Over the past decade, researchers have shifted away from a literal interpretation of Alois Alzheimer’s groundbreaking discovery of plaques and tangles as the likely cause of Alzheimer disease. After years of argument—mostly in the 1990s—about whether plaques or tangles were the culprit, the answer appears to be, “Both and neither.” How can that be true? Scientists have recognized that both constituent proteins of those hallmark pathologies—the amyloid-β (Aβ) peptide and tau—play essential roles in the development of the disease, relegating the Baptist-Tauist divide solidly to the past. That’s the “both” part. But scientists also increasingly agree that the microscopically visible protein deposits are not the worst offenders: hence, the “neither.” Instead, they blame smaller, oligomeric forms of the Aβ peptide that they believe exist in a complex equilibrium with higher-order protofibrils along a path to aggregation. These, they say, damage synapses and interfere with cognitive function. In short, they say plaques are bad, but oligomers are worse. For tau, this story isn’t nearly as far along, but trends suggest that it may well develop along similar lines. (And ditto for α-synuclein.)

The Society for Neuroscience conference, held 3-7 November in San Diego, was a testament to how deeply the science of Aβ oligomers has taken hold in the field. There were some 35 presentations about Aβ species variably called oligomers, ADDLs, AβOs, or protofibrils. Speakers increasingly cited the “Amyloid Oligomer Hypothesis” rather than the “Amyloid Hypothesis” in the introductory slide of their talk. Indeed, a range of presentations from a diverse group of labs reported data largely concurrent with its essential tenet that AD begins with synaptic dysfunction caused by soluble Aβ species. Here are selected highlights.

Perhaps the most direct support came from Ganesh Shankar, an M.D.-Ph.D. student working with a team of colleagues in Dennis Selkoe’s laboratory and Cindy Lemere at Brigham and Women’s Hospital, Boston, Dominic Walsh’s group and Ciaran Regan’s group, both at University College in Dublin, Ireland, and with Bernardo Sabatini at Harvard Medical School. In a sparsely attended slide session on the last afternoon of the conference, Shankar expanded on what a poster presented by Shaomin Li from the same team had foreshadowed days before. The scientists isolated soluble Aβ species from cortex of human AD brain, and report that oligomers as small as a dimer recapitulated the synaptotoxic effect the scientists had previously published for similar small oligomers secreted by cultured cells.

Prior studies from several laboratories have consistently found synaptotoxic effects for various forms on Aβ oligomers (e.g., Walsh et al., 2002—from conditioned media of 7PA2 Chinese hamster ovary cells; Lambert et al., 1998—from synthetic Aβ42; Lesne et al., 2006—from Tg2576 mouse brain). Yet these studies begged the question of how relevant to human Alzheimer disease all this can be until human Aβ oligomers are in hand. To address this question, Shankar and colleagues obtained postmortem cortical tissue from several patients with late-onset AD (one of whom had had no clinical AD but pathological AD upon autopsy). As controls, the scientists used cortex from patients with Lewy body dementia (LBD—they get parkinsonism and dementia at about the same time and are thought to have mixed pathologies), Down syndrome (who have typical AD-type amyloid pathology), and frontotemporal and multi-infarct dementia (who do not). Readily detectable soluble Aβ showed up in cortex from all clinically demented AD patients but not in one cognitively normal person who had the plaque pathology. It also showed up in the Down brain, and to a much smaller extent in the LBD brain. Curiously, soluble extracts from normal control brains appeared to contain very little or no soluble monomeric Aβ by this immunoprecipitation/Western blot assay, even though the brain presumably produces some all the time.

These AD cortical extracts were made merely in TBS buffer without detergent, and they showed primarily monomer at a weight of 4 kDa and dimer at 8 kDa. Extracts made in parallel with detergent also had monomer and dimer in them. Shankar showed experiments suggesting that besides the dimer, soluble Aβ extract from human AD brain also contains complexes having a larger molecular weight—either Aβ aggregated with itself or bound to other proteins—but that these fall apart upon treatment with detergent. This is a technical difference with studies on ADDLs and Aβ*56, both of which are reported to be SDS-stable. Shankar said that his colleagues and he searched for SDS-stable species in the human extracts but so far have been unable to find any that are larger than trimer. Shankar and colleagues used detection by two antibodies that detect the free N- and C-terminus of Aβ, respectively, and also used mass spectrometry, to ascertain that the dimers contained true Aβ, and to exclude any other Aβ-containing APP cleavage fragments that might be contained in the extracts.

Next, the scientists applied their preparations to tests of LTP and spine integrity that they had developed previously. The TBS extracts from AD brains blocked LTP induction, whereas extracts from the other diseases and age-matched controls did not, Shankar reported. (The Down’s extract was not tested.) Immunodepletion of Aβ restored LTP, meaning the effect was specific to Aβ. The effect was potent, acting in the picomolar range. By enriching Aβ through immunoprecipitation, eluting with SDS buffer and then running on size exclusion chromatography, only the fraction enriched for Aβ dimers inhibited LTP significantly; the monomer had no effect. The soluble AD brain extract also facilitated long-term depression, reducing neuronal excitability after a period of stimulation. The main point, Shankar said, is that Aβ dimer extracted from human AD brain is sufficient to disrupt the molecular basis of learning and memory. It is not the only form, but the smallest form that can be toxic.

Various anti-Aβ antibodies are in clinical trials at present, and one debate in the field revolves around which type of antibody might be most potent. Shankar and colleagues indirectly addressed this debate by testing which of the three classes of anti-Aβ antibody used in those trials—N-terminal, mid-region, C-terminal—was best able to rescue the detrimental effect on LTD of the human AD extract. In a subtraction experiment, where the investigators selectively depleted the extract with only one kind of antibody, N-terminal antibodies best protected LTP, Shankar reported. This electrophysiology result concurred with associated biochemistry, in that the N-terminal antibodies also captured the most Aβ from the extract. (Not all immunotherapy clinical trials, however, are based on the premise of directly counteracting Aβ oligomers in brain; some aim to draw down Aβ from the periphery, or target Aβ more generally.)

Beyond LTP and LTD, do these human oligomers really matter to the structure of synapses? There is strong consensus in the field that synapses in AD-relevant brain areas gradually decrease in number early on as people develop cognitive symptoms (Davies et al., 1987; Scheff et al., 2007). At Neuroscience, Shankar showed evidence that the human AD oligomers reduced the density of dendritic spines in cultured brain slices in much the same way as cell-secreted oligomers do (Shankar et al., 2007). Furthermore, Shankar showed data on a rat behavioral test. The human AD extract impaired learning in a passive avoidance paradigm. It did so when infused 3 hours after the rats had initially learned, the period that other studies have identified as the time when synapses undergo remodeling following learning.

Finally, the researchers reported taking a hard crack at the Aβ dimers. Acting on a hunch that the dimers might represent a seed for plaque formation, the team isolated mature, cored plaques and removed as many associated components from them as possible by repeated washes in detergent and TBS buffer. This left behind insoluble, microscopic cores that stained with Congo red. These cores did not inhibit LTP. They were very hardy, but when the scientists blasted them apart with highly concentrated formic acid, Aβ dimers were released, and those did inhibit LTP. Taken together, these investigators interpret their data to mean that soluble Aβ oligomers from typical AD patients, starting with dimers, disrupt synaptic function in humans, and that insoluble cores sequester these species. As to plaques, they represent a reservoir of soluble Aβ in a given brain region, Shankar said. For their part, the dimers would seem to be a relevant substrate for both research into the molecular pathways of synaptic impairment, and also for testing prospective therapeutic agents preclinically, Shankar added.

Other labs need to replicate these findings. When asked whether the recipe for isolating the human oligomers was technically difficult, he replied: “No, it’s pretty much standard biochemistry. But rigorous clinical and histopathological information on the patients should be available before attempting it, so close interaction with a brain bank is key.”—Gabrielle Strobel.

This is Part 1 of a three-part story. See Parts 2 and 3.


Make a Comment

To make a comment you must login or register.

Comments on this content

  1. The work of Shankar and colleagues provides new evidence supporting the concept that soluble Aβ oligomers disrupt synaptic function in Alzheimer disease. The recent publication from Rowan, Wang, and their colleagues (Rowan et al., 2007) suggesting that synaptic dysfunction caused by Aβ oligomers is mediated by TNF-α is highly relevant. This new publication extends Rowan and Wang’s previous work, which suggested that β amyloid inhibition of LTP is mediated via TNF (Wang et al., 2005). In Rowan and Wang’s most recent paper, experimental evidence is presented that pretreatment with a biologic inhibitor of TNF-α to neutralize TNF-α prevented Aβ inhibition of LTP induction at medial perforant pathway synapses.

    These are observations of great importance, because they help bridge the gap between the amyloid hypothesis and the neuroinflammatory hypothesis of AD. These interrelated mechanisms may help explain the positive clinical effects my colleagues and I have observed using anatomically targeted anti-TNF treatment in AD (Tobinick et al., 2006) and underscores the need to further investigate this treatment approach (see also Tweedie et al., 2007).


    . Synaptic memory mechanisms: Alzheimer's disease amyloid beta-peptide-induced dysfunction. Biochem Soc Trans. 2007 Nov;35(Pt 5):1219-23. PubMed.

    . Beta-amyloid inhibition of long-term potentiation is mediated via tumor necrosis factor. Eur J Neurosci. 2005 Dec;22(11):2827-32. PubMed.

    . TNF-alpha modulation for treatment of Alzheimer's disease: a 6-month pilot study. MedGenMed. 2006;8(2):25. PubMed.

    . TNF-alpha inhibition as a treatment strategy for neurodegenerative disorders: new drug candidates and targets. Curr Alzheimer Res. 2007 Sep;4(4):378-85. PubMed.

  2. Regarding the data presented by Dennis Selkoe's group: when they isolate the oligomeric material from AD brain, they separate it by a Sephadex75 column. The material goes with the void in that column (“fraction 4”), which means that it is larger than 65 kDa. Our synthetic protofibrils behave in the same way on the same column. Selkoe’s group then immunoprecipitate the material and run it on Western blot, where it appears as a dimer.

    There are at least two explanations for the difference observed:

    1. The dimer is broken down from a larger oligomeric species through immunoprecipitation and Western blot.

    2. The dimer is bound to a larger protein which gives a molecular weight of more than 65 kDa.

    [Editor's note: see Oligo report Part 2]

  3. I would like to further detail some of the statements present in Dr. Pimplikar's comments regarding our studies using human brain tissues (20 Non-Cognitively Impaired, 10 Mild Cognitively Impaired and 10 AD selected from the Religious Order Study by Dr. David Bennett, director of the program).

    It is true that during my Minisymposium talk, I reported a greater than 3-fold increase in Aβ*56 levels in brains of individuals clinically diagnosed with MCI or AD. Aβ*56 levels in both MCI and AD groups were not different compared to each other, suggesting that Aβ*56 may be a molecule initiating Aβ-induced cognitive decline. I also mentioned that we did not observe changes in levels of soluble monomeric Aβ, nor in levels of Aβ trimers.
    Finally, we reported that cerebral levels of Aβ*56 are inversely correlated with MMSE score, while soluble Aβ monomers, Aβ trimers, or amyloid burden were not associated with neurological status.

    As for our poster presentation, we demonstrated that Aβ*56 was not associated with changes in levels of synaptophysin or drebrin (among other pre- and postsynaptic markers), suggesting that Aβ*56 is not triggering synaptic loss (contrary to 7PA2 CM-derived Aβ dimers/trimers). In addition, we reported that in human tissues, trimeric Aβ levels were dependent on monomeric Aβ levels. Finally, Aβ dimers were not detected in soluble fractions of protein extracts from human or transgenic mouse brains.

  4. At this moment, perhaps the greatest contribution to the field of AD would be focusing on efforts to further define and compare the various preparations of amyloid-β (Aβ) aggregates, continuing research in the vein of recent SfN presentations from the Ashe/Cleary and Selkoe groups. The nomenclature for oligomers is inconsistent at best. The Aβ assemblies/aggregates preparations studied by particular investigators are defined by numerous methods, including neurotoxic activities, isolation technique (primarily size exclusion chromatography), size estimation by SDS or native PAGE, and several imaging techniques. In addition, reactivity with various Aβ conformation-specific antibodies is now also being used to identify specific species of Aβ. Thus, comparison of results across different preparations of Aβ oligomers is virtually impossible. Establishing a common series of definitions and encouraging future publications to work within these established parameters would greatly advance the study of the relationship between Aβ structure and function.

  5. Alzheimer Disease, Aβ Oligomers, and Shrek
    Gabrielle Strobel and Alzforum should be congratulated on bringing to our attention the excitement the “amyloid oligomer hypothesis” has generated in the AD field. Her three-part presentation (Oligomers Live Up to Bad Reputation) summarizes the enormous amount of data presented at the meeting and leaves little doubt that “oligomer” is the buzzword of today.

    That the three oligomeric forms of Aβ (7PA2 derived small oligomers; high “n”-oligomers termed ADDLs; and “star”-oligomers) exhibit deleterious effects at various concentrations, in various experimental paradigms, is not surprising, and perhaps, not significant. After all, the literature of the 1990s is littered with reports of Aβ monomers or fibrils being toxic to cells. What is important (to come out of the San Diego meeting) is that two studies found the presence of Aβ oligomers in the AD brains but not in the control tissues. Surely, this should silence the critics, right?

    In oral and poster presentations, Lesne et al. reported increased levels of Aβ*56 oligomers in brains from AD patients but apparently did not find a similar increase in the small oligomers. Conversely, Shankar et al. found increased levels of Aβ dimers/trimers in AD patients but detected no star oligomers. If Aβ oligomers are the causative factors of the disease and exist in AD brains, one wonders whether this is too much of a coincidence that these two groups observed only “their” type of oligomers from AD brains.

    An important assumption underlying both these studies is that the oligomeric forms of Aβ, detected at the end of a Western blot protocol, already exist in the diseased tissue prior to homogenization/isolation/purification/detection. However, Aβ is an amphipathic peptide (hydrophobic at one end and hydrophilic at the other), and work by Teplow, Bitan, and colleagues has conclusively shown how easily Aβ can create different higher-order multimers depending on experimental conditions. Some may consider this to be hypercritical, but the observation that two productive and respected groups in the field find only their favorite oligomers in AD brains should raise alarm bells: do oligomers exist in vivo or are they created by the very experimental manipulations that are used to detect their presence? Surely, two different protocols are likely to yield two different higher-order forms.

    It is a common observation that naysayers are often irritating and sometimes wrong, but objective, dispassionate criticism is essential for the relevance of the amyloid hypothesis (which has been instrumental in promoting AD research) to AD pathogenesis. Shrek, the large, green, intimidating giant also caught our attention and we have come to like it. However, unless ogres exist in real life, Shrek adds little value to our lives other than entertainment.


    . Neurotoxic protein oligomers--what you see is not always what you get. Amyloid. 2005 Jun;12(2):88-95. PubMed.

  6. Our lab has begun looking at Aβ oligomers in our mouse model and in vitro. To add to this series, our findings presented at the SfN meeting can be summarized as follows:

    1. We observe entry into the cell cycle (as evidenced by expression of cell cycle proteins and DNA replication by FISH) of selected neuronal populations in our APP YAC transgenic mouse model of AD at 6 months of age. This cell cycle entry is dependent upon amyloidogenic processing of APP and occurs about 6 months prior to Aβ deposition.

    2. We can identify the presence of Aβ oligomers at this age (bands on SDS-PAGE) recognized by both 6E10 and the oligomer-specific antibodies NU1 and A11, including the presence of dimers and trimers as well as higher-MW Aβ species.

    3. In-vitro preparations of oligomeric Aβ (prepared in Hams F12 media or purified via SEC) and, to a much lesser extent, monomeric Aβ, induced concentration-dependent aberrant neuronal cell cycle entry as measured by BrdU incorporation and expression of cell cycle proteins, in primary cortical neurons. Oligomeric Aβ also induced loss of neurites.

    4. Exposure to increasing concentrations of in-vitro preparations of oligomeric Aβ induced altered morphology of primary microglia, consistent with activation and similar to that observed with lipopolysaccharide (LPS). Conditioned media from oligomer-exposed or LPS-stimulated microglia also induced neuronal cell cycle entry, suggesting that Aβ oligomers may act both directly on neurons and perhaps indirectly through activation of microglia.

    [Editor's note: See also ARF conference story on cell cycle symposium.]

  7. Editor’s note: The Alzforum editors invited Bill Klein of Northwestern University’s Cognitive Neurology and Alzheimer’s Disease Center in Evanston, Illinois, to round off this series of SfN conference news and commentary. Readers who came late to the story can kick back and use Klein’s perspective on the biology and structure of Aβ oligomers as their frame of reference for this current coverage. Below, Klein offers an informal overview of some milestones, along with his take on today’s central questions. If these remarks whet your appetite, you’ll find an in-depth discussion of the broader topic in Klein’s chapter in Synaptic Plasticity and the Mechanism of Alzheimer’s Disease, Selkoe, Dennis J.; Triller, Antoine; Christen, Yves (Eds.), due out January 2008 from Springer.

    Oligomers as Alzheimer’s toxins.
    Thanks to the work of many labs, we now know that soluble Aβ oligomers are long-lived, neurologically active molecules, not simply intermediates in fibrillogenesis. Oligomers accumulate in brains and CSF of individuals afflicted by AD, where they are believed responsible for dementia-producing neuron damage (see, e.g., the “Pathway to Harm,” in the Progress Report on Alzheimer’s Disease published by the Department of Health and Human Services). Neurologically active oligomers have been given many names—we initially called them ADDLs, for Aβ-derived diffusible ligands. (The pronunciation, by the way, is “addles,” as in “Not paying attention to the Alzheimer Research Forum addles the brain.”) Their oligomeric structure endows ADDLs with the capacity to attack particular synapses, mainly at spines and near NMDA receptors. In essence, extracellular ADDLs act as gain-of-function pathogenic ligands. This capacity for highly specific synaptic targeting provides a putative mechanism to explain why AD is a disease of memory. Binding disrupts synaptic plasticity, causes overexpression of the memory-linked immediate early gene Arc, and triggers pathological changes in synapse shape and composition. Because ADDL binding also instigates synapse loss, oxidative damage, AD-type tau hyperphosphorylation, and selective nerve cell death, the attack on synapses provides a plausible mechanism unifying memory dysfunction with major features of AD neuropathology. Most recently, ADDLs were shown to trigger downregulation of synaptic insulin receptors, providing a mechanism to explain insulin resistance in AD brain (“type 3 diabetes”; Zhao et al., 2007). Acting as novel neurotoxins that putatively account for memory loss and neuropathology, ADDLs present significant targets for disease-modifying therapeutics in AD, with proof-of-concept already evident from animal models.

    Structurally speaking, Aβ is the peptide from hell.
    I’ve lost track of who wrote this first, but the truth still holds. Aβ is one-third hydrophobic, two-thirds hydrophilic, and almost 100 percent erratic in its biochemistry, which is why the problem of how Aβ produces Alzheimer’s dementia is still unsolved after almost 25 years. Darwin may have been thinking of something like Aβ when he said, “Nature will tell you a direct lie when she can.”

    The challenges in working with Aβ are exemplified by the remarkable contentiousness of the early 1990s regarding whether Aβ was or was not toxic—some evidence said yes, other evidence said no. Fortunately, despite Aβ’s recalcitrance, insightful research can outsmart it. Seminal work by Christian Pike and Carl Cotman, Alfredo Lorenzo and Bruce Yankner, and their colleagues, resolved the controversy by showing that Aβ preparations are indeed toxic but only if monomers undergo a process of self-assembly. Since their toxic solutions showed amyloid fibrils, it was concluded that toxicity required these emergent fibrils. The apparent requirement for fibrils, certainly consistent with Occam’s razor, provided major support for the original amyloid cascade hypothesis. It made good sense at the time that Alzheimer’s was a pathology of nerve cell death instigated by large insoluble amyloid fibrils.

    In keeping with this concept, the first reports of SDS-stable oligomers in AD brain regarded oligomers simply as subunits responsible for ongoing formation of fibrils. Colin Masters, Konrad Beyreuther, and colleagues found that formic acid extracts of isolated amyloid plaques contained dimers and tetramers, as well as a pH-sensitive presence of larger oligomers (Masters et al., 1985). Janusz Frackowiak et al. in 1994 showed SDS-stable dimers and tetramers in extracts of meningeal blood vessels; they concluded that during amyloid formation in AD vessel walls, non-fibrillar oligomers accumulate (Frackowiak et al., 1994). However, 1994 also brought a glimmer of an extremely different concept for structure and toxicity. Tomiichiro Oda and colleagues, working in Tuck Finch’s lab, mixed Aβ with clusterin (ApoJ) and found large fibrils were blocked from forming (Oda et al., 1994). They anticipated a protective effect. To their surprise, their non-pelleting material robustly impaired the ability of PC12 cells to metabolize MTT. They wrote, “Inhibition of Aβ aggregation and enhancement of Aβ toxicity by clusterin suggest new mechanisms in AD.” Supporting this possibility, Alex Roher and colleagues reported that dimers chemically extracted from amyloid deposits were capable of killing neurons via a mechanism requiring microglia (Roher et al., 1996).

    Stimulated by the clue provided by Oda et al., our group collaborated closely with Tuck Finch and Grant Krafft in studies that lead to the first ADDL paper (Lambert et al., 1998). Together, we identified small Aβ oligomers as a new type of neurotoxin structurally distinct from amyloid fibrils and protofibrils. We coined the name ADDLs to broadly cover this new class of soluble Aβ-derived molecules showing potent CNS neurotoxicity. We described three different preparative methods, and all yielded solutions of neurotoxic Aβ assemblies that were totally free of fibrils and protofibrils. By atomic force microscopy, ADDLs comprised globular structures roughly comparable in size to soluble proteins smaller than 50 kDa. SDS-PAGE with Tris-glycine gels indicated the assemblies were made of SDS-resistant tetramers and pentamers. Currently, Western blots using Tris-tricine gels with BioRad markers routinely show trimers, tetramers, and 12mers. Native, non-denaturing gels also showed fibril-free oligomers, so oligomers were neither SDS-induced nor products of protofibril breakdown. A later study confirmed that fresh Aβ42, never put in aqueous solution, migrates as monomer in SDS-PAGE (Chromy et al., 2003). In terms of structure-function, the bottom line was that certain conditions promote assembly of Aβ into potent CNS neurotoxins that comprise long-lived soluble oligomers.

    The Lambert paper revealed an aspect of ADDL activity that, because of its clear relevance to memory mechanisms, made this new toxin particularly exciting. Within minutes, and greatly in advance of cellular degeneration, ADDLs inhibited long-term potentiation (LTP). We hypothesized that memory dysfunction in early Alzheimer’s was the result of impaired synaptic plasticity caused by ADDLs.

    At the end of our Discussion we wrote, “ADDLs thus have profound neurological effects well in advance of tissue damage. If Aβ derivatives such as ADDLs prove to be part of Alzheimer’s pathogenesis, these results suggest that it would be theoretically feasible to halt or reverse the disease during its early stages.” Our prediction that memory loss could be reversed was confirmed 4 years later in mouse model passive immunization experiments by Steve Paul’s group at Eli Lilly (Dodart et al., 2002) and Karen Ashe’s group at Minnesota (Kotilinek et al., 2002). Most recently, the use of conformation-sensitive antibodies targeting non-monomeric Aβ in memory recovery experiments by Trojanowski and Lee specifically supported the pathogenic role of oligomers (Lee et al., 2006).

    A PubMed search of amyloid-β oligomer(s) now yields over 500 hits. Knowing what is implied structurally by “toxic oligomer” has become a central issue. There clearly is a wide variety of preparations, resultant structures, and even nomenclature. With respect to ADDLs, the name is generic and broadly encompasses oligomers associated with dementing activity. It was chosen to sharply distinguish AD-relevant globular assemblies from fibrillar toxins.

    A neurologically active ADDL preparation typically comprises two classes of oligomers by HPLC-SEC in aqueous buffer. Some migrate in a peak near putative 12mers and some migrate in a peak near putative 3- and 4mers. The relative abundance of the peaks varies from preparation to preparation. The two peaks also can be separated by ultrafiltration. Both peaks are SDS-resistant but not completely SDS-stable, as indicated originally in the Lambert paper. Silver stains from each peak show prominent 4mers, 3mers, and monomers; this is the case even for the larger peak, which contains oligomers that do not pass a 50 kDa cutoff filter in aqueous solution. Ultrafiltration actually indicates that ADDL preparations in physiological buffer contain almost undetectable amounts of monomer (compared, e.g., with positive controls using Aβ40 monomers). The fraction comprising larger oligomers, when analyzed by Western blots with our conformation-sensitive antibodies, shows 12mers. In overexposed Western blots of ADDLs, it is possible to detect a full spectrum of oligomers up to 24mers, although particular species (3- and 4mers, and 12mers) are favored in a temperature-dependent manner (Klein, 2002). We note that some non-toxic Aβ preparations also show oligomers (Chromy et al., 2003), so quality control monitoring both structure and function is essential. Greatly increased SDS-resistance in the 12mer fraction is promoted by incubation with certain prostaglandins or levuglandins (Boutaud et al., 2002; Boutaud et al., 2006) or by copper and H202 (Atwood et al., 2004).

    The fraction with 12mers is particularly interesting to us, because it contains the most striking ligand activity, detected as binding to particular synapses in cell biology experiments (Lacor et al., 2004; 2007). The 12mer fraction also gives the most robust induction of drebrin loss and synaptic spine degeneration (Lacor et al., 2007). Regarding our fraction of 3- and 4mers, current data suggest much less binding and pathogenicity, although absence of evidence is not necessarily evidence of absence. We are investigating possible neuronal responses to the small ADDLs. The findings of Walsh, Selkoe, and colleagues are extremely significant and certainly substantiate a robust pathogenicity for trimers produced by cellular metabolism.

    Oligomers in AD brain.
    The salient issue is the nature of toxins in human brain—what are the neurologically significant oligomers in AD-afflicted brain tissue? Can these molecules explain why early Alzheimer’s is a disease of memory, and can they account for AD’s other major pathologies?

    Our studies of human brain samples using physiological buffer for gentle extraction showed that human brain-derived ADDLs and synthetic ADDLs are biologically and structurally equivalent (Gong et al., 2003). Whether obtained from Alzheimer-affected brain or prepared in vitro, ADDLs act in cell biology experiments as specific ligands that bind to particular synapses (Lacor et al., 2004), a specificity that strongly suggests equivalent conformation in solution. After binding, both ADDLs stimulate AD-type hyperphosphorylation of tau (De Felice et al., 2007). With respect to direct structural comparisons (Gong et al., 2003), human ADDLs are readily detected by conformation-sensitive antibodies that we previously generated against synthetic ADDLs (Lambert et al., 2001). The antibodies, which recognize assembled Aβ but not monomers, bind human and mouse brain-derived and synthetic ADDLs in blots, in solution, and when attached as ligands to synapses, consistent with conformational equivalence. The antibodies also prevent both types of ADDLs from binding to synapses and triggering tau hyperphosphorylation. With respect to precedents for conformational-sensitive antibodies, several years earlier Austin Yang, Charlie Glabe, and colleagues found that antibodies they produced could discriminate in Western blots between monomers and larger structures they called insoluble amyloidogenic fragments—which we now would recognize as Aβ oligomers. They noted their blots were consistent with conformational epitopes being uniquely present in the larger structures (Yang et al.,1995).

    When we used 2D gel analysis to look further at structure, we found that both brain-derived and synthetic ADDLs show pIs of 5.6. Of greatest interest, given current attention to oligomer size, both brain-derived and synthetic ADDLs were found to comprise prominent 12mers (54 kDa).

    The fact that AD brain manifests 12mers that act as pathogenic synaptic ligands is particularly significant given the subsequent detection of neurologically active 12mers in transgenic mouse models (Lesne et al., 2006). The mouse 12mers become detectable roughly coincident with the onset of memory dysfunction. The 12mers in the mouse model have been referred to as Aβ*56 by Sylvain Lesne, Karen Ashe, and colleagues. The extent to which the mouse 12mer differs from the pathogenic human 12mer has not yet been determined.

    Isolation of neurologically active dimers from Alzheimer’s brain tissue reported at the SfN meeting by Ganesh Shankar, Dennis Selkoe, and colleagues, and summarized by Gabrielle for ARF in this news series, is the latest important addition to our understanding of the involvement of oligomers in dementia.

    For all the oligomers—whether dimers, 12mers, or other pathogenic species still to be characterized—we’ll need to learn how they form, why they accumulate, how they target particular neurons, and the extent to which they give a unifying mechanism that explains AD memory loss and brain pathology. These are still big questions for small toxins….

    As if that were not enough, one final point makes these questions even more important. We should remember that Aβ oligomers represent a whole new type of toxic structure derived from a fibrillogenic protein. Now a number of other fibrillogenic, disease-causing proteins have been found to generate sub-fibrillar, oligomeric toxins (see Gabrielle’s latest SfN story about α-synuclein and also Klein WL. [2006] Cytotoxic intermediates in the fibrillation pathway: Aβ oligomers in Alzheimer’s disease as a case study. In Protein Misfolding, Aggregation, and Conformational Diseases. Vol. 1. V. Uversky, ed. Kluwer Academic/Plenum Publishers, New York, NY). Even innocuous proteins can turn toxic when forced to form oligomers (Vieira et al., 2007). Toxic Aβ oligomers were just the first of many.


    . Amyloid plaque core protein in Alzheimer disease and Down syndrome. Proc Natl Acad Sci U S A. 1985 Jun;82(12):4245-9. PubMed.

    . Non-fibrillar beta-amyloid protein is associated with smooth muscle cells of vessel walls in Alzheimer disease. J Neuropathol Exp Neurol. 1994 Nov;53(6):637-45. PubMed.

    . Purification and characterization of brain clusterin. Biochem Biophys Res Commun. 1994 Nov 15;204(3):1131-6. PubMed.

    . Morphology and toxicity of Abeta-(1-42) dimer derived from neuritic and vascular amyloid deposits of Alzheimer's disease. J Biol Chem. 1996 Aug 23;271(34):20631-5. PubMed.

    . Diffusible, nonfibrillar ligands derived from Abeta1-42 are potent central nervous system neurotoxins. Proc Natl Acad Sci U S A. 1998 May 26;95(11):6448-53. PubMed.

    . Self-assembly of Abeta(1-42) into globular neurotoxins. Biochemistry. 2003 Nov 11;42(44):12749-60. PubMed.

    . Immunization reverses memory deficits without reducing brain Abeta burden in Alzheimer's disease model. Nat Neurosci. 2002 May;5(5):452-7. PubMed.

    . Reversible memory loss in a mouse transgenic model of Alzheimer's disease. J Neurosci. 2002 Aug 1;22(15):6331-5. PubMed.

    . Targeting amyloid-beta peptide (Abeta) oligomers by passive immunization with a conformation-selective monoclonal antibody improves learning and memory in Abeta precursor protein (APP) transgenic mice. J Biol Chem. 2006 Feb 17;281(7):4292-9. PubMed.

    . Self-assembly of Abeta(1-42) into globular neurotoxins. Biochemistry. 2003 Nov 11;42(44):12749-60. PubMed.

    . Prostaglandin H2 (PGH2) accelerates formation of amyloid beta1-42 oligomers. J Neurochem. 2002 Aug;82(4):1003-6. PubMed.

    . PGH2-derived levuglandin adducts increase the neurotoxicity of amyloid beta1-42. J Neurochem. 2006 Feb;96(4):917-23. PubMed.

    . Copper mediates dityrosine cross-linking of Alzheimer's amyloid-beta. Biochemistry. 2004 Jan 20;43(2):560-8. PubMed.

    . Synaptic targeting by Alzheimer's-related amyloid beta oligomers. J Neurosci. 2004 Nov 10;24(45):10191-200. PubMed.

    . Abeta oligomer-induced aberrations in synapse composition, shape, and density provide a molecular basis for loss of connectivity in Alzheimer's disease. J Neurosci. 2007 Jan 24;27(4):796-807. PubMed.

    . Alzheimer's disease-affected brain: presence of oligomeric A beta ligands (ADDLs) suggests a molecular basis for reversible memory loss. Proc Natl Acad Sci U S A. 2003 Sep 2;100(18):10417-22. PubMed.

    . Alzheimer's disease-type neuronal tau hyperphosphorylation induced by A beta oligomers. Neurobiol Aging. 2008 Sep;29(9):1334-47. PubMed.

    . Intracellular A beta 1-42 aggregates stimulate the accumulation of stable, insoluble amyloidogenic fragments of the amyloid precursor protein in transfected cells. J Biol Chem. 1995 Jun 16;270(24):14786-92. PubMed.

    . A specific amyloid-beta protein assembly in the brain impairs memory. Nature. 2006 Mar 16;440(7082):352-7. PubMed.

    . Soluble oligomers from a non-disease related protein mimic Abeta-induced tau hyperphosphorylation and neurodegeneration. J Neurochem. 2007 Oct;103(2):736-48. PubMed.

  8. We have read with great interest all the recent reports and comments on the toxicity of Aβ oligomers. We would like to start our own comment with a citation from Dr. Klein’s recent comment on this topic:

    “For all the oligomers—whether dimers, 12mers, or other pathogenic species still to be characterized—we’ll need to learn how they form, why they accumulate, how they target particular neurons,...”

    Our lab has been interested in characterizing axonal transport in neurodegenerative diseases, in particular in Alzheimer disease (AD). We wanted to ask whether the Aβ deposition in AD might result from a deficient axonal transport, a question that Dr. Larry Goldstein’s lab—and other labs as well—are also trying to answer. Experimentally, this question is difficult to address in animal models of AD, due to the difficulty of identifying early modifications in individual neurons. To circumvent this problem, we have employed a cell culture system, where CAD cells (a mouse neuronal cell line derived from the locus coeruleus) [1] produce and accumulate within their processes large amounts of Aβ, similar to what may occur in brain neurons, in the initial phases of AD [2]. Using this system, we showed that accumulation of Aβ likely begins within neurites, prior to any detectable signs of neurodegeneration or abnormal vesicular transport, and long before Aβ deposits are detected extracellularly. We found that neuritic accumulation of Aβ is restricted to a small population of neighboring cells that express normal levels of APP, but show redistribution of BACE1 to the neurites, where it colocalizes with Aβ and markers of late endosomes and autophagic vacuoles. Importantly, cells that accumulate Aβ appeared in isolated islets, indicating that Aβ accumulation is initiated in a small number of neurons, probably by intracellular determinants that alter APP metabolism and lead to Aβ aggregation.

    We found the fact that CAD cells occasionally produce and accumulate large amounts of Aβ remarkable, since these are cells that express normal levels of APP (they are not transfected, and only express endogenous APP). Previously, such neuritic accumulations have been mostly found in neuronal cultures derived from mice that largely overexpress mutated human APP (e.g., Tg2576 mice). We found it also very interesting that CAD cells that show neuritic Aβ accumulations may contain such high levels of Aβ that it is detectable by Western blots in whole cell lysates, without immunoprecipitation. Most importantly, this Aβ is oligomeric (we found dimers, trimers, as well as higher-number oligomers detectable with Dr. Charles Glabe’s anti-oligomer antibody, A11).

    We are now using the CAD cell system to investigate what leads to the formation of these accumulations of Aβ at the neurite terminals, and—most importantly—why these accumulations appear in clusters of cells. Do these cells originate from a few progenitors present in the culture? Are there intrinsic or extrinsic factors that determine this phenotype? Certainly, CAD cells appear to recapitulate some of the biochemical processes leading to Aβ aggregation, and may thus provide an experimental in vitro system for studying the molecular pathobiology of AD.


    . Characterization of a CNS cell line, CAD, in which morphological differentiation is initiated by serum deprivation. J Neurosci. 1997 Feb 15;17(4):1217-25. PubMed.

    . Neuritic deposits of amyloid-beta peptide in a subpopulation of central nervous system-derived neuronal cells. Mol Cell Biol. 2006 Jul;26(13):4982-97. PubMed.

Comments on Primary Papers for this Article

No Available Comments on Primary Papers for this Article


News Citations

  1. San Diego: Oligomers Live Up to Bad Reputation, Part 2
  2. San Diego: Oligomers Live Up to Bad Reputation, Part 3

Paper Citations

  1. . Naturally secreted oligomers of amyloid beta protein potently inhibit hippocampal long-term potentiation in vivo. Nature. 2002 Apr 4;416(6880):535-9. PubMed.
  2. . Diffusible, nonfibrillar ligands derived from Abeta1-42 are potent central nervous system neurotoxins. Proc Natl Acad Sci U S A. 1998 May 26;95(11):6448-53. PubMed.
  3. . A specific amyloid-beta protein assembly in the brain impairs memory. Nature. 2006 Mar 16;440(7082):352-7. PubMed.
  4. . A quantitative morphometric analysis of the neuronal and synaptic content of the frontal and temporal cortex in patients with Alzheimer's disease. J Neurol Sci. 1987 Apr;78(2):151-64. PubMed.
  5. . Synaptic alterations in CA1 in mild Alzheimer disease and mild cognitive impairment. Neurology. 2007 May 1;68(18):1501-8. PubMed.
  6. . Natural oligomers of the Alzheimer amyloid-beta protein induce reversible synapse loss by modulating an NMDA-type glutamate receptor-dependent signaling pathway. J Neurosci. 2007 Mar 14;27(11):2866-75. PubMed.

Further Reading


  1. . Detection of a biomarker for Alzheimer's disease from synthetic and clinical samples using a nanoscale optical biosensor. J Am Chem Soc. 2005 Feb 23;127(7):2264-71. PubMed.
  2. . Alzheimer's disease-affected brain: presence of oligomeric A beta ligands (ADDLs) suggests a molecular basis for reversible memory loss. Proc Natl Acad Sci U S A. 2003 Sep 2;100(18):10417-22. PubMed.


  1. San Diego: Oligomers Live Up to Bad Reputation, Part 2