binding of igg to amyloid βa4 peptide via the heavy-chain hinge region with preservation of antigen...

Journal ofNeurotmmunology, 48 (1993) 199-204 199 © 1993 Elsevier Science Publishers B.V. AII rights reserved 0165-5728/93/$06.00

JNI 02468

Binding of IgG to amyloid/3A4 peptide via the heavy-chain hinge region with preservation of antigen binding

David Huang a, Margaret Martin b, Danny Hu c, Allen D. Roses a,b, Dmitry Goldgaber d and Warren J. Strit tmatter *'a'b

a Department of Neurobiology, Duke Unwersity Medical Center, Durham, NC 27710, USA b Department of Medicine (Neurology) Box 2900, Joseph and Kathleen Bryan Neurobiology Buildmg, Duke University Medical Center, Durham,

NC 27710, USA c Duke Untversity, Durham, NC 27710, USA

d Department of Psychiatry and Behavioural Science, State University of New York at Stony Brook, NYl1794-8101, USA

(Received 5 April 1993) (Revision received 30 June 1993)

(Accepted 1 July 1993)

Key words: Alzheimer's Disease; IgG; Senile plaque; /3A4 amyloid peptide

Summary

Amyloid /3A4 peptide is found in the extracellular region of the senile plaque and in the angiopathy of Alzheimer's disease. Several other proteins, including IgG, also reside in these abnormal structures. In an attempt to understand how these structures are assembled and to determine how proteins are recruited, interactions of various proteins with synthetic /3A4 peptide have been examined in vitro. Purified IgG binds directly to synthetic /3A4 peptide with high avidity. The domain between amino acids 12-28 of/3A4 binds IgG. /3A4 peptide binds the hinge region of the immunoglobulin heavy chain, and preserves the ability of the immunoglobulin to bind antigen. A protein which does not bind directly to/3A4 peptide can be targetted to the senile plaque and angiopathy by binding to IgG, which avidly binds/3A4 peptide.

Introduction

The senile plaque and congophilic angiopathy are abundant lesions in the brains of patients with Alzheimer's disease (Selkoe, 1991). They both contain complex extracellular structures whose biochemical composition and mechanism of assembly have been characterized only partially. Their role in the pathogenesis of the disease is not known. A prominent constituent of the senile plaque and angiopathy is amyloid /3A4 peptide. /3A4 peptide is produced by proteolytic processing of the amyloid precursor protein (APP) (Younkin, 1991), and contains 39-43 amino acids. Several additional proteins have been localized to the plaque and angiopathy, including apolipoprotein E (Namba et al., 1991; Wisniewski et al., 1992;

* Corresponding author. Phone (919) 684 5963, Fax (919) 684 6514.

Strittmatter et al., 1993a), a-l-antichymotrypsin (Abraham et al., 1988), complement factors Clq and C3q (Eikelenboom et al., 1982; McGeer, 1989), APP (Beyreuther et al., 1991) and lgG (Ishii et al., 1975, 1976; Ihara et al., 1981). The functional interactions of plaque proteins in either the pathogenesis of the disease or in the response to the disease are not known. Alternatively the presence of IgG in senile plaques could result from extravasation post mortem. To study the binding of proteins to the /3A4 peptide we devel- oped an in vitro assay in which synthetic /3A4 peptide is covalently immobilized to a membrane matrix (Strit- tmatter et al., 1993a). Using this assay we have previously characterized the binding of apolipoprotein E and amyloid precursor protein to synthetic /3A4 peptide (Strittmatter et al., 1993a,b). Here we demonstrate that IgG directly and avidly binds synthetic /3A4 amyloid, resisting dissociation by either sodium dodecyl sulfate or guanidine hydrochloride. Binding of the IgG heavy chain to /3A4 peptide occurs through the IgG

200

hinge region. Such binding of IgG to /3A4 peptide still permits IgG to bind its antigen protein. Therefore proteins which do not directly bind to /3A4 may be targetted to the plaque and angiopathy by interaction with IgG.

Materials and methods

Affinity-purified human IgG and IgG fragments (Fab, Fc, F(ab ' ) 2) were purchased from Cappel- Organon Teknika, Durham, NC. Peroxidase-conjugated goat antibodies to human IgG (heavy and light chains) were purchased from Pierce, Rockford, IL. Sheep anti-human serum albumin (HSA) IgG, sheep anti-ferritin IgG, and peroxidase-conjugated sheep IgG against human serum albumin were purchased from The Binding Site, Ltd, San Diego, CA. /3A4 peptides were from Bachem, Torrance, CA. The synthesis of peptides E.H. and H.M. was described previously (Stri t tmatter et al., 1993a). Human cerebrospinal fluid (CSF), from diagnostic lumbar punctures after in- formed consent, was obtained from the Kathleen Bryan Brain Bank at Duke University Medical Center.

All peptides were covalently bound to 13-mm Immo- bilon AV membrane discs as described previously, with 100/zg peptide (Stri t tmatter et al., 1993a). To characterize the binding of immunoglobulins to the immobilized peptides, affinity-purified human IgG (0.3 ng), Fab (0.45 ng), Fc (0.15 ng) or F(ab ' ) 2 (0.45 ng) in 150 /~I PBS (pH 7.4) were incubated with immobil ized/3A4 peptide, other peptides, or with ethanolamine at room tempera ture for 30 min. The membranes were washed in a Millex filter holder (Millipore) with 3.0 ml phos- phate-buffered saline (PBS), followed by 700 /zl 10% sodium dodecyl sulfate (SDS). Retained proteins were

then eluted from the membranes by boiling 5 min in 150 p.1 Laemmli with/3-mercaptoethanol . 45/xl of each sample was loaded on a 12% polyacrylamide gel. Elec- trophoresis and Western transfer were performed as described (Strit tmatter et al., 1993a). The Western transfer membrane was blocked with 40 ml Blotto (5% dried milk in Tris-buffered saline, pH 7.6) with 0.1% Tween at room temperature for 1 h. The membrane was incubated with peroxidase-conjugated goat anti- human IgG antibody which recognizes both heavy and light chains (diluted 1 : 4000 in Blotto) at 4°C overnight. The membrane was rinsed four times with 20 ml Blotto and was then washed three times with 40 ml Blotto for 5 min. Antibody was visualized using an enhanced chemoluminescence detection kit (Amersham) and ex- posure of the membrane to Hyperfi lm (Amersham), as described previously (Strit tmatter et al., 1993a).

Studies examining the binding of IgG to both antigen and /3A4 were conducted by incubating 1.0 /xl cerebrospinal fluid (which contains abumin) with either 0.5 mg sheep anti-human serum albumin antibody or sheep anti-ferritin antibody (control) in 150/xl PBS for 30 min at room temperature. The mixtures were then incubated 30 min with either immobilized /3A4 peptide or immobilized ethanolamine (control) at room temperature. The membranes were first washed with 3.0 ml PBS, followed by 1.0 ml 6 M guanidine hydrochloride, and 2.0 ml PBS. The retained proteins were then eluted by boiling 5 min in 150 /xl Laemmli buffer. Proteins were then electrophoresed and transferred to Immobilon P membrane as described above. The Western transfer membrane was incubated with peroxidase-conjugated sheep ant i-HSA antibody (1:10000 dilution in Blotto) overnight at 4°C. After washing the membranes as described above, immunolabeled albumin was visualized by chemoluminescence. All of the

1 2 3 4 I Load I ~A I Eth II LOadl pA I Eth ]1L0adl ~A I Eth I ,oad' I Eth

5 Load I ~A I Eth

Fig. 1. Binding of IgG and IgG domains to immobilized/~A40_2s ) and to immobihzed ethanolamine. Affinity-purified IgG (panel 1); Fab (panels 2 and 4); F c (panel 3); or F(ab') 2 (panel 5) were applied directly to a polyacrylamide gel (Load), or were incubated with /3A4o_as ) or with ethanolamine (Eth) previously immobilized to Immobilon AV membrane discs. The membranes were then washed with PBS and guanldine hydrochloride, as described in Materials and methods. The membranes were then boiled m Laemmli buffer, and the eluted proteins

electrophoresed, transferred to Immobilon P, and visuahzed as described.

experiments shown in the figures have been replicated at least once.

Results

Purified IgG binds directly to /3A4 peptide, and binding requires the hinge region of the heavy chain. Figure 1, Panel 1, illustrates that purified IgG binds to immobilized /3A4(a_28 ) peptide, and does not bind to the ethanolamine-blocked control membrane. Various domains of the immunoglobulin molecule were examined for binding to immobilized 13A4 peptide. Fab (Panels 2 and 4) and F c (Panel 3) did not bind to /3A4o_z8 ) peptide. F(ab') 2, containing the hinge region of the heavy chain, binds to immobilized /3A4 peptide (Panel 5). To further characterize the avidity of /3A4 peptide binding, F(ab') 2 was incubated with immobilized/3A4(~_28 ) or with immobilized ethanolamine (Fig. 2A), washed with PBS (Lane 1), and then followed by 10% SDS (Lane 2), 4 M urea (Lane 3), or 6 M guanidine hydrochloride (Lane 4). The F(ab')2 that bound to/3A4 peptide was not eluted by urea, and was only partially eluted by SDS or guanidine hydrochlo-

a, 1 2 3 4

I~ AI E t h [ ~ ' ~ " ~ ~ 1 - - ' ~ ] - ' ~ I ~

bm 1 2 3 4 A' 'hll t"ll

Fig. 2. Binding of affinity-purified F(ab') 2 (a), or CSF IgG (b) to /3A4 peptide following various washes. Affinity-purified F(ab') 2 (a), or cerebrospinal fluid, which contains IgG (b), were incubated with either /3A4(l_28 ) or ethanolamine previously immobilized to Immo- bilon AV membrane discs. The discs were then washed with 3.0 ml PBS (panel 1), followed by 700 /zl 5% SDS (panel 2), 700 /xl 4 M urea (panel 3), or 6 M guanidine hydrochloride (panel 4). Proteins were then eluted by boiling in Laemmli buffer, electrophoresed, and

immunoglobulin visualized as described.

201

Fig. 3. Binding of purified IgG to various flA4 peptides, and to other peptides. Affinity-purified IgG was incubated with /3A40_zs), /3A4(12_28), /3A40_4o), H.M. peptide ('hydropathic mimic' peptide), E.H. peptide ('even-hydropathic' peptide) or ethanolamine (Eth), previously immobilized to Immobilon AV membrane discs. The membranes were then washed with PBS and 6 M guanidine hydrochloride. Retained proteins were then eluted by boiling in Laemmli

buffer, electrophoresed, and immunoglobulin visualized.

ride. Similar results were obtained with IgG in CSF, shown in Fig. 2B. Incubation of cerebrospinal fluid, which contains IgG, with immobilized /3A4(l_z8 ) or with immobilized ethanolamine, and washed under the conditions shown in Fig. 2A, reveals that CSF IgG avidly binds/3A4. In a previous study, Pardridge et al. (1987) examined human cerebrospinal fluid for protein immunoreactive with an antibody against synthetic ~A4(t_28), and demonstrated that CSF IgG was identi- fied by this antibody, suggesting either cross-reactivity, or the presence of/3A4 peptide on IgG.

Both IgG and F(ab') e bind with high avidity to full length/3A4 peptide (/~A(l_40)). As illustrated in Fig. 3, affinity-purified IgG bound t o /~A4(1_40), flA40_2s), and 13A4o2_28 ). Thus the binding requires amino acids 12-28 of/3A4. Peptide H.M. ('hydro-mimic' peptide) is a 17-amino acid peptide with a hydropathy profile similar to /3A4o2_28), but with different amino acids, while peptide E.H. ('even-hydro' peptide) is a 17-amino acid peptide containing the same amino acids as ~Ao2_28 ) but with a scrambled sequence (Strittmatter et al., 1993a). IgG bound to peptide H.M., but only bound minimally to peptide E.H. (Fig. 3). Similar results were obtained with purified F(ab') 2 (not shown). The lack of binding of IgG and F(ab') 2 to the scrambled amino acid sequence (peptide E.H.) suggests a certain degree of specificity of interaction. Since IgG and F(ab') 2 bound to peptide H.M., which has the same hydropathic profile as /3A(lz_z8), binding may require specific steric, hydrophobic, or charge interactions.

With/3A4 bound to the IgG hinge domain, IgG can in principle interact with other antigens. This presents the possibility that IgG can target an antigen to the senile plaque by first binding its specific antigen and then binding to/3A4 peptide. To test this we examined the ability of albumin to bind to /3A4 peptide directly and to bind to/3A4 peptide indirectly through anti-albumin lgG. As shown in Fig. 4, albumin in cerebrospinal fluid was weakly bound to both immobilized /3A4 peptide or to ethanolamine-bound control membranes. Cerebrospinal fluid was then first incubated

202

1 2 3 4 5 ] ' Ethl[ 13A ' Eth[ I 13A ' Eth l l 13A ' Eth II 13A ' Ethl

m

Fig. 4. Binding of CSF albumin to immobihzed /3A4~i_2s ~ (panel 1), and after pre-incubation with anti-albumin IgG (panel 4). Cerebrospinal fluid was incubated with immobilized /3A4~t_2s ) o r with lmmoblhzed ethanolamlne (Eth) Immobilon AV discs (panel 1), or was first pre-incubated with anti-albumin IgG (panel 4), or with antl-ferritm IgG (panel 5). As controls, anti-albumin IgG alone (panel 2) and anti-ferritin lgG alone (panel 3) (without cerebrospinal fluid), were incubated with these membranes. The membranes were washed in PBS and 6 M guanidine, and retained proteins eluted by boiling in Laemmli buffer After electrophoresls and Western transfer, albumin was detected with the anti-albumin

antibody.

with either an anti-albumin IgG or with an anti-ferritin IgG (as a control) and was then incubated with immobilized /3A4 peptide. Prior incubation with anti-albumin IgG markedly increased the amount of albumin bound to the immobilized /3A4 peptide (Fig. 4, lane 4) in contrast to pre-incubation with the anti-ferritin IgG control (lane 5). Neither of the antibodies were immunoreactive for albumin (Fig. 4, lanes 2 and 3) and both antibodies bound equally well to immobilized /3A4 peptide (not shown). Thus albumin, which binds weakly to /3A4 peptide in this assay can be targetted to /3A4 peptide (and therefore to the senile plaque and angiopathy) by specific IgG.

lgG already bound to /3A4 peptide can still recog- nize and bind antigen. Both anti-albumin IgG and anti-ferritin IgG were first bound to immobilized /3A4

F-U-]

Fig. 5. Binding of CSF albumin to /3A4(1_28 ) (lanes 1 and 4), and to anti-albumin IgG (lanes 2 and 5) or anti-ferritin IgG (lanes 3 and 6), previously bound to immobilized /3A4. Anti-albumin IgG or anti-ferritin IgG were incubated with /3A4 peptide immobilized to Immo- bilon AV discs, and washed with PBS. The discs were then incubated with cerebrospinal fired, and washed with PBS (lanes 1, 2, 3) or with PBS and 6 M guanidlne hydrochloride (lanes 4, 5, 6), Proteins retained by each membrane were then eluted by boiling in Laemmli buffer, and albumen was detected with an anti-albumin antibody

after electrophoresis and Western transfer.

peptide, and were then incubated with cerebrospinal fluid. The anti-albumin IgG continued to bind albumin (Fig. 5, lane 2), while the anti-ferritin antibody did not (lane 3). The binding of albumin to /3A4 peptide via specific IgG was maintained even after wash with 6 M guanidine hydrochloride (lane 5). Data in Figs. 4 and 5 suggest that IgG is capable of high avidity binding to both antigen and to /3A4 peptide, and that the tempo- ral sequence of interaction does not appear important.

Discussion

Since the senile plaque and congophilic angiopathy are complex structures containing many proteins, the recruitment of these proteins may be important in the pathogenesis of the disease. The ability of IgG to bind both its specific antigen and /3A4 peptide appears to be one mechanism by which proteins which do not directly bind ~A4 can be targetted to these structures. Understanding the detailed molecular mechanism by which the hinge region of IgG binds /3A4 peptide may present the opportunity to specifically, and selectively, block this interaction.

We have previously demonstrated that the amount of /3A4 amyloid in the senile plaque and angiopathy in AD brains is directly dependent on the A P O E genotype (Schmechel et al., 1993) and that apoE4 binds /3A4 with faster kinetics, and more avidly, than apoE3 (Roses et al., 1993; Strittmatter et al., 1993a,c). Thus the localization of /3A4 in AD plaques may be through decoration of apoE. The binding of IgG to /3A4 may provide a sandwich mechanism for the deposition of other plaque constituents in AD. APOE4 is associated with late-onset AD (Saunders et al., 1993) and is not found in autosomal dominant early-onset AD (chro- mosome 21 or 14 type), hereditary Creutzfeld-Jakob

diseases, Down's syndrome, and familial amyloid (FAP) (Saunders et al., in preparation), yet apoE is deposited in amyloid plaques in these diseases. Alternate mechanisms of plaque formation with the involvement of some of the same constituents may involve primary causation by several different triggers. The ability of IgG to sandwich diverse constituents to plaques suggests immunological involvement in some plaque-forming diseases.

Fibril-forming proteins have often been studied in vitro using models of self-association. In most diseases the plaques contain constituents whose role is usually undefined. The interaction of IgG as a potential sandwich in plaque formation in vivo, perhaps targetting pathogenic or plaque-enhancing factors, may be critical in several diseases. Understanding the detailed molecular mechanism by which the hinge region binds peptides may present potential therapeutic targets.

Acknowledgements

This work was conducted with the support of Na- tional Institutes of Health L.E.A.D. Award 5R35 AG- 07922 to A.D.R. and National Institutes of Health Alzheimer's Disease Research Center 5P50 AG-05128, and numerous contributions, each less than $100, to the Joseph and Kathleen Bryan Alzheimer's Disease Research Center. We would like to thank Wendell F. Rosse, M.D. for helpful discussions.

References

Abraham, C.R., Selkoe, D.J. and Potter, H. (1988) Immunochemical identification of the serine protease inhibitor al-antichymotry p- sin in the brain amyloid deposits of Alzheimer's disease. Cell 52, 487-501.

Beyreuther, K. and Masters, C.L.(1991) Amyloid precursor protein (APP) and /3A4 amyloid in the etiology of Alzheimer's disease: Precursor-product relationships in the derangement of neuronal function. Brain Pathol. 1, 241-251.

Eikelenboom, P. and Stam, F.C. (1982) Immunoglobulins and complement factors in senile plaques. Acta Neuropathol. 57, 239-242.

Ihara, Y., Kurizaki, H., Nukina, N. et al. (1981) Presence of immunoglobulin light chain in the cores of senile plaques - an unlabelled antibody peroxidase-antiperoxidase (PA) study. Neu- rol. Med. 15, 292-295.

203

Ishii, T. and Haga, S. (1976) Immuno-electron microscopic localization of immunoglobulins in amyloid fibrils of senile plaques. Acta Neuropathol. 36, 243-249.

Ishii, T., Haga, S. and Shimizu, F. (1975) Identification of compo- nents of immunoglobulins in senile plaques by means of fluores- cent antibody technique. Acta Neuropathol. 32, 157-162.

McGeer, P.L., Akiyama, H., ltagaki, S. and McGeer, E.G. (1989) Immune system response in Alzheimer's disease. Can. J. Neurol. Sci. 16, 516-527.

Namba, Y., Tomonaga, M., Kawasaki, H., Otomo, E. and Ikeda, K. (1991) Apolipoprotein E immunoreactivity in cerebral amyloid deposits and neurofibrillary tangles in Alzheimer's disease and kuru plaque amyloid in Creutzfeldt-Jakob disease. Brain Res. 541, 163-166.

Pardridge, W.M., Vinters, H.V., Miller, B.L., Tourtellotte, W.W., Eisenberg, J.B. and Yang, J. (1987) High molecular weight Alzheimer's disease amyloid peptide immunoreactivity in human serum and CSF is an immunoglobulin G. Biochem. Biophys. Res. Commun. 145, 241-248.

Roses, A.D., Pericak-Vance, M.A., Saunders, A.M., Schmechel, D., Goldgaber, D. and Strittmatter, W.J. (1993) Complex genetic disease: can genetic strategies in Alzheimer's disease and new genetic mechanisms be applied to epilepsy? Epilepsia, in press.

Saunders, A.M., Strittmatter, W.J., Schmechel, D., St. George-Hy- slop, P.H., Pericak-Vance, M., Joo, S.H., Rosi, B.L., Gusella, J.F., Crapper-MacLachlan, D.R., Growden, J., Alberts, M., Hulette, C., Crain, B., Goldgaber, D. and Roses, A.D. (1993) Association of apolipoprotein E allele e4 with late-onset familial and sporadic Alzheimer's disease. Neurology 43, 1467-1472.

Schmechel, D.E., Saunders, A.M., Strittmatter, W.J., Crain, B., Hulette, C., Joo, S.H., Pericak-Vance, M.A., Goldgaber, D. and Roses, A.D. (1993) Increased amyloid /3-peptide deposition as a consequence of apolipoprotein E genotype in late-onset Alzheimer disease. Proc. Natl. Acad. Sci. USA 90, 9649-9653.

Selkoe, D.J. (1991) The molecular pathology of Alzheimer's disease. Neuron 6, 487-498.

Strittmatter, W.J., Saunders, A.M., Schmechel, D., Pericak-Vance, M., Enghild, J., Salveson, G.S. and Roses, A.D. (1993a) Apolipoprotein E: High-avidity binding to /3-amyloid and increased frequency of type 4 allele in late-onset familial Alzheimer disease. Proc. Natl. Acad. Sei. USA 90, 1977-1981.

Strittmatter, W.J., Huang, D.Y., Bhasin, R., Roses, A.D. and Goldgaber, D. (1993b) Avid binding of/3A amyloid peptide to its own precursor. Exp. Neurol. 122, 327-335.

Strittmatter, W.J., Weisgraber, K.H., Huang, D., Dong, L.-M., Salvesen, G.S., Pericak-Vance, M., Schmechel, D., Saunders, A., Goldgaber, D. and Roses, A.D. (1993c) Binding of human apolipoprotein E to flA4 peptide: Isoform-specific effects and implications for late-onset Alzheimer disease. Proc. Natl. Acad. Sci. USA 90, 8098-8102.

Wisniewski, T. and Frangione, B. (1992) Apolipoprotein E: a patho- logical chaperone protein in patients with cerebral and systemic amyloid. Neurosci. Lett. 135, 235-238.

Younkin, S.G. (1991) Processing of the Alzheimer's disease /3A4 amyloid protein precursor. Brain Pathol. 1,253-262.

binding of igg to amyloid βa4 peptide via the heavy-chain hinge region with preservation of antigen...

Documents