Anti-amyloid immunotherapy remains one of the front-line strategies for the development of Alzheimer therapeutics. Both passive and active immunization are currently under active development for human clinical application. Antibodies that target the amino terminus of Aβ seem particularly interesting. Not only does this region appear to be an immuno-dominant site, but antibodies that recognize epitopes in this region also seem particularly effective in reversing AD pathogenesis in transgenic animals and in depolymerizing amyloid fibrils in vitro. In this article, Chris Dealwis and colleagues report the crystal structures of two monoclonal antibodies that target the amino terminus of Aβ.

These antibodies, PFA1 and PFA2, are remarkably specific for the EFRHD sequence at residues 3-7 of the Aβ peptide, as substitution of an alanine residue at any position nearly eliminates antibody binding. The crystal structures of the Fab complex with the peptide DAEFRHDS reveals that a WWDDD motif in the heavy chain complementarity determining region (CDR) of the antibodies forms salt...  Read more

Anti-amyloid immunotherapy remains one of the front-line strategies for the development of Alzheimer therapeutics. Both passive and active immunization are currently under active development for human clinical application. Antibodies that target the amino terminus of Aβ seem particularly interesting. Not only does this region appear to be an immuno-dominant site, but antibodies that recognize epitopes in this region also seem particularly effective in reversing AD pathogenesis in transgenic animals and in depolymerizing amyloid fibrils in vitro. In this article, Chris Dealwis and colleagues report the crystal structures of two monoclonal antibodies that target the amino terminus of Aβ.

These antibodies, PFA1 and PFA2, are remarkably specific for the EFRHD sequence at residues 3-7 of the Aβ peptide, as substitution of an alanine residue at any position nearly eliminates antibody binding. The crystal structures of the Fab complex with the peptide DAEFRHDS reveals that a WWDDD motif in the heavy chain complementarity determining region (CDR) of the antibodies forms salt bridges, hydrogen bonds, and hydrophobic contacts with the EFRHD sequence of Aβ.

Although both PFA1 and PFA2 are remarkably specific for the EFRHD sequence, the authors show that a similar sequence (AKFRHD) derived from the human protein GRIP1 also reacts with the monoclonal antibodies. This raises the possibility of undesirable cross-reactivity with other human proteins; however, the structure of the antigen combining site suggests that one could redesign the CDRs to eliminate undesired cross-reactivity.

The amino terminus of Aβ is also interesting because it seems to contain a conformational switch associated with aggregation and is the site that some conformation-dependent antibodies recognize. The fact that antibodies directed against this region depolymerize amyloid fibrils suggests that antibody binding induces a structure that is incompatible with the amyloid fibril lattice (1). The amino terminus is also the site of a conformation-dependent epitope recognized by the M16 polyclonal antisera that is specific for Aβ aggregates and fibrils, but does not recognize Aβ monomer or APP (2).

References:
1. Frenkel D, Balass M, Katchalski-Katzir E, Solomon B. High affinity binding of monoclonal antibodies to the sequential epitope EFRH of beta-amyloid peptide is essential for modulation of fibrillar aggregation. J Neuroimmunol. 1999 Mar 1;95(1-2):136-42. Abstract

2. Soreghan B, Pike C, Kayed R, Tian W, Milton S, Cotman C, Glabe CG. The influence of the carboxyl terminus of the Alzheimer Abeta peptide on its conformation, aggregation, and neurotoxic properties. Neuromolecular Med. 2002;1(1):81-94. Abstract

View all comments by Charles Glabe

The paper of Gardberg et al. (1) describes, in an elegant and convincing way, the molecular basis of immunotherapy with Aβ peptide anti-N-terminal antibodies. They report the isolation of two mAbs (PFA1 and PFA2) raised against stabilized protofibrils of Aβ, which recognize Aβ monomers, protofibrils, and fibrils. Importantly, they report the structures of their antigen binding fragments (Fabs) in complex with the Aβ(1-8) peptide DAEFRHDS.

As previously shown, immunization against the EFRH sequence rescues cognitive function in mouse models of Alzheimer disease. The EFRH epitope is available for antibody binding when Aβ peptide is either in solution or is an aggregate, and locking of this epitope by antibodies affects the dynamics of all the molecules, preventing self-aggregation as well as enabling resolubilization of already formed aggregates (2-4). All these prior findings illustrate the importance of understanding the structural basis of antibody recognition of this sequence.

Among the proposed mechanisms of immunotherapy, the catalytic dissolution via antibodies,...  Read more

The paper of Gardberg et al. (1) describes, in an elegant and convincing way, the molecular basis of immunotherapy with Aβ peptide anti-N-terminal antibodies. They report the isolation of two mAbs (PFA1 and PFA2) raised against stabilized protofibrils of Aβ, which recognize Aβ monomers, protofibrils, and fibrils. Importantly, they report the structures of their antigen binding fragments (Fabs) in complex with the Aβ(1-8) peptide DAEFRHDS.

As previously shown, immunization against the EFRH sequence rescues cognitive function in mouse models of Alzheimer disease. The EFRH epitope is available for antibody binding when Aβ peptide is either in solution or is an aggregate, and locking of this epitope by antibodies affects the dynamics of all the molecules, preventing self-aggregation as well as enabling resolubilization of already formed aggregates (2-4). All these prior findings illustrate the importance of understanding the structural basis of antibody recognition of this sequence.

Among the proposed mechanisms of immunotherapy, the catalytic dissolution via antibodies, which act as chaperones catalyzing the structural change of the Aβ peptide from the β-strand to an alternative conformation less prone to aggregation, has an important role. Consistent with this mechanism, the efficacy of a given mAb depends on the Aβ sequence element it binds; thus, the mAb 6C6, which recognizes the Aβ N- terminus, is three times more effective in disaggregating Aβ fibrils than the mAb 1C2 directed to other regions. Antibody binding to Aβ is required in either monomer or aggregated forms. The authors suggest that the most appropriate antibodies are those equally capable of recognizing all assembly forms of Aβ peptides, and this is particularly pointed out in this study. Indeed, antibodies against the EFRHD sequence recognize Aβ in all these conformations. As previously shown (5), only antibodies against this sequence are able to dissolve already formed aggregation.

The high specificity of such antibodies to epitope EFRHD results from studies on crystallization of Fab fragments with Aβ. The Fab fragments exhibit binding to Aβ monomers in the 20–40 nM range, and this binding is significantly impaired or eliminated in Aβ(1–40) mutants where a single residue in the 3–7 segment is replaced with alanine.

The accumulated experience of many efforts to obtain antibodies to Aβ suggests that the N-terminus is the immuno-dominant epitope of this peptide. Furthermore, if aggregated forms of Aβ are to be targeted in therapy, antibodies to the N-terminus will probably be required, given the poor accessibility of other portions of the sequence in aggregates.

As cross-reactivity occurred between these antibodies and other unrelated proteins, a smaller amount of high-affinity antibodies to the N-terminal epitope of Aβ peptide is required for a successful immunotherapy in Alzheimer disease.

References:
1. Gardberg AS, Dice LT, Ou S, Rich RL, Helmbrecht E, Ko J, Wetzel R, Myszka DG, Patterson PH, Dealwis C. Molecular basis for passive immunotherapy of Alzheimer's disease. Proc Natl Acad Sci U S A. 2007 Sep 25; [Epub ahead of print] Abstract

2. Frenkel D, Balass M, Solomon B. N-terminal EFRH sequence of Alzheimer's beta-amyloid peptide represents the epitope of its anti-aggregating antibodies. J Neuroimmunol. 1998 Aug 1;88(1-2):85-90. Abstract

3. Frenkel D, Balass M, Katchalski-Katzir E, Solomon B. High affinity binding of monoclonal antibodies to the sequential epitope EFRH of beta-amyloid peptide is essential for modulation of fibrillar aggregation. J Neuroimmunol. 1999 Mar 1;95(1-2):136-42. Abstract

4. Solomon B. Clinical immunologic approaches for the treatment of Alzheimer's disease Expert Opinion on Investigational Drugs. 2007;16(6):819-828. Review. Abstract

5. Solomon B, Koppel R, Frankel D, Hanan-Aharon E. Disaggregation of Alzheimer -amyloid by site-directed mAb. Proc. Natl. Acad. Sci. USA. 1997;94: 4109-4112. Abstract

View all comments by Beka Solomon

I agree with Charles and Beka that this is excellent work. It's actually long overdue in the AD field, which is at the same time crowded and lacking some essential experts.

Nevertheless, I am more critical than my two learned friends and colleagues about the real meaning of this study for immunotherapy in AD. After careful reading—and discussion with an expert or two—we came to the conclusion that this paper sails under the wrong flag. A more apt title might have read: "Molecular Basis for Recognition of Epitope EFRHD on the Amyloid-β Peptide by a Monoclonal Antibody."

The data highlight in exquisite detail the structure of the peptide-antibody immune complex. But they do not address—and therefore do not answer—the primary question in AD immunotherapy: why do N-terminal-specific antibodies dissociate amyloid peptide aggregates, and thereby improve the cognitive functions of AD mice (and hopefully patients as well)?

I am convinced that this excellent paper will help considerably in paving the way to answer the first part of that question. At the same time, I cannot...  Read more

I agree with Charles and Beka that this is excellent work. It's actually long overdue in the AD field, which is at the same time crowded and lacking some essential experts.

Nevertheless, I am more critical than my two learned friends and colleagues about the real meaning of this study for immunotherapy in AD. After careful reading—and discussion with an expert or two—we came to the conclusion that this paper sails under the wrong flag. A more apt title might have read: "Molecular Basis for Recognition of Epitope EFRHD on the Amyloid-β Peptide by a Monoclonal Antibody."

The data highlight in exquisite detail the structure of the peptide-antibody immune complex. But they do not address—and therefore do not answer—the primary question in AD immunotherapy: why do N-terminal-specific antibodies dissociate amyloid peptide aggregates, and thereby improve the cognitive functions of AD mice (and hopefully patients as well)?

I am convinced that this excellent paper will help considerably in paving the way to answer the first part of that question. At the same time, I cannot resist adding this extra level of complexity of why Mabs against "conformational" epitopes are most effective in doing what they do (Muhs et al., 2007)?

References:
Muhs A, Hickman DT, Pihlgren M, Chuard N, Giriens V, Meerschman C, Van Der Auwera I, Van Leuven F, Sugawara M, Weingertner MC, Bechinger B, Greferath R, Kolonko N, Nagel-Steger L, Riesner D, Brady RO, Pfeifer A, Nicolau C. Liposomal vaccines with conformation-specific amyloid peptide antigens define immune response and efficacy in APP transgenic mice. Proc Natl Acad Sci U S A. 2007 Jun 5;104(23):9810-5. Abstract

View all comments by Fred Van Leuven

This study reports the three-dimensional structure, to near atomic resolution, of both an Abeta antibody and the complex with its antigen.

View all comments by Paul Coleman

Hoyer and coworkers have solved a structure of the amyloid-β peptide in complex with a phage-display selected affibody using NMR spectroscopy. The affibody is responsible for stabilizing the Aβ monomer by inhibiting fibril formation. The Aβ adopts a parallel β-hairpin structure, where the two β-strands consisting of residues 15-22 (strand A) and 30-36 (strand B) form intramolecular hydrogen bonds between each other. Strand A is stabilized by a short strand from the affibody which runs anti-parallel, while strand B is stabilized by a short strand that is parallel to it.

From a large body of data, we know that fibrils exhibit a “cross-β” pattern in x-ray fiber diffraction (1). This is associated with a fundamental structure consisting of extended β-sheet networks in which peptide chains are displayed perpendicular to the fibril axis, while the hydrogen bonding direction of the sheet is parallel to the fibril axis (2,3). Hence, the direction of the hydrogen bonds of a conventional β-hairpin as observed in the current study will not fit the bill.

The authors acknowledge...  Read more

Hoyer and coworkers have solved a structure of the amyloid-β peptide in complex with a phage-display selected affibody using NMR spectroscopy. The affibody is responsible for stabilizing the Aβ monomer by inhibiting fibril formation. The Aβ adopts a parallel β-hairpin structure, where the two β-strands consisting of residues 15-22 (strand A) and 30-36 (strand B) form intramolecular hydrogen bonds between each other. Strand A is stabilized by a short strand from the affibody which runs anti-parallel, while strand B is stabilized by a short strand that is parallel to it.

From a large body of data, we know that fibrils exhibit a “cross-β” pattern in x-ray fiber diffraction (1). This is associated with a fundamental structure consisting of extended β-sheet networks in which peptide chains are displayed perpendicular to the fibril axis, while the hydrogen bonding direction of the sheet is parallel to the fibril axis (2,3). Hence, the direction of the hydrogen bonds of a conventional β-hairpin as observed in the current study will not fit the bill.

The authors acknowledge this fact, and propose that what they observe might be an intermediate structure. In fact, the authors suggest that their structure may amount to the toxic soluble oligomers associated with AD. If this is true, the strands A and B will have to rotate 90 degrees about their axis in order to form the in-register parallel β structure that is likely to form the fibrils. While one cannot rule out this possibility, one must be cautious with such interpretations. It is equally possible that the affibody might be responsible for forming a stable, mixed β-sheet structure. In any event, the affibodies can be useful for passive immunotherapy via the peripheral sink theory (4), as they bind Aβ monomers. Hence, this is an interesting study and the main conclusions require further validation.

References:
1. Sunde M, Blake C. The structure of amyloid fibrils by electron microscopy and X-ray diffraction. Adv Protein Chem. 1997;50:123-59. Abstract

2. Petkova AT, Leapman RD, Guo Z, Yau WM, Mattson MP, Tycko R. Self-propagating, molecular-level polymorphism in Alzheimer's beta-amyloid fibrils. Science. 2005 Jan 14;307(5707):262-5. Abstract

3. Guo JT, Wetzel R, Xu Y. Molecular modeling of the core of Abeta amyloid fibrils. Proteins. 2004 Nov 1;57(2):357-64. Abstract

4. Solomon B. Intravenous immunoglobulin and Alzheimer's disease immunotherapy. Curr Opin Mol Ther. 2007 Feb;9(1):79-85. Abstract

View all comments by Chris Dealwis

Capturing Aβ Using Engineered Affinity Proteins
Alzheimer disease (AD) is associated with the amyloid-β protein (Aβ) which assembles into toxic oligomers, protofibrils, and fibrils, and is the major component of amyloid plaques in the AD brain. Substantial evidence implicates the early stages of Aβ assembly in the onset of the disease. Many different strategies that aim at preventing Aβ molecules from formation of toxic assemblies are currently under investigation.

The present study by Hoyer et al. was motivated by novel therapeutic strategies that explore ways to create a peripheral sink mechanism by administering an Aβ binding molecule, a ligand, with the capacity to reduce Aβ in the central nervous system by channeling it into the plasma. As the Aβ1-40 binding molecule, Hoyer et al. proposed to apply an engineered affinity protein (affibody), ZAβ3, based on the Z domain derived from the staphylococcal protein A. In their paper, Hoyer et al. presented 16 different ligands which were previously shown to bind both Aβ1-40 and Aβ1-42 and to form dimers through the...  Read more

Capturing Aβ Using Engineered Affinity Proteins
Alzheimer disease (AD) is associated with the amyloid-β protein (Aβ) which assembles into toxic oligomers, protofibrils, and fibrils, and is the major component of amyloid plaques in the AD brain. Substantial evidence implicates the early stages of Aβ assembly in the onset of the disease. Many different strategies that aim at preventing Aβ molecules from formation of toxic assemblies are currently under investigation.

The present study by Hoyer et al. was motivated by novel therapeutic strategies that explore ways to create a peripheral sink mechanism by administering an Aβ binding molecule, a ligand, with the capacity to reduce Aβ in the central nervous system by channeling it into the plasma. As the Aβ1-40 binding molecule, Hoyer et al. proposed to apply an engineered affinity protein (affibody), ZAβ3, based on the Z domain derived from the staphylococcal protein A. In their paper, Hoyer et al. presented 16 different ligands which were previously shown to bind both Aβ1-40 and Aβ1-42 and to form dimers through the intermolecular disulfide bonding between Cys28. Using a combination of powerful methods—isothermal titration calorimetry, circular dichroism, and heteronuclear single quantum correlation (HSQC) NMR spectroscopy—Hoyer et al. derived detailed structural information on the observed ZAβ3:Aβ1-40 complex.

Dimeric ZAβ3 was shown to bind monomeric Aβ1-40 with 1:1 stoichiometry. Binding of ZAβ3 was coupled to folding of Aβ1-40 and the ligand itself. Aβ1-40 was shown to undergo a conformational transition into a β-hairpin conformation upon binding. ZAβ3 dimer, composed of two β-strands and four α-helices wrapping around the Aβ1-40 monomer, created a large hydrophobic tunnel-like cavity. Aβ1-40 within this cavity adopted a β-hairpin with two strands, 17-23 and 30-36, connected by intramolecular backbone hydrogen bonds. Importantly, the cavity of ZAβ3 dimer was inaccessible to water, and the nonpolar faces of the Aβ1-40 β-hairpin inside the cavity were thus mostly shielded from the solvent. The N-terminal region of Aβ1-40, 1-15, and the terminal regions 1-13 and 57-58 of both ZAβ3 subunits were not well defined by the NMR constraints, implying lack of any ordered structure in these regions.

The significance of the ZAβ3 ligand as an inhibitor of Aβ1-40 assembly is that it binds to the hydrophobic part of Aβ1-40 and thereby disrupts Aβ’s potential for further β-sheet extension. To make that possible, the ZAβ3 ligand was engineered such that the two solvent-exposed short β-strands of the ZAβ3 dimer have little propensity to form either α-helical or β-sheet structure.

Hoyer et al. discussed the β-hairpin monomer conformation of Aβ1-40 in relation to the conformation of individual Aβ1-40 molecules within a fibril as proposed in a fibril model by Petkova et al. (1). While the hook-like structure of the β-hairpin monomer resembles the hook-like structure in the fibril model, the hydrogen bonding in the latter is rotated by a right angle with respect to the former. Such a rotation requires a significant structural change as it involves the breaking of intramolecular hydrogen bonds (to destabilize the β-hairpin monomer) and the formation of new intermolecular hydrogen bonds between neighboring Aβ1-40 molecules.

Hoyer et al. suggest that the β-hairpin monomer conformation may be a key conformation for fibril formation to take place. Here, I will give some insights on the significance of the β-hairpin Aβ1-40 conformation from the point of view of a computational physicist. Even though the β-hairpin conformation of Aβ1-40 monomer may be accessible, it may not be energetically favorable under normal aqueous conditions. On the contrary, in aqueous solutions Aβ1-40 monomer adopts a collapsed-coil conformation with less that 15 percent of total secondary structure (2,3). The interactions driving the folding into a collapsed-coil conformation are primarily due to the effective hydrophobicity as also corroborated by computational studies (4). The fact that Hoyer et al. observed a β-hairpin monomer structure of Aβ1-40 within the ZAβ3:Aβ1-40 complex is very much to do with the particular local environment associated with Aβ1-40 folding. The fact that the observed cavity formed by the ZAβ3 dimer is water inaccessible is critical for understanding Aβ1-40 β-hairpin formation. With no water present, there is no hydrophobic effect, so Aβ1-40 folding cannot be driven by the hydrophobic collapse as in the case of aqueous environment. In the absence of the hydrophobic effect, molecular dynamics simulations indeed show that a β-hairpin and other β-sheet monomer structures with up to four β-strands are formed, leading to formation of planar β-sheet dimers (5). This same study also showed that in water, these planar β-sheet monomer and dimer conformations would be equally favorable to both Aβ1-40 and Aβ1-42—a result that is difficult to reconcile with the fact that Aβ1-42 forms larger assemblies faster than does Aβ1-40. While the computational study of Aβ oligomer formation indicated the presence of a turn/loop centered at G25-S26 in both Aβ1-40 and Aβ1-42, with the local structure that resembles the β-hairpin, the intramolecular hydrogen bonding is at best weak, consistent with the importance of effective hydrophobicity in oligomer formation (4). Finally, the presence of the intramolecular hydrogen bonding in a β-hairpin conformation may actually increase the free energy barrier for transition from a monomeric β-hairpin into a fibril-like structure because breaking the hydrogen bonds in a β-hairpin is energetically unfavorable. Thus, it would be energetically more favorable to form a fibril-like structure from collapsed-coil monomer/oligomer conformations, which already possess a turn structure at G25-S26, held together by hydrophobically driven forces rather than intramolecular hydrogen bonds.

In conclusion, Aβ1-40 monomer forming the β-hairpin structure upon dimeric ZAβ3 binding may create a large free energy barrier preventing Aβ1-40 to form toxic assemblies and thus provide a new and exciting therapeutic strategy.

References:
1. Petkova AT, Ishii Y, Balbach JJ, Antzutkin ON, Leapman RD, Delaglio F, Tycko R. A structural model for Alzheimer's beta -amyloid fibrils based on experimental constraints from solid state NMR. Proc Natl Acad Sci U S A. 2002 Dec 24;99(26):16742-7. Abstract

2. Zhang S, Iwata K, Lachenmann MJ, Peng JW, Li S, Stimson ER, Lu Y, Felix AM, Maggio JE, Lee JP. The Alzheimer's peptide a beta adopts a collapsed coil structure in water. J Struct Biol. 2000 Jun;130(2-3):130-41. Abstract

3. Kirkitadze MD, Condron MM, Teplow DB. Identification and characterization of key kinetic intermediates in amyloid beta-protein fibrillogenesis. J Mol Biol. 2001 Oct 5;312(5):1103-19. Abstract

4. Urbanc B, Cruz L, Yun S, Buldyrev SV, Bitan G, Teplow DB, Stanley HE. In silico study of amyloid beta-protein folding and oligomerization. Proc Natl Acad Sci U S A. 2004 Dec 14;101(50):17345-50. Abstract

5. Urbanc B, Cruz L, Ding F, Sammond D, Khare S, Buldyrev SV, Stanley HE, Dokholyan NV. Molecular dynamics simulation of amyloid beta dimer formation. Biophys J. 2004 Oct;87(4):2310-21. Abstract

View all comments by Brigita Urbanc