology 236 (2011) 27–38

ultrasonicatedfibril fragments.Weconfirmed the fragmentedfibrilsusingTEM image analysis before use. As shown in Fig. 5b, BAM-10 recognizedthe monomer, prefibril oligomers, fibrils, or ultrasonicated fibril frag-

ments. In contrast, TAT-B6-C15 showed marked binding activity toprefibril oligomers but very faint spots to fibrils and sonicated fibrilfragments. As the concentration of fibril ends was much higher in

ba

B6-scFv phage

Aββ42 fibril

Fig. 2. Transmission electron microscopy (TEM) of Aβ42 fibrils bound with B6 scFv-phages. a, Aβ42 fibrils were stained with the mouse anti-Aβ antibody (BAM-10). The reactedantibodies were visualized with biotinylated anti-mouse IgG and streptavidin colloidal gold. b, Aβ42 fibrils were stained with B6 scFv-phage. B6 scFv-phage was reacted to an Aβ42fibril-coated grid and stained with phosphotungstic acid. The scale bar is 100 nm.

32 K. Tanaka et al. / Journal of Neuroimmunology 236 (2011) 27–38

ultrasonicatedfibril fragments than thatoffibrils in these spots, these faintbindings appeared to be negligible and nonspecific. These resultsindicated that TAT-B6-C15 did not bind to Aβ42 fibrils.

Aβ42 oligomers shown in Fig. 4c contained various Aβ42 species.Therefore, to examine the target molecular species of TAT-B6-C15 moreprecisely, Aβ42 conformers generated 1.5 h after the onset of Aβfibrillation were separated into two fractions, i.e., high-molecularoligomers (A)with approximately 200 kDa and low-molecular oligomerscontainingmonomers, trimers, and tetramers (B) with around 12 kDa bySuperdex 75 (Fig. 5c).

SPR analysiswas performed to confirm the results of dot hybridization(Fig. 5b). This experiment indicated the qualitative rather than thequantitative features of their binding reaction, since Aβ42 conformers areheterogeneous in molecular species. TAT-B6-C15 was chemically immo-

Abs

orba

nce

@ 4

05 n

m

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7 B6 scFvcontrol scFvgelatin

TAT-B6-C15

TAT-B6-S15 TAT

Peptide Amino acid sequen

a

b

Fig. 3. Synthetic B6-C15 mimotope peptides. a, Synthetic peptides. B6-C15 is a circular mimB6-S15 is a serine derivative of B6-C15. Biotin was conjugated to the N-terminus of the peptisequence (Wadia and Dowdy, 2005). b, Both TAT-B6-C15 and TAT-B6-S15 are recognizedplates. B6 scFv (1 μg/40 μl/ well) was added to each well. This binding was detected by ant

bilized on a CM5 chip.When fraction Awas analyzed on SDS-PAGE, Aβ42conformers were resolved in tetramers and monomers in addition tohigh-molecular oligomers (inset of Fig. 5c), indicating the existence ofboth SDS-stable and unstable Aβ42 conformers in fraction A. In contrast,fractionBcontainedmainly tetramers in addition tomonomers. As shownin Fig. 5d, fraction A strongly bound to TAT-B6-C15, while fraction Bshowed no binding activity. In contrast, Aβ42 fibrils and ultrasonicatedfibrils showed no binding activity to TAT-B6-C15 at all.

3.6. TAT-B6-C15 specifically binds to Aβ42 prefibril oligomers but not toother aggregated proteins

To determine the binding feature of TAT-B6-C15 to other aggregatingproteins, Aβ40 fibrils and recombinant human prion protein (rhPrP)

ce Characteristic

otope peptide. Disulfide bonding of B6-C15 was confirmed by mass-spectrum analysis.de using a spacer of 11 mer amino acids. The spacer sequence was derived from the TATwith B6 scFv. Biotinylated TAT-B6-C15 was immobilized on neutravidin-coated ELISAi-E tag IgG and AP-conjugated anti-mouse IgG.

c0 6 24 96

buffer

0 6 24 96

TAT-B6-C15

0 6 24 96

TAT

0 6 24 96

Cryj2

6.516.5

25

32.5

47.5

62

83

175

Marker

hour

a

b

fibril

trimertetramer

oligomer

monomer

0 25 50 75 100 1250

500

1000

1500

2000

2500

3000

3500

TAT-B6-C15TATCryj2buffer

Flu

ores

cenc

e in

tens

ity

Time (hour)

0 5 10 200

1000

2000

3000

4000

5000

Flu

ores

cenc

e in

tens

ity

TAT-B6-C15 dose (µM)

Fig. 4. TAT-B6-C15 inhibits Aβ42 fibrillation. a, TAT-B6-C15 inhibits Aβ42 fibrillation.Aβ42 (40 μM) was incubated for 96 h, as described in Experimental Procedures, in theabsence or presence of various peptides (20 μM). Aβ42 fibrillation was monitored bythe ThT assay. Cryj2 was used as an unrelated peptide. b, Dose-response of theinhibitory activity of TAT-B6-C15 on Aβ42 fibrillation. The TAT-B6-C15 peptide wastested in the range from 0 to 20 μMunder identical conditions to those in panel (a). Thedata was at 96 h. c, SDS-PAGE of Aβ42 conformers. Aβ42 fibrillation was performed inthe absence or presence of TAT, Cryj2 or TAT-B6-C15 as described in panel (a).Immunoblotting of SDS-PAGE (5–20%) was performed using BAM-10 and HRP-conjugated anti-mouse IgG.

33K. Tanaka et al. / Journal of Neuroimmunology 236 (2011) 27–38

conformers were examined. As B6 scFv bound to Aβ42 but not Aβ40fibrils (Yoshihara et al., 2008), we examined the binding specificity ofTAT-B6-C15 tovariousAβ42orAβ40 conformers. As shown in Fig. 6, bothAβ42 and Aβ40 were recognized by BAM-10 regardless of being in theprefibril oligomer orfibril form. In contrast, TAT-B6-C15bound to neitherAβ40prefibril oligomersnorAβ40fibrils,while it bound toAβ42prefibriloligomers but not to Aβ42 fibrils. In the cases of the prion protein, boththe monomer and fibril forms of rhPrP were recognized with murineanti-rhPrP IgG, SAF32. On the other hand, TAT-B6-C15 showed nobinding activity to rhPrP conformers, indicating that the binding is notdue to the aggregated nature of proteins.

3.7. TAT-B6-C15 inhibits Aβ42-induced cytotoxicity to SH-SY5Y cell line

As primary toxic species, soluble Aβ42 oligomers are better knownthan fibrils; therefore, we examined the effect of TAT-B6-C15 onAβ42-induced cytotoxicity assay using a human neuroblastoma cellline, SH-SY5Y. Cells treated with all-trans retinoic acid for 72 h werecultured for 18 h with or without soluble Aβ42 in the presence orabsence of TAT-B6-C15 or TAT peptide. Cell viability was thenexamined (Fig. 7). In the absence of soluble Aβ42, TAT-B6-C15 hadno significant effect on cell viability when compared with cells alone,whereas TAT appeared to promote cell viability marginally. Whencells were cultured in the presence of Aβ42, cell viability wassignificantly reduced to 39%. In this condition, the addition of TAT-B6-C15 markedly recovered the cell viability to the control level. Theaddition of TAT had no influence on Aβ42-induced cytotoxicity. In thisassay, fibril forms of Aβ42 did not influence the cell viability asdescribed (Yoshihara et al., 2008). These results suggested that TAT-B6-C15 bound to soluble Aβ42 oligomers, resulting in the conversionof toxic oligomers to non-toxic oligomers (Ladiwala et al., 2010).

3.8. B6-C15 phage induces an anti-Aβ42 conformer antibody response inmice

Biochemical characterization of the B6-C15 motif suggested thatB6-C15 might mimic an Aβ42 tertiary structure. Therefore, weexamined the immune response elicited with a B6-C15 phage inmice. Phages were purified by CsCl-ultracentrifugation. LPS contam-ination was estimated to be 0.15 EU/ml at 1011 virion/ml using anEndotoxin assay kit. Mice were simply administered i.p. or s.c. with aB6-C15 phage in PBS solution. Ten days after the last injection, serawere collected. When BALB/c mice were immunized i.p. with 1 μg ofthe B6-C15 phage, they produced a significant level of anti-Aβ42 fibrilIgG antibodies as well as anti-B6-C15 peptide antibodies 10 days afterthe last injection although a phage displays only 5 copies of the B6-C15 motif. The wild-type M13 phage showed background responses.Using BALB/cmice immunized s.c. with a B6-C15 phage plus alum fourtimes, T-cell proliferation assays were performed against the B6-C15peptidemotif, soluble Aβ42, or M13 phage. As shown in Fig. 8b, no B6-C15-reactive or soluble Aβ42-reactive cell proliferation was observed,while the M13 phage strongly induced cell proliferation. As we usedMyD88-knockout mice having a C57BL/6 genetic background inanother study (Hashiguchi et al., 2010), we examined the antibodyresponses of C57BL/6 as well as BALB/c mice against B6-C15 phageimmunization. The antibody responses of C57BL/6 were relativelylower than those of BALB/c mice, and s.c. immunization of the B6-C15phage induced a response comparable to i.p. immunization (Fig. 8c).Therefore, we examined the anti-Aβ42 conformer IgG response by s.c.injection with varying immunizing doses of the B6-C15 phage. Asshown in Fig. 8c, mice produced IgG antibodies reactive to Aβ42 fibril,prefibril Aβ42 oligomers, or soluble Aβ42 in an immunizing dose-dependent manner. They showed a background response to gelatin asa control protein. When the anti-B6-C15 peptide antibody or anti-M13 phage IgG was examined in these mice, the results showed thatthey produced much higher levels of the B6-C15 peptide-reactiveantibodies than the anti-Aβ42 conformer antibodies (Fig. 8d).Regarding anti-M13 phage antibodies, the B6-C15 phage induced astrong IgG response comparable to that of the wild phage.

4. Discussion

We established three human scFv antibodies, including B6 scFv,with specific binding activity to Aβ42 fibrils but not to freshlydissolved Aβ42 (Yoshihara et al., 2008). Epitope mapping of B6 scFvwas previously carried out by an epitope-defined antibody competi-tion assay and using Ph.D.-12 library displaying random linear 12

a

mon

omer

olig

omer

fibr

il

BAM-1

0

TAT-B6-

C15

TAT

fibr

il f

ragm

ent

(Son

icat

ed f

ibri

l)

b

0

1.5

3

6

24

48

Aß

conf

orm

ers

gene

rate

d at

eac

h ti

me

(hou

r)

BAM-1

0

TAT-B6-

C15

TATB6 s

cFv

Contro

l scF

v

d

c

Time (sec.)

-5

0

5

10

15

20

25

30

0 50 100 150 200 250 300 350 400 450 500Time (sec.)

0 50 100 150 200 250 300 350 400 450 500

Time (sec.)0 50 100 150 200 250 300 350 400 450 500

Time (sec.)0 50 100 150 200 250 300 350 400 450 500

93.5 nM

23.4 nM 46.8 nM

Frac. A (High molecular oligomer)

Res

pons

e (R

U)

-5

0

5

10

15

20

25

30

Res

pons

e (R

U)

-5

0

5

10

15

20

25

30

Res

pons

e (R

U)

-25

-15

-5

5

15

25 55.4 nM 110.8 nM Frac. B (Low molecular oligomer)

Res

pons

e (R

U)

250 nM

62.5 nM 125 nM

Fibril fragment (Sonicated fibril)

250 nM

62.5 nM 125 nM

Fibril

0

5

10

15

20 30 40 50

Mea

n A

bso

rban

ce @

257

nm Prefibril oligomer

Marker

670 kDa

158 kDa

44 kDa

1.35 kDa

17 kDa

A B

Time (min.)

A B

monomer

tetramer

highmolecularoligomer

Fig. 5. TAT-B6-C15 binds to prefibril Aβ42 oligomers but not to monomers or fibrils. a, Immunoblotting of Aβ42 conformers. Aβ42 (40 μM)was incubated as described in Fig. 4a. Thegenerated Aβ42 conformers were periodically hand-spotted (1 μl/spot) onto a nitrocellulose filter followed by incubation with detection probes such as TAT-B6-C15 (200 nM), TAT(200 nM), B6 scFv (0.2 μg/ml), or BAM-10 (0.2 μg/ml). These probes were detected by HRP-conjugated streptavidin, anti-E-tag, and / or HRP-conjugated anti-mouse IgG. b, TAT-B6-C15 does not bind to monomers, fibrils, or ultrasonicated fibrils of Aβ42. Ultrasonicated fibril fragments were prepared as described in Experimental Procedures. c, Size-exclusionchromatography profile of Aβ42 conformers. Aβ42 (40 μM) was incubated for 1.5 h and then applied on a Superdex 75 prep-grade column. The inset indicates the SDS-PAGE (15%)patterns of fraction A or B. d, SPR analysis of Aβ42 conformers. TAT-B6-C15 was chemically immobilized on a CM5 chip.

34 K. Tanaka et al. / Journal of Neuroimmunology 236 (2011) 27–38

amino acid sequences, demonstrating that the B6 scFv epitope locatesat the C-terminus of Aβ42 (Fig. 1c, Ref. Yoshihara et al., 2008).

In this paper, we have characterized the mimotopes of these Aβ42fibril-specific scFvs using a cysteine-constrained peptide motif-displaying phage library, Ph.D.-C7C.

Three distinct Aβ42-specific scFvs (B6, B7, and D1) havesignificantly similar motifs including a common mimotope peptide,B6-C15, shared with these antibodies. Comparing every motif, G atposition 2 and PWM at position 5–7 are highly conserved in thesemotifs. Homology analysis showed very weak homology to the

Fig. 6. TAT-B6-C15 specifically binds to Aβ42 prefibril oligomers but not to other aggregated proteins. Aβ42, Aβ40, or rhPrP conformers were prepared as described in ExperimentalProcedures. The conformers were periodically hand-spotted (1 μl/spot) onto a nitrocellulose filter followed by incubation with TAT-B6-C15 (200 nM), BAM-10 (0.2 μg/ml), or SAF32(0.2 μg/ml). Detection was performed by HRP-conjugated streptavidin and/or HRP-conjugated anti-mouse IgG.

35K. Tanaka et al. / Journal of Neuroimmunology 236 (2011) 27–38

sequence of 23–31 of the C-terminus, which overlaps with the firstGXXXG dimerizing motif of APP, corresponding to the loop regionbetween β1 and β2 strands at the C-terminus of Aβ42.

To investigate the relationship of the B6-C15 mimotope withAβ42-fibril specificity sharedwith B6, B7, and D1 scFvs, we chemicallysynthesized the B6-C15mimotope. We first synthesized the disulfide-bonded B6-C15 mimotope, which was flanked with G at the C-terminal and biotinylated RR at the N-terminal. However, the biotinresidue of this peptide was not recognized by streptavidin. MOEanalyses suggested the need of a spacer due to steric hindrance of RRof B6-C15 on the binding of biotin to streptavidin (Supplementary

Cel

l via

bilit

y (%

con

trol

)

buffer

(Con

trol)TAT

TAT-B6-

C15

solu

ble Aß42

solu

ble Aß42

+ T

AT

solu

ble Aß42

+ T

AT-B6-

C15

Aß42 fi

bril

*

**

***

0

20

40

60

80

100

120

140

Fig. 7. TAT-B6-C15 inhibits Aβ42-mediated cytotoxicity. SH-SY5Y cells (1×104/well)treated with all-trans-retinoic acid were cultured with or without freshly prepared Aβ42(10 μM) in the presence or absence of TAT-B6-C15 or TAT (5 μM) for 18 h. Cell viability wasestimated by WST assay and expressed as % control: (the absorbance at 450 nm of viablecells of the experimental group/the absorbance at 450 nmof viable cells of the control groupof cells alone) x 100. Data represent mean±SEM of two independent experiments.Statistical analysis was performed by student t-test. *, pN0.4; **, pb0.001; ***, pb0.0001versus soluble Aβ42-treated group.

Fig. 1). Therefore, we inserted the TAT sequence between biotin andthe B6-C15 mimotope. This modification was expected to conferfunctions such as a spacer to avoid steric interference as well as toensure the hydrophilicity and permeability of the blood-brain barrierto this peptide. The functional influence of the TAT sequence in vivowas not examined in this study.

4.1. Biochemical characterization of B6-C15

B6 scFv recognized TAT-conjugated B6-C15 as well as its serinederivative, indicating that disulfide bonding is not essential. As shownin Fig. 4, TAT-B6-C15 inhibited Aβ42 fibrillation in a dose-dependentfashion.

To identify the target Aβ species of TAT-B6-C15, dot hybridizationand SPR analyses were performed. These analyses clearly showed thatTAT-B6-C15 bound to prefibril oligomers but not freshly dissolvedAβ42 species (Fig. 5a,d). TAT-B6-C15 did not show any bindingactivity to fibril Aβ42 in dot and SPR experiments (Fig. 5b,d). Thisfinding was unexpected because B6-C15 was defined to be amimotope of Aβ fibril-specific human B6 scFv (Yoshihara et al.,2008). To examine the possibility that the concentration of the ends offibril Aβ42 decreased by fibrillation, resulting in a level below thedetection limit by biotinylated TAT-B6-C15 plus streptavidin-HRP, weexamined the binding activity of ultrasonicated Aβ42 fibrils in dot blotand SPR analyses (Fig. 5b,d). In SPR, the binding step quicklyprogresses in second, whereas it takes hours in dot hybridization orELISA. Although this difference in experimental conditions may giverise to a background in ELISA or dot experiments, dot blothybridization showed an equally faint spot at both fibrils andultrasonicated fragmented fibrils despite the fact that the fragmentedfibrils displayed an overwhelming number of fibril ends relative tofibrils, indicating that these faint bindings were nonspecific. SPRanalysis clearly showed that TAT-B6-C15 did not have binding activityto fragmented fibrils. In these experiments, TAT-B6-C15 exhibitedmarked binding activity to Aβ42 prefibril oligomers. As shown inFig. 6, TAT-B6-C15 showed no binding activity to any Aβ40conformers, including monomers, oligomers, and fibrils. When Aβ40fibrillation was performed for 500 h in the presence of TAT-B6-C15, aslight delay in Aβ40 fibrillation was observed, but no inhibition wasobserved at 350 h, indicating the crucial interaction of the C-terminalamino acids of Aβ42 with TAT-B6-C15. Furthermore, TAT-B6-C15 didnot show any binding activity to the monomer or fibril form of rhPrP

d

0

0.5

1

1.5

2

Abs

orba

nce

@ 4

05nm

TAT-B6-C15 peptide WT phage gelatin

30 10 1 10Immunogen amount (µg)

B6-C15 phage WT phage

c

0.3

0.25

0.2

0.15

0.1

0.05

0

Abs

orba

nce

@ 4

05nm

3H-t

hym

idin

e in

corp

orat

ion

(cpm

)

Dose (µg / well)

0.01 0.1 1 10 1000

1000

2000

3000

Aß42

TAT-B6-C15

B6-C15 phage

WT phage

a b

B6-C15 phage WT phage

Immunogen amount (µg)

1 1

Aß42 fibrilTAT-B6-C15TAT gelatin

30 10 1 10100

0.02

0.04

0.06

0.08

0.1

Abs

orba

nce

@ 4

05nm

Immunogen amount (µg)

B6-C15 phage WT phage

i.p. s.c.

Aß42 fibrilPrefibril Aß42 oligomersoluble Aß42 gelatin

Fig. 8. B6-C15phage induces anti-Aβ42 conformer antibody responses inmice. a, BALB/cmice produced ananti-Aβ42fibril antibody by immunizing aB6-C15phage.Mice (3mice/group)wereadministered i.p. with 1 μg of the B6-C15 phage or 1 μg of theWT phage in PBS on days 0, 30, and 76. Ten days after the last injection, serum antibodies (diluted 1:50with PBS)were estimatedusing Aβ42 fibril-, TAT-B6-C15-, TAT-, or gelatin-coated ELISA plates. Results were represented as a mean value with SD. b, Splenocytes derived from B6-C15 phage-immunized BALB/c miceshowed no response to the B6-C15 peptide motif. Mice were immunized s.c. with 30 μg of the B6-C15 phage with alum on days 0, 14, 28, and 42. Five weeks after the last immunization,splenocyteswere stimulatedwith theB6-C15phage (ranging from0.02 to5 μg/well),WTphage (0.02 to5 μg/well), solubleAβ42 (0.08 to6 μg/well), or TAT-B6-C15 (0.08 to20 μg/well) for 72 handmonitored by the addition of 0.5 μCi [3H] thymidine as described in theMaterials andmethods section. c, C57BL/6mice produced an anti-Aβ42 conformer antibody response by i.p. or s.c.injection of the B6-C15 phage in an immunizing dose-dependent manner. Mice (3 mice/ group) were immunized i.p. with 10 μg or s.c. with 30, 10, or 1 μg of the B6-C15 phage or 10 μg of thewild-type (WT) phage in PBS on days 0, 16, 50, and 70. Ten days after the last immunization, anti-Aβ42 IgG in sera (diluted 1:50) was determined by Aβ42 fibril-, prefibril Aβ42 oligomers-,solubleAβ42-, orgelatin-coatedELISAplates. d,Anti-M13phageorB6-C15peptideantibody responseofmice shown inpanel c.Anti-M13phage-or TAT-B6-C15 IgG in sera (diluted1:2000)wasdetermined by the same procedure as that described above.

36 K. Tanaka et al. / Journal of Neuroimmunology 236 (2011) 27–38

(Fig. 6). Thus, TAT-B6-C15 shows fine specificity to the prefibriloligomer of Aβ42.

It is postulated that the β2 strand of the Aβ42 fibril forms acohesive end at the growing head and the Aβ42 fibril grows bycontinuous intermolecular interaction with the β2 strand to the β1strand of the adjacent Aβ42molecule (Luhrs et al., 2005). As shown inFig. 2b, the B6 scFv phage appears to bind to ends or gaps but notalongside Aβ42 fibrils. As B6-C15 appears to be a B6 mimotope withweak homology to Aβ23-31, it is conceivable that B6-C15 breaks intothis region, resulting in the termination of Aβ42 oligomer growth.

When HPLC-purified Aβ42 prefibril oligomers were analyzedunder SDS-PAGE, they were resolved in Aβ42 bands that consistedof SDS-unstable and SDS-stable species (Fig. 5c). The pathologicallyfunctional relationships between SDS-unstable and SDS-stable Aβ42oligomers are unclear. TAT-B6-C15 bound to prefibril Aβ42 oligomersbut not to Aβ42 fibrils (Fig. 5a,b,d), suggesting that TAT-B6-C15 mighthave been competitively accommodated to the open region of Aβ23-31 of prefibril Aβ42 oligomers against their fibril elongation whilehardly accessing this loop region of fibrils or fragmented fibrils

because the fibril structure was highly shielded by compact in-teractions between β2 and β1 strands. Furthermore, it is likely thatTAT-B6-C15 was hardly retained at the growing ends of Aβ42 fibrils inthe absence of accommodating structures for TAT-B6-C15.

4.2. Physiological and immunological characterizations of B6-C15

It is of note that TAT-B6-C15 inhibited soluble Aβ42-inducedcytotoxicity inWST assay using SH-SY5Y cells (Fig. 7), suggesting thatTAT-B6-C15 might have converted soluble Aβ42 into non-toxicconformers as off-pathway conformers (Ladiwala et al., 2010).

We recently demonstrated that mice show remarkable anti-M13phage IgG production in a primary response by an i.p. administrationof the phage in phosphate-buffered saline and that this antibodyresponse is completely MyD88-dependent. We observed no moresignificant influence in the anti-M13 phage antibody response ofTLR4−/− mice than in that of wild-type C57BL/6 mice, suggesting thatLPS contaminated in the CsCl-purified M13 phage preparation had noinfluence on the magnitude of the antibody responses (Hashiguchi et

37K. Tanaka et al. / Journal of Neuroimmunology 236 (2011) 27–38

al., 2010). As these knockout mice are generated from C57BL/6 genebackground mice, we presented the data of antibody responses inC57BL/6 as well as BALB/c mice (Fig. 8). As shown in Fig. 8a, wedemonstrated that, when mice were administered i.p. with a PBSsolution of B6-C15 phages, they produced significant levels of IgGantibody response directed to Aβ42 conformers although only 5copies of the B6-C15 motif were displayed on a phage. When thesesera were absorbed with a fibril Aβ42-coated plate, the supernatantstill showed the binding activity to the B6-C15 peptide, indicating thatthese binding activities could not be attributed to the multi-specificityof an identical antibody molecule (unpublished data). As shown inFig. 8c, mice immunized with the B6-C15phage produced anti-Aβ42fibril, prefibril oligomers, or soluble Aβ42 antibodies. Regarding thenature of these Aβ42-directed specificities, it remains to be examinedwhether this reactivity reflects cross-reactive antibodies or thebinding activities of antibodies with distinct specificity, whichdiscriminates fibril, prefibril, or soluble Aβ42 from one another.These results suggested that the g8p-display-based B6-C15 phagemight become more immunogenic (Solomon, 2007). Furthermore, itwas clearly demonstrated that mice immunized with the B6-C15phage showed no Aβ42-specific T-cell response (Fig. 8b).

Safety issues and the potential significance of bacteriophages fortherapeutic agents have been proposed (Frenkel and Solomon, 2002;Brussow, 2005; Kurzepa et al., 2009). In accordance with thesereports, no harmful effect of phage administration was confirmed bymonitoring the body weights and behaviors of mice in this study(unpublished data). The habitual presence of the M13 phage wasdemonstrated by detecting marked titers of anti-M13 IgG antibodiesin every serum collected from non-laboratory's individuals andcommercial human AB sera for cell culture (unpublished data).

Thus, we here showed the biochemical and immunological mimicnatures of the B6-C15 moiety to Aβ42 conformers, suggesting theusefulness of TAT-B6-C15 or B6-C15 phage for the development oftherapeutics for Alzheimer's disease.

Acknowledgments

We thank Dr. Tomoki Yoshihara for initial work on this study. Thiswork was supported in part by the Research for PromotingTechnological Seed of Japan Science and Technology Agency andGrant-in-aid for Scientific Research (B) 21310144 from Japan Societyfor the Promotion of Science.

Appendix A. Supplementary data

Supplementary data to this article can be found online atdoi:10.1016/j.jneuroim.2011.04.010.

References

Bacskai, B.J., Kajdasz, S.T., Christie, R.H., Carter, C., Games, D., Seubert, P., Schenk, D., Hyman,B.T., 2001. Imaging of amyloid-beta deposits in brains of living mice permits directobservation of clearance of plaques with immunotherapy. Nat. Med. 7, 369–372.

Bard, F., Cannon, C., Barbour, R., Burke, R.L., Games, D., Grajeda, H., Guido, T., Hu, K.,Huang, J., Johnson-Wood, K., Khan, K., Kholodenko, D., Lee, M., Lieberburg, I.,Motter, R., Nguyen, M., Soriano, F., Vasquez, N., Weiss, K., Welch, B., Seubert, P.,Schenk, D., Yednock, T., 2000. Peripherally administered antibodies against amyloidbeta-peptide enter the central nervous system and reduce pathology in a mousemodel of Alzheimer disease. Nat. Med. 6, 916–919.

Billings, L.M., Oddo, S., Green, K.N., McGaugh, J.L., LaFerla, F.M., 2005. IntraneuronalAbeta causes the onset of early Alzheimer's disease-related cognitive deficits intransgenic mice. Neuron 45, 675–688.

Brody, D.L., Holtzman, D.M., 2008. Active and passive immunotherapy for neurodegenerativedisorders. Annu. Rev. Neurosci. 31, 175–193.

Brussow, H., 2005. Phage therapy: the Escherichia coli experience. Microbiology 151,2133–2140.

Chyung, J.H., Selkoe, D.J., 2003. Inhibition of receptor-mediated endocytosis demonstratesgeneration of amyloid beta-protein at the cell surface. J. Biol. Chem. 278, 51035–51043.

DeMattos, R.B., Bales, K.R., Cummins, D.J., Dodart, J.C., Paul, S.M., Holtzman, D.M., 2001.Peripheral anti-A beta antibody alters CNS and plasma A beta clearance and

decreases brain A beta burden in a mouse model of Alzheimer's disease. Proc NatlAcad Sci USA 98, 8850–8855.

Dodart, J.C., Bales, K.R., Gannon, K.S., Greene, S.J., DeMattos, R.B., Mathis, C., DeLong, C.A.,Wu, S., Wu, X., Holtzman, D.M., Paul, S.M., 2002. Immunization reverses memorydeficits without reducing brain Abeta burden in Alzheimer's disease model. Nat.Neurosci. 5, 452–457.

Frenkel, D., Solomon, B., 2002. Filamentous phage as vector-mediated antibody deliveryto the brain. Proc Natl Acad Sci USA 99, 5675–5679.

Fukumoto, T., Torigoe, N., Kawabata, S., Murakami, M., Uede, T., Nishi, T., Ito, Y., Sugimura, K.,1998. Peptide mimics of the CTLA4-binding domain stimulate T-cell proliferation. Nat.Biotechnol. 16, 267–270.

Gejima, R., Tanaka, K., Nakashima, T., Hashiguchi, S., Ito, Y., Yoshizaki, K., Sugimura, K.,2002. Human single-chain Fv (scFv) antibody specific to human IL-6 with theinhibitory activity on IL-6-signaling. Hum. Antibodies 11, 121–129.

Harmeier, A., Wozny, C., Rost, B.R., Munter, L.M., Hua, H., Georgiev, O., Beyermann, M.,Hildebrand, P.W., Weise, C., Schaffner, W., Schmitz, D., Multhaup, G., 2009. Role ofamyloid-beta glycine 33 in oligomerization, toxicity, and neuronal plasticity. J. Neurosci.29, 7582–7590.

Hartman, R.E., Izumi, Y., Bales, K.R., Paul, S.M., Wozniak, D.F., Holtzman, D.M., 2005.Treatment with an amyloid-beta antibody ameliorates plaque load, learningdeficits, and hippocampal long-term potentiation in a mouse model of Alzheimer'sdisease. J. Neurosci. 25, 6213–6220.

Hashiguchi, S., Nakashima, T., Nitani, A., Yoshihara, T., Yoshinaga, K., Ito, Y., Maeda, Y.,Sugimura, K., 2003. Human Fc epsilon RIalpha-specific human single-chain Fv (scFv)antibody with antagonistic activity toward IgE/Fc epsilon RIalpha-binding. J. Neurosci.133, 43–49.

Hashiguchi, S., Yamaguchi, Y., Takeuchi, O., Akira, S., Sugimura, K., 2010. Immunologicalbasis of M13 phage vaccine: regulation under MyD88 and TLR9 signaling. BiochemBiophys Res Commun.

Holtzman, D.M., 2008. Alzheimer's disease: Moving towards a vaccine. Nature 454,418–420.

Jan, A., Gokce, O., Luthi-Carter, R., Lashuel, H.A., 2008. The ratio of monomeric to aggregatedformsof Abeta40 andAbeta42 is an important determinant of amyloid-beta aggregation,fibrillogenesis, and toxicity. J. Biol. Chem. 283, 28176–28189.

Janus, C., Pearson, J., McLaurin, J., Mathews, P.M., Jiang, Y., Schmidt, S.D., Chishti, M.A.,Horne, P., Heslin, D., French, J., Mount, H.T., Nixon, R.A., Mercken, M., Bergeron, C.,Fraser, P.E., St George-Hyslop, P., Westaway, D., 2000. A beta peptide immunizationreduces behavioural impairment and plaques in a model of Alzheimer's disease.Nature 408, 979–982.

Kayed, R., Head, E., Thompson, J.L., McIntire, T.M., Milton, S.C., Cotman, C.W., Glabe, C.G.,2003. Common structure of soluble amyloid oligomers implies common mecha-nism of pathogenesis. Science 300, 486–489.

Klyubin, I., Betts, V.,Welzel, A.T., Blennow, K., Zetterberg, H.,Wallin, A., Lemere, C.A., Cullen,W.K., Peng, Y., Wisniewski, T., Selkoe, D.J., Anwyl, R., Walsh, D.M., Rowan, M.J., 2008.Amyloid beta protein dimer-containing human CSF disrupts synaptic plasticity:prevention by systemic passive immunization. J. Neurosci. 28, 4231–4237.

Kotilinek, L.A., Bacskai, B., Westerman, M., Kawarabayashi, T., Younkin, L., Hyman, B.T.,Younkin, S., Ashe, K.H., 2002. Reversible memory loss in a mouse transgenic modelof Alzheimer's disease. J. Neurosci. 22, 6331–6335.

Kurzepa, A., Dabrowska, K., Skaradzinski, G., Gorski, A., 2009. Bacteriophage in-teractions with phagocytes and their potential significance in experimentaltherapy. Clin. Exp. Med. 9, 93–100.

Ladiwala, A.R., Lin, J.C., Bale, S.S., Marcelino-Cruz, A.M., Bhattacharya, M., Dordick, J.S.,Tessier, P.M., 2010. Resveratrol selectively remodels soluble oligomers and fibrils ofamyloid Abeta into off-pathway conformers. J. Biol. Chem. 285, 24228–24237.

Lee, E.B., Leng, L.Z., Zhang, B., Kwong, L., Trojanowski, J.Q., Abel, T., Lee, V.M., 2006.Targeting amyloid-beta peptide (Abeta) oligomers by passive immunization with aconformation-selective monoclonal antibody improves learning and memory inAbeta precursor protein (APP) transgenic mice. J. Biol. Chem. 281, 4292–4299.

LeVine III, H., 1993. Thioflavine T interactionwith synthetic Alzheimer's disease beta-amyloidpeptides: detection of amyloid aggregation in solution. Protein Sci. 2, 404–410.

Li, K., Chen, Y., Li, S., Nguyen, H.G., Niu, Z., You, S., Mello, C.M., Lu, X., Wang, Q., 2010.Chemical modification of M13 bacteriophage and its application in cancer cellimaging. Bioconjug. Chem. 21, 1369–1377.

Luhrs, T., Ritter, C., Adrian, M., Riek-Loher, D., Bohrmann, B., Dobeli, H., Schubert, D.,Riek, R., 2005. 3D structure of Alzheimer's amyloid-beta(1–42) fibrils. Proc NatlAcad Sci USA 102, 17342–17347.

Luhrs, T., Zahn, R., Wuthrich, K., 2006. Amyloid formation by recombinant full-lengthprion proteins in phospholipid bicelle solutions. J. Mol. Biol. 357, 833–841.

Morgan, D., Diamond, D.M., Gottschall, P.E., Ugen, K.E., Dickey, C., Hardy, J., Duff, K.,Jantzen, P., DiCarlo, G., Wilcock, D., Connor, K., Hatcher, J., Hope, C., Gordon, M.,Arendash, G.W., 2000. A beta peptide vaccination prevents memory loss in ananimal model of Alzheimer's disease. Nature 408, 982–985.

Munter, L.M., Voigt, P., Harmeier, A., Kaden, D., Gottschalk, K.E., Weise, C., Pipkorn, R.,Schaefer, M., Langosch, D., Multhaup, G., 2007. GxxxG motifs within the amyloidprecursor protein transmembrane sequence are critical for the etiology of Abeta42.Embo J. 26, 1702–1712.

Oddo, S., Vasilevko, V., Caccamo, A., Kitazawa, M., Cribbs, D.H., LaFerla, F.M., 2006.Reduction of soluble Abeta and tau, but not soluble Abeta alone, amelioratescognitive decline in transgenic mice with plaques and tangles. J. Biol. Chem. 281,39413–39423.

Salloway, S., Sperling, R., Gilman, S., Fox, N.C., Blennow, K., Raskind, M., Sabbagh, M.,Honig, L.S., Doody, R., van Dyck, C.H., Mulnard, R., Barakos, J., Gregg, K.M., Liu, E.,Lieberburg, I., Schenk, D., Black, R., Grundman, M., 2009. A phase 2 multipleascending dose trial of bapineuzumab in mild to moderate Alzheimer disease.Neurology 73, 2061–2070.

38 K. Tanaka et al. / Journal of Neuroimmunology 236 (2011) 27–38

Schenk, D., Barbour, R., Dunn, W., Gordon, G., Grajeda, H., Guido, T., Hu, K., Huang, J.,Johnson-Wood, K., Khan, K., Kholodenko, D., Lee, M., Liao, Z., Lieberburg, I., Motter,R., Mutter, L., Soriano, F., Shopp, G., Vasquez, N., Vandevert, C., Walker, S., Wogulis,M., Yednock, T., Games, D., Seubert, P., 1999. Immunization with amyloid-betaattenuates Alzheimer-disease-like pathology in the PDAPP mouse. Nature 400,173–177.

Schneeberger, A., Mandler, M., Otawa, O., Zauner, W., Mattner, F., Schmidt, W., 2009.Development of AFFITOPE vaccines for Alzheimer's disease (AD)–from concept toclinical testing. J. Nutr. Health Aging 13, 264–267.

Shankar, G.M., Li, S., Mehta, T.H., Garcia-Munoz, A., Shepardson, N.E., Smith, I., Brett, F.M.,Farrell, M.A., Rowan, M.J., Lemere, C.A., Regan, C.M., Walsh, D.M., Sabatini, B.L.,Selkoe, D.J., 2008. Amyloid-beta protein dimers isolated directly from Alzheimer'sbrains impair synaptic plasticity and memory. Nat. Med. 14, 837–842.

Smith, G.P., Scott, J.K., 1993. Libraries of peptides and proteins displayed on filamentousphage. Methods Enzymol. 217, 228–257.

Solomon, B., 2007. Active immunization against Alzheimer's beta-amyloid peptideusing phage display technology. Vaccine 25, 3053–3056.

Stine Jr., W.B., Dahlgren, K.N., Krafft, G.A., LaDu, M.J., 2003. In vitro characterization ofconditions for amyloid-beta peptide oligomerization and fibrillogenesis. J. Biol.Chem. 278, 11612–11622.

Wadia, J.S., Dowdy, S.F., 2005. Transmembrane delivery of protein and peptide drugs byTAT-mediated transduction in the treatment of cancer. Adv. Drug Deliv. Rev. 57,579–596.

Wilcock, D.M., Rojiani, A., Rosenthal, A., Subbarao, S., Freeman,M.J., Gordon,M.N.,Morgan, D.,2004. Passive immunotherapy against Abeta in aged APP-transgenic mice reversescognitive deficits and depletes parenchymal amyloid deposits in spite of increasedvascular amyloid and microhemorrhage. J. Neuroinflammation 1, 24.

Wilcock, D.M., Alamed, J., Gottschall, P.E., Grimm, J., Rosenthal, A., Pons, J., Ronan, V.,Symmonds, K., Gordon, M.N., Morgan, D., 2006. Deglycosylated anti-amyloid-betaantibodies eliminate cognitive deficits and reduce parenchymal amyloid withminimal vascular consequences in aged amyloid precursor protein transgenic mice.J. Neurosci. 26, 5340–5346.

Yoshihara, T., Takiguchi, S., Kyuno, A., Tanaka, K., Kuba, S., Hashiguchi, S., Ito, Y.,Hashimoto, T., Iwatsubo, T., Tsuyama, S., Nakashima, T., Sugimura, K., 2008.Immunoreactivity of phage library-derived human single-chain antibodies toamyloid beta conformers in vitro. J. Biochem. 143, 475–486.

Yoshinaga, K., Matsumoto, M., Torikai, M., Sugyo, K., Kuroki, S., Nogami, K., Matsumoto,R., Hashiguchi, S., Ito, Y., Nakashima, T., Sugimura, K., 2008. Ig L-chain shuffling foraffinity maturation of phage library-derived human anti-human MCP-1 antibodyblocking its chemotactic activity. J. Biochem. 143, 593–601.