Amyloid-β, the peptide that forms the hallmark plaques of Alzheimer disease, can exist in a variety of forms, including monomers, soluble oligomers, and fully fledged fibrils. The oligomers, or protofibrils, are generally thought to be intermediates on the way to true fibrils (but see ARF related news story), and have been shown to be the most toxic to cells (see ARF related news story and ARF news story). But what is the basis for this toxicity? In an accelerated publication in the Journal of Biological Chemistry, Charlie Glabe and colleagues from the University of California, Irvine, show that protofibrils formed from several different amyloidogenic peptides are all toxic in the same way—they permeabilize lipid bilayers, increasing the flux of ions across the membrane.

Glabe and colleagues were drawn to the plasma membrane as a likely site of toxicity because amyloid aggregates are known to be toxic no matter which side of the membrane they are on. First author Rakez Kayed and colleagues tested various forms of Aβ1-42 species for their ability to alter conductance of a planar lipid bilayer. They found that monomers, dimers, and fibrils had no effect, but oligomers, specifically those forming spherical particles of 3-5 nanometers in diameter, dramatically increased current through the membrane. The effect was rapidly reversed when the authors added an anti-Aβ oligomer antibody. This suggests two things, the authors write. One, when they interact with the membrane, oligomers of Aβ do not change to a conformation unrecognized by the antibody. And two, the oligomers are likely associating with the periphery of the membrane, not becoming embedded in it.

How these oligomers affect conductance is unclear. However, they do not appear to be targeted to specific ion channels because Kayed and colleagues found that the conductance elicited is not ion selective. Previous reports have indicated that Aβ can form ion channels in the membrane and that these are inhibited by Congo red (see for example Kagan et al., 2002), but the authors found that this dye had no effect on the conductance. “Rather than forming pores or channels, soluble oligomers appear to enhance the ability of ions to move through the lipid bilayer on their own,” write the authors. The difference between these and previous experiments that found Aβ channel conductance could be the absence of significant numbers of soluble amyloid oligomers in the earlier work, the authors suggest.

The effect on lipid bilayer conductance is not limited to oligomers of Aβ. When Kayed and colleagues carried out the same experiments with oligomers of α-synuclein, IAPP, a poly Q peptide, and a prion peptide, they found that soluble oligomers of these peptides, but again not low-molecular-weight species or fibrils, caused dramatic increases in membrane conductance. Hence, “this effect is specific for the particular conformation associated with spherical oligomers,” note the authors. The results also suggest a common mechanism of pathogenesis for a variety of amyloidogenic peptides.—Tom Fagan


  1. The authors report that soluble oligomers of Aβ and other protein misfolding disease related peptides such as islet amyloid polypeptide (IAPP, diabetes), PrP106-126 (prion diseases), polyglutamine (PG, Huntington’s and other triplet repeat diseases) and α-synuclein (Parkinson’s disease) can dramatically increase the ionic permeability of planar lipid bilayers. This result is a valuable confirmation of the work of numerous other labs showing that at least eight amyloid* peptides can increase the permeability of synthetic (Aβ, Arispe et al., 1993; IAPP, Mirzabekov et al., 1996; PrP106-126, Lin et al., 1997; atrial natriuretic factor, Kourie et al., 2001a; β-2-microglobulin, Hirakura and Kagan, 2001; serum amyloid A, Hirakura et al., 2002; polyglutamine, Hirakura et al., 2000b; HypF, Relini et al., 2004; α-synuclein, Volles and Lansbury, 2002; Azimova and Kagan, 2003; PrP82-146, Bahadi et al., 2003; calcitonin, Stipani et al., 2001) and natural membranes (Bhatia et al., 2000; Furukawa et al., 1994; Ibrahim et al., 2001; Janson et al., 1999; Kawahara et al., 1997; Kim et al., 2002; Simmons and Schneider, 1993). An antibody developed by the authors, which recognizes only the soluble oligomer conformation, is used to neutralize and even reverse the permeability increase. This suggests that the soluble oligomeric conformation, previously shown to be required for cytotoxicity (Kayed et al., 2003), is critical for the permeability increase. This supports early results which suggested that amyloid monomers and fibrils could not permeabilize membranes, but that oligomeric intermediates could (Sanderson et al., 1997; Hirakura et al., 1998; 1999, 2000; Volles and Lansbury, 2002).

    *I shall loosely use the term “amyloid” to include polyglutamine and α-synuclein as well as the more classical amyloid peptides. Unlike the many previous reports of amyloid peptides increasing membrane permeability, in these experiments the authors find no evidence of discrete channel formation and suggest that the peptides increase permeability through a different molecular mechanism.

    Overall, this work lends additional weight to the “channel hypothesis” of amyloid diseases, first proposed by Arispe et al. (1993a). The experiments demonstrate once again that Aβ1-40 and 1-42, and other amyloids, can increase membrane permeability, and this solidifies the connection between the cytotoxic soluble oligomer conformation and the membrane permeabilizing conformation. This is in agreement with earlier studies which showed that only Aβ variants which could permeabilize membranes could also be cytotoxic (Pike et al., 1993) or capable of inhibiting long-term potentiation (LTP) in rat hippocampus (Chen et al., 2000). It also concurs with experiments showing that “fresh and globular” (i.e., oligomeric) Aβ was cytotoxic and capable of membrane permeabilization, while aggregated or monomeric Aβ was not (Lin et al., 1999; Zhu et al., 2000; Bhatia et al., 2000).

    The channel hypothesis suggests that membrane permeabilization by amyloid peptides is the mechanism of amyloid cytotoxicity and amyloid-induced cellular dysfunction. (See Kagan et al., 2002, for review.) The locus of critical membrane damage by amyloid peptides remains unclear, but both the plasma membrane and the mitochondrial membrane have been suggested as prime targets. Neuronal plasma membranes require constant ionic gradients to maintain electrical activity. The “leakage” induced by amyloid peptides could easily run down these ionic gradients and lead to depolarization of neurons. This leakage would also tax the neuron’s energy stores as ATP was used to regenerate ionic gradients. The energetically vulnerable cells in areas such as the hippocampus would be most sensitive to this kind of metabolic insult.

    Ionic gradients and membrane potentials are also key to the functioning of mitochondria, the cell’s energy factories. It is of interest that many proteins involved in the regulation of apoptosis such as Bax and Bcl-2 are also able to increase membrane permeability via channel formation (Reed, 2000). Perhaps amyloid peptides somehow interfere with the regulation of apoptosis by altering mitochondrial membrane permeability.

    Panov et al. (2002) has shown that mitochondria from Huntington’s disease (HD) patients are depolarized, and that polyglutamine (PG) is capable of depolarizing mitochondria. In addition, the PG tract of the mutant huntingtin protein was found to be localized to mitochondrial membranes, and the mitochondrial depolarizing effects of PG can be replicated with a non-huntingtin fusion protein containing PG.

    It is more complicated to know the meaning of the failure of Kayed et al. to observe discrete ion channels. Over a dozen reports by six different laboratories have observed discrete ionic channels by electrical measurements (Arispe et al., 1993,a, b, 1996; Mirzebekov et al., 1996; Lin et al., 1997; Kourie et al., 2001; Hirakura and Kagan, 2001; Hirakura et al., 2002, 2000b; Azimova and Kagan, 2003; Stipani et al., 2001; Janson et al., 1999; Kawahara et al., 1997). Four other reports by three separate groups have observed annular pore-like structures using electron microscopy (Lashuel et al., 2002, 2003; Malisauskas et al., 2003; Zhu et al., 2002). Another report has observed pore-like sieving of permeants by amyloid peptides in lipid vesicles (Volles and Lansbury, 2002). How can the observations of Kayed et al. be reconciled with this large mass of contrary data?

    One possibility is that the lipid environment surrounding amyloid peptides is different. Kayed et al. used a mixture of phosphatidyl ethanolamine (PE), phosphatidylserine (PS), phosphatidylcholine (PC) and cholesterol. (1:1) as their lipid. Other experiments used a variety of other lipid mixtures. It is unfortunately well known from the earliest days of single channel observations that channel forming peptides will form electrophysiologically observable single channels in some lipids but not others (Wooley and Wallace, 1992). There is no adequate theory to explain this well reproduced observation. It is similarly possible that single channels were formed in these mixtures, but were too small in electrical conductance or too fast in opening and closing to be observed.

    A second possibility involves the boundary conditions of planar lipid bilayer formation. Kayed et al. employed “solvent-free” membranes formed from the union of two lipid monolayers. Some of the other laboratories used “solvent containing” membranes. These two types of membranes differ in various physical/chemical properties such as thickness, fluidity, capacity, etc. These properties could affect the ability to observe discrete channels.

    A third source of variation could lie in the peptide preparation and state of aggregation. The relatively pure and homogeneous preparation of oligomers of Kayed et al. may differ in subtle ways from the more heterogeneous amyloid preparations of the other investigators. The molecular weight of 90-110 kd suggests that 20-25 Aβ molecules comprise each “oligomer.” These might be larger oligomers than those used in other previous electrophysiological experiments.

    It must also be noted that planar lipid bilayers are usually composed of lipids different from those in natural neuronal membranes, and that they always lack the proteins, gangliosides, sterols, etc., found in natural membranes. These elements could play a key role in the detailed mechanism by which membrane permeability is enhanced. Cholesterol, for example, plays a critical receptor role in the channel forming ability of the polyene antifungal agents nystatin and amphotericin B as well as the Clostridial perfringolysin O cytotoxin (Bolard, 1986; Ramachandran et al., 2004). The cholesterol content in the membranes of Kayed et al. could explain their failure to see discrete pores.

    The conductance observed by Kayed et al. was also not blocked by Zn+2, as other amyloid channel experiments have reported. Since Zn+2 blocks Aβ cytotoxicity, this finding is puzzling. Congo red failed to affect the amyloid induced permeability, as previously reported, but it is unclear if it was able to inhibit it prior to contact with the membrane, as was previously shown (Hirakura et al., 1999). Since Congo red can slow the course of prion disease in animals (Ingrosso et al., 1995), it would be useful to know if it can prevent the formation of soluble oligomers and their membrane permeabilization effects.

    The ability of channel blocking compounds to inhibit the deleterious effects of amyloid peptides also argues for the existence of discrete channels. Not only can the channel blocker Zn+2 protect cells from Aβ toxicity, but the PrP106-126 channel blocker quinacrine can protect cells (Farelly et al., 2003) and may have some positive clinical effects in humans with Creutzfeld-Jakob disease (Nakajima et al., 2004).

    Recent studies have also shown that several distinct types of Aβ and PrP106-126 channels can be characterized electrophysiologically (Kourie and Culverson, 2000; Kourie et al., 2001b), and that cholesterol content in the membrane controls cytotoxicity and channel forming ability in a similar manner (Arispe and Doh, 2002). Indeed, Arispe (2004) has demonstrated that certain peptides specifically interact with the Aβ pore region. The concurrence of electrophysiologic, ultrastructural, and cytoxicity evidence seems to strongly support channel formation as the pathogenic mechanism of amyloid diseases.

    Finally, the permeability increase reported here differs from previous reports in showing no ionic selectivity (i.e., they report a reversal potential of zero in ionic gradients). This is a curious finding. Even a permeability increase of no intrinsic ionic selectivity should show significant cation/anion selectivity due to the substantial positive potential in the middle of the lipid bilayer relative to aqueous solution (This potential is induced by the dipole moments of the lipid molecules, and would render cations much less likely to cross the membrane). A competing effect is the concentration of cations at the membrane surface by the net negative surface potential generated by the charged lipid head groups. It would be an amazing coincidence for the intrinsic selectivity of the amyloid peptides to exactly cancel out these effects.

    If channels are not responsible for the permeability increases reported, what is? A “carpet mechanism” has been proposed but not proven for antibiotic peptides which kill bacteria (Shai, 2002). Kayed et al. suggest that amyloid peptides “stretch" the bilayer thus making it more permeable to ions. If this mechanism could be shown to occur, it would represent a novel finding in biology.

    In summary, this work provides additional support for the notion that membrane permeabilization (whether or not by channel formation) by amyloid peptides is the cellular mechanism of amyloid disease. A substantial body of evidence points to channel formation as the molecular mechanism. Additional biophysical work in vivo and in vitro will be required to elucidate the details of this mechanism.


    . Alzheimer disease amyloid beta protein forms calcium channels in bilayer membranes: blockade by tromethamine and aluminum. Proc Natl Acad Sci U S A. 1993 Jan 15;90(2):567-71. PubMed.

    . Giant multilevel cation channels formed by Alzheimer disease amyloid beta-protein [A beta P-(1-40)] in bilayer membranes. Proc Natl Acad Sci U S A. 1993 Nov 15;90(22):10573-7. PubMed.

    . Zn2+ interaction with Alzheimer amyloid beta protein calcium channels. Proc Natl Acad Sci U S A. 1996 Feb 20;93(4):1710-5. PubMed.

    . Ion channels formed by a fragment of α-synuclein (NAC) in lipid membranes. Biophys J. 2003 Feb;84(2):53a.

    . Cu2+-induced modification of the kinetics of A beta(1-42) channels. Am J Physiol Cell Physiol. 2003 Oct;285(4):C873-80. PubMed.

    . Fresh and globular amyloid beta protein (1-42) induces rapid cellular degeneration: evidence for AbetaP channel-mediated cellular toxicity. FASEB J. 2000 Jun;14(9):1233-43. PubMed.

    . How do the polyene macrolide antibiotics affect the cellular membrane properties?. Biochim Biophys Acta. 1986 Dec 22;864(3-4):257-304. PubMed.

    . Impairment of hippocampal long-term potentiation by Alzheimer amyloid beta-peptides. J Neurosci Res. 2000 Apr 1;60(1):65-72. PubMed.

    . Amyloid beta protein-induced irreversible current in rat cortical neurones. Neuroreport. 1994 Oct 27;5(16):2016-8. PubMed.

    . Pore formation by beta-2-microglobulin: a mechanism for the pathogenesis of dialysis associated amyloidosis. Amyloid. 2001 Jun;8(2):94-100. PubMed.

    . Channel formation by serum amyloid A: a potential mechanism for amyloid pathogenesis and host defense. Amyloid. 2002 Mar;9(1):13-23. PubMed.

    . Membrane perturbation by the neurotoxic Alzheimer amyloid fragment beta 25-35 requires aggregation and beta-sheet formation. Biochem Mol Biol Int. 1998 Nov;46(4):787-94. PubMed.

    . Alzheimer amyloid abeta1-42 channels: effects of solvent, pH, and Congo Red. J Neurosci Res. 1999 Aug 15;57(4):458-66. PubMed.

    . Amyloid peptide channels: blockade by zinc and inhibition by Congo red (amyloid channel block). Amyloid. 2000 Sep;7(3):194-9. PubMed.

    . A helix-loop-helix peptide at the upper lip of the active site cleft of lysozyme confers potent antimicrobial activity with membrane permeabilization action. J Biol Chem. 2001 Nov 23;276(47):43767-74. PubMed.

    . The mechanism of islet amyloid polypeptide toxicity is membrane disruption by intermediate-sized toxic amyloid particles. Diabetes. 1999 Mar;48(3):491-8. PubMed.

    . The channel hypothesis of Alzheimer's disease: current status. Peptides. 2002 Jul;23(7):1311-5. PubMed.

    . Alzheimer's disease amyloid beta-protein forms Zn(2+)-sensitive, cation-selective channels across excised membrane patches from hypothalamic neurons. Biophys J. 1997 Jul;73(1):67-75. PubMed.

    . Common structure of soluble amyloid oligomers implies common mechanism of pathogenesis. Science. 2003 Apr 18;300(5618):486-9. PubMed.

    . Amyloid beta peptide induces cytochrome C release from isolated mitochondria. Neuroreport. 2002 Oct 28;13(15):1989-93. PubMed.

    . Properties and modulation of alpha human atrial natriuretic peptide (alpha-hANP)-formed ion channels. Can J Physiol Pharmacol. 2001 Aug;79(8):654-64. PubMed.

    . Diversity of amyloid beta protein fragment [1-40]-formed channels. Cell Mol Neurobiol. 2001 Jun;21(3):255-84. PubMed.

    . Neurodegenerative disease: amyloid pores from pathogenic mutations. Nature. 2002 Jul 18;418(6895):291. PubMed.

    . Mixtures of wild-type and a pathogenic (E22G) form of Abeta40 in vitro accumulate protofibrils, including amyloid pores. J Mol Biol. 2003 Sep 26;332(4):795-808. PubMed.

    . Channel formation by a neurotoxic prion protein fragment. J Biol Chem. 1997 Jan 3;272(1):44-7. PubMed.

    . Amyloid beta protein (1-40) forms calcium-permeable, Zn2+-sensitive channel in reconstituted lipid vesicles. Biochemistry. 1999 Aug 24;38(34):11189-96. PubMed.

    . Amyloid protofilaments from the calcium-binding protein equine lysozyme: formation of ring and linear structures depends on pH and metal ion concentration. J Mol Biol. 2003 Jul 18;330(4):879-90. PubMed.

    . Pore formation by the cytotoxic islet amyloid peptide amylin. J Biol Chem. 1996 Jan 26;271(4):1988-92. PubMed.

    . Early mitochondrial calcium defects in Huntington's disease are a direct effect of polyglutamines. Nat Neurosci. 2002 Aug;5(8):731-6. PubMed.

    . Neurodegeneration induced by beta-amyloid peptides in vitro: the role of peptide assembly state. J Neurosci. 1993 Apr;13(4):1676-87. PubMed.

    . Membrane-dependent conformational changes initiate cholesterol-dependent cytolysin oligomerization and intersubunit beta-strand alignment. Nat Struct Mol Biol. 2004 Aug;11(8):697-705. PubMed.

    . Mechanisms of apoptosis. Am J Pathol. 2000 Nov;157(5):1415-30. PubMed.

    . Monitoring the process of HypF fibrillization and liposome permeabilization by protofibrils. J Mol Biol. 2004 May 14;338(5):943-57. PubMed.

    . Aggregates of a beta-amyloid peptide are required to induce calcium currents in neuron-like human teratocarcinoma cells: relation to Alzheimer's disease. Brain Res. 1997 Jan 2;744(1):7-14. PubMed.

    . Mode of action of membrane active antimicrobial peptides. Biopolymers. 2002;66(4):236-48. PubMed.

    . Amyloid beta peptides act directly on single neurons. Neurosci Lett. 1993 Feb 19;150(2):133-6. PubMed.

    . Channel formation by salmon and human calcitonin in black lipid membranes. Biophys J. 2001 Dec;81(6):3332-8. PubMed.

    . Vesicle permeabilization by protofibrillar alpha-synuclein is sensitive to Parkinson's disease-linked mutations and occurs by a pore-like mechanism. Biochemistry. 2002 Apr 9;41(14):4595-602. PubMed.

    . Model ion channels: gramicidin and alamethicin. J Membr Biol. 1992 Aug;129(2):109-36. PubMed.

    . Surface-catalyzed amyloid fibril formation. J Biol Chem. 2002 Dec 27;277(52):50914-22. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Amyloid Formation—Taking Things One Monomer at a Time
  2. "Arctic" APP Mutation Supports Protofibril Role
  3. Post-Conference News: Oligomers, Protofibrils, and Fibrils—They Are All Bad

Paper Citations

  1. . The channel hypothesis of Alzheimer's disease: current status. Peptides. 2002 Jul;23(7):1311-5. PubMed.

Further Reading

Primary Papers

  1. . Permeabilization of lipid bilayers is a common conformation-dependent activity of soluble amyloid oligomers in protein misfolding diseases. J Biol Chem. 2004 Nov 5;279(45):46363-6. PubMed.