Growing evidence suggests that neurons begin faltering decades in advance of cognitive symptoms in Alzheimer’s disease, but identifying the molecular culprits at play during this “invisible” phase has proven challenging. This is especially true of oligomeric forms of amyloid-β. Research reported in the March 11 JAMA Neurology proposes a step forward in the quest. Karen Ashe and colleagues at the University of Minnesota in Minneapolis, in collaboration with researchers at the University of Gothenburg in Sweden, have measured Aβ*56 in human cerebrospinal fluid (CSF). Aβ*56—convincingly to some, controversially to others, as much of the data on Aβ oligomers—is a species that reportedly wipes out memory in AD mice (see ARF related news story on Lesné et al., 2006).

In the new CSF analysis, Aβ*56 levels correlated with known CSF markers of neurodegeneration. This was the case only in cognitively normal seniors—not in people with AD or mild cognitive impairment (MCI). The findings imply that other factors elevate tau in symptomatic individuals. “This may help explain why anti-Aβ therapies have not been working in people who are symptomatic,” Ashe told Alzforum. Other scientists raised questions about the procedures used to measure the oligomers, which have been difficult to detect in human material, especially CSF.

In prior analyses with Tg2576 transgenic mice, an AD model generated in the Ashe lab, she and colleagues identified Aβ*56 as a putative 56 kDa dodecamer that correlates with cognitive impairment in this mouse strain. Aβ*56 shows up in their brains just as memory problems emerge, and levels of the oligomer track with the animal’s cognitive performance (ARF related news story on Lesné et al., 2006). Smaller Aβ species, including monomers, trimers, and hexamers, appear in Tg2576 mice prior to cognitive impairment. In this field, labs tend to stick to their own methods and protocols rather than adopt and compare each other’s for the purpose of multiple independent replication. This has so far prevented most reported oligomer species, including Aβ*56, from developing broad-based momentum.

The current paper, however, reports the first detection of Aβ*56 in human tissue. Lead author Maureen Handoko and colleagues report measuring this oligomer, as well as Aβ trimers, in the CSF of 48 cognitively impaired seniors (26 AD, 22 MCI), 49 age-matched controls with normal cognition, and 10 younger controls. The specimens came from a large, cross-sectional MCI study headed by coauthor Anders Wallin at the University of Gothenburg. One-fourth of each 1 ml CSF sample remained in Sweden, where researchers in the lab of coauthor Kaj Blennow measured CSF Aβ1-42, total tau, and phospho-tau 181 by using enzyme-linked immunosorbent assays (ELISA). The remaining 750 microliters went to the University of Minnesota, where Ashe’s group used immunoprecipitation (IP) and Western blotting to measure Aβ trimers and Aβ*56 in triplicate using 240-microliter aliquots. The Minnesota researchers were blinded to the participants’ clinical status—which the Gothenburg scientists revealed after the IP/Western data were quantitated and tabulated, Ashe said.

Among the cognitively normal volunteers, older individuals tended to have more Aβ trimers and Aβ*56 in their CSF. Moreover, Aβ*56 levels correlated with CSF tau and CSF phospho-tau levels, whereas Aβ1-42 did not. “This argues against fibrillar Aβ being coupled with tau, at least in the asymptomatic phase of disease,” Ashe said. About 10 percent of the cognitively normal older adults had an elevated CSF tau/Aβ1-42 ratio. The association was weaker in people with MCI or AD.

On a methodological level, the detection of Aβ oligomers in human CSF can be considered a feat in and of itself. Recent advances in sandwich ELISAs have enabled scientists to measure oligomers in human brain samples with high specificity, but these techniques did not detect oligomers in CSF (see ARF related news story). Ashe said her lab spent two years developing “a highly sensitive and specific immunoblot assay that can detect as little as 25-50 picograms of Aβ.” The Minneapolis group uses standard antibodies but has optimized transfer time, blocking conditions, and reagent concentrations. “All these things can drastically affect how clean your blots are,” Ashe said, noting that the detection limit for most Western blots is around 500 picograms—10 times less sensitive than her protocol. The detailed methods will appear in another manuscript, Ashe said.

The current study did not measure Aβ dimers, which some consider the most neurotoxic form of Aβ (see ARF related news story on Shankar et al., 2008; ARF news story on McDonald et al., 2010, and Villemagne et al., 2010). In unpublished work she is writing up for publication, Ashe found that Aβ dimers occur in human CSF at a ~10-fold lower concentration than Aβ trimers and Aβ*56; this ratio reverses in the brain. On her immunoblot, “you would need at least 1-2 milliliters of CSF to measure dimers,” Ashe said. Moreover, she said reliable visualization of Aβ dimers would require separate protein gels with different acrylamide concentration than those used in this study.

Some scientists were concerned about the biochemical evidence for Aβ oligomers in the present study. “It is not clear from the paper whether the detected species are Aβ rather than an APP fragment,” wrote David Brody of Washington University School of Medicine, St. Louis, Missouri, in an e-mail to Alzforum (see full comment below). He noted that levels of soluble APP in the CSF have been shown to be ~100-fold higher than Aβ (Nitsch et al., 1995). The paper reports the IP/Western blot data as densitometry light units; it does not show original blots. Some of the blots will appear in another paper currently in press in Brain.

The study does not control for the possibility that the IP/Western procedure itself could induce artifactual aggregation of Aβ, Brody noted. In a recent study (Esparza et al., 2013), he and colleagues found that monomeric Aβ can aggregate at high local concentrations, such as those occurring after immunoprecipitation. SDS, a detergent the researchers used to prepare samples for Westerns, can also induce Aβ aggregation (see Watt et al., 2013). Ashe acknowledges the potential for detergents such as SDS to trigger anomalous formation of oligomers. However, “if you do an immunoprecipitation from CSF, that eliminates this possibility because there is no detergent in CSF,” she said. On the concern about mistakenly detecting APP fragments, she said her lab has several lines of unpublished evidence indicating that they are, in fact, measuring Aβ. For example, when the researchers ran undiluted CSF on a size exclusion chromatography (SEC) column, Aβ*56 and Aβ monomers showed up in separate fractions on SDS-PAGE. Furthermore, they see no signal on Western blots probed with APP antibodies that recognize epitopes outside the Aβ region. “I’m very convinced that what we’re looking at is Aβ,” Ashe said.—Esther Landhuis


  1. The topic is important, and the results are intriguing.

    The issue of Aβ oligomerization is a complex one, and several methodological questions could be raised:

    1. The immunoprecipitation with 6E10 and Western blotting with 6E10 may not distinguish between Aβ and soluble APP fragments. It is not clear from the paper whether the detected species are Aβ rather than an APP fragment. Of note, the levels of soluble APP in the CSF are about 100-fold higher than Aβ (see Nitsch et al., 1995, Table 2), so even a minor APP fragment would give a lot of signal in the 6E10-based assay.

    2. It will be important to control for the possibility that the immunoprecipitation and Western blotting assay procedures themselves induce artifactual aggregation of Aβ (see Esparza et al., 2013, Fig. 2 L). Monomeric Aβ can aggregate at high local concentrations, such as those that occur after immunoprecipitation.

    3. Additional controls of interest regarding the assay include test-retest reproducibility, dilutional linearity, and spike-recovery linearity.

    4. A future direction could involve quantification of the amount of protein corresponding to the signals detected on Western blot relative to known biochemical standards.


    . Cerebrospinal fluid levels of amyloid beta-protein in Alzheimer's disease: inverse correlation with severity of dementia and effect of apolipoprotein E genotype. Ann Neurol. 1995 Apr;37(4):512-8. PubMed.

    . Amyloid-β oligomerization in Alzheimer dementia versus high-pathology controls. Ann Neurol. 2013 Jan;73(1):104-19. PubMed.

    View all comments by David Brody
  2. The measurements of CSF oligomeric Aβ were done by immunoprecipitation/Western blot using as little as 240 μL per determination (ran as triplicate, then averaged). Using such a low volume suggests it may be possible to integrate similar measurements in longitudinal studies. It is disappointing that CSF levels of Aβ dimers could not be determined due to the experimental design (the acrylamide content in the gel cannot resolve small species). Concentrations of Aβ1-42 and tau/pt181-tau were determined by ELISA, as it is traditionally done for biomarker studies.

    Both oligomeric Aβ species (Aβ*56 and Aβ trimers) detected in the CSF correlated with tau/ptau concentrations in aged, unimpaired subjects, while presumably monomeric Aβ1-42 did not. By extrapolation, it could indicate that the elevation of trimer-based Aβ oligomers seen in aging and AD is linked to abnormal tau changes. This interpretation is consistent with the notion that Aβ*56 and Aβ trimers may initiate the disease process during the latent phase of AD (i.e., preclinical AD).

    In impaired individuals (including MCI and AD groups), CSF Aβ trimers and Aβ1-42 correlated to tau/ptau concentrations (in opposite fashion). These observations could indicate that: 1) soluble Aβ trimers might still drive abnormal tau changes in the AD brain, but their effects are reduced (based on comparing correlation coefficients); 2) Aβ*56 might not be as "active" a toxin in AD as in preclinical AD; and 3) tau changes become less dependent on oligomeric Aβ as disease progresses (none of these scenarios are mutually exclusive). Whether these scenarios would apply to other oligomeric forms of Aβ remains to be examined.

    Overall, the data presented are consistent with the hypothesis that Aβ*56 and Aβ trimers (and presumably trimer-based oligomers) play important roles in the prodromal stages of the disease. These findings are also in agreement with alterations in CSF Aβ levels occurring early in asymptomatic individuals.

    View all comments by Sylvain Lesne
  3. Development of valid and quantitative assays for oligomeric Aβ species is considered by many to be the Holy Grail in the AD fluid biomarker field. Such assays are technologically challenging for many reasons. Several groups have reported such assays, but establishing an assay’s validity has been problematic, and none has stood the test of time. Dr. Ashe’s group has been interested in oligomeric Aβ species for several years, reporting in 2006 the presence of the Aβ*56 form in Tg2576 mice and its memory-disrupting ability when injected into rats. In this current paper, they report the presence of Aβ trimers as well as the Aβ*56 species in human CSF samples using a combined immunoprecipitation and immunoblotting procedure. Although the assay appears not to be quantitative (instead, semi-quantitative), the reported percent coefficient of variation among triplicates was good (Assuming the assay is indeed valid for detecting these specific oligomeric species, the data are very interesting and support several current hypotheses regarding Aβ metabolism in AD. The observation that these species were elevated in cognitively normal individuals at risk for developing AD dementia (defined as those with a high tau/Aβ42 ratio) suggests that they are involved in the very early (preclinical) stages of the disease. It would have been nice to see the relationship between monomeric Aβ42 (the proposed analyte quantified in the INNOTEST ELISA assays) and the various oligomeric species, as well as the distribution of levels of these species in the various clinical groups. It would also be interesting to know whether levels were elevated in MCI (or control) individuals who later progressed to AD, but were not elevated in MCI (or control) individuals who remained stable (hypothesized not to have underlying AD). The analysis of individuals as a function of the tau/Aβ42 ratio at baseline suggests that this would be the case, and I expect future studies that include longitudinal clinical follow-up in this cohort will be able to shed light on this issue.

    The relationship between levels of Aβ trimers/Aβ*56 and tau/ptau in cognitively normal individuals and its attenuation in symptomatic cases is arguably the most interesting finding of this study. Such a relationship is not observed between levels of monomeric Aβ42 and tau, suggesting a potential unique role of oligomeric species in the pathologic cascade at the earliest stages of the disease. While interesting in its own right to those striving to understand the normal evolution of AD pathobiology, this finding also has potentially important implications for the design and evaluation of therapies targeting Aβ and/or amyloid. Hopefully, the assay can be developed into a truly quantitative platform with higher throughput that will permit the evaluation of large numbers of samples from individuals who have been well characterized clinically over time and with multiple biomarker assessments. Then, perhaps, we can add a line to the left of the Aβ42/amyloid line in the hypothesized biomarker trajectory schematic.

    View all comments by Anne Fagan
  4. It is our hope that the publication of our JAMA Neurology paper and the accompanying Alzforum story will motivate other laboratories to study Aβ*56. We certainly recognize that the existence of this species as an authentic oligomer that occurs in vivo is controversial. Perhaps, though, the following considerations will encourage skeptics and believers alike to take a closer look at Aβ*56.

    The existence of specific Aβ oligomers as real entities, rather than artifacts, has been questioned because of the possibility that they are artificially generated through exposure to detergents, such as SDS. Several lines of evidence argue against this possibility.

    1. When proteins in undiluted CSF are first separated by size-exclusion chromatography (SEC) and then analyzed by Western blot, Aβ*56 and Aβ monomers are seen in separate fractions eluted from the SEC column. If Aβ*56 was artifactually generated from monomers during the process of gel electrophoresis, one would expect to see both of these species in the same fractions from the SEC column.

    2. Using the same extraction and detection protocols to measure the oligomers, we have observed that different mouse lines overexpressing APP consistently display line-specific sets of oligomeric Aβ species.

    3. The Aβ oligomers that are particular to a given mouse line accumulate with age in an orderly fashion (Lesné et al., 2006; Larson and Lesné, 2012).

    4. Aβ dimers and Aβ*56 in micro-dissected tissue are differentially distributed (Liu et al., 2011a).

    5. Finally, perhaps the most compelling evidence in favor of the existence of Aβ*56 is the strong correlation between levels of this oligomer and markers of compromised neuronal function in both brain and CSF, and in humans as well as APP transgenic mice.

    Levels of brain Aβ*56 correlate with cognitive impairment in multiple lines of transgenic mice (Lesné et al., 2006; Cheng et al., 2007); a transient (~three-week) dip in the levels of this oligomer during the period of the most rapid plaque deposition in Tg2576 mice is accompanied by a temporary recovery of cognitive function (Lesné et al., 2008) . In human subjects who were cognitively normal at the time of death, Aβ*56, but not other Aβ oligomers, correlated negatively with the postsynaptic markers drebrin and Fyn kinase, and positively with pathological conformers of tau (Lesné et al., in press). In the CSF of clinically unimpaired subjects, Aβ*56 correlated strongly with levels of total tau and tau phosphorylated at threonine 181—putative markers of neuronal injury (Handoko et al., 2013). It seems to us very unlikely that an entity generated artificially during extraction or blotting would consistently correlate with other indicators of an unhealthy brain.

    We would expect an Aβ species that initiates the amyloid cascade to stimulate downstream processes (network dysfunction, generation of toxic tau species, neuroinflammation, aberrant cell-cycle re-entry?) that eventually lead to neuron death. We find it difficult to reconcile the slow progression of AD with acute Aβ-induced toxicity in cell culture models. We would argue that such acute toxicity is an experimental artifact, caused by biologically irrelevant (perhaps artificially generated?) species of Aβ, unrealistically high concentrations of Aβ, spatial or temporal patterns of exposure to Aβ that are not reflective of those that occur in situ, or elevated susceptibility to Aβ toxicity. Consistent with what we would expect of an Aβ species that triggers the amyloid cascade, Aβ*56 does not kill cells—APP transgenic mice that generate Aβ*56 do not exhibit widespread neurodegeneration, and exogenous administration of Aβ*56 to healthy host animals results in reversible memory dysfunction. These observations suggest that Aβ*56 interferes with synaptic function or plasticity; elucidating the mechanisms of action of Aβ*56 is a major focus of our laboratory.

    We fully acknowledge that detecting Aβ*56 on blots can be difficult. Many other proteins run at ~55 kDa on SDS-PAGE, including the IgG heavy chain. It is critical to immunodeplete samples of endogenous mouse IgG when using mouse anti-Aβ antibodies followed by secondary anti-mouse antibodies for detection. Blocking of blots and protein extraction methods are also key factors. We (Liu et al., 2011b) and others (Sherman and Lesné, 2011) have published detailed methods that should ensure success, if followed.

    Dr. Ashe has already alluded to the approaches used to determine whether Aβ*56 is an oligomer—some of these have been published (Lesné et al., 2006) and others will be found in a paper currently in press in Brain (Lesné et al., in press). We continue to refer to Aβ*56 as a “putative dodecamer,” recognizing that its identity is not yet firmly established. Whatever this band represents, it correlates with cognitive deficits in multiple APP transgenic mouse lines, and now with markers of disease/synaptic dysfunction in humans. We strongly believe that these findings make Aβ*56 worthy of further study.

    To really test the hypothesis that Aβ*56 is necessary to trigger the amyloid cascade, we must determine whether selectively decreasing the levels of Aβ*56 or interfering with its interactions with its cellular targets reduces the risk of symptomatic AD. Such studies await the identification of the targets of Aβ*56 and/or development of reagents that selectively target specific Aβ oligomers.


    . Accelerating amyloid-beta fibrillization reduces oligomer levels and functional deficits in Alzheimer disease mouse models. J Biol Chem. 2007 Aug 17;282(33):23818-28. PubMed.

    . Correlation of Specific Amyloid-β Oligomers With Tau in Cerebrospinal Fluid From Cognitively Normal Older Adults. JAMA Neurol. 2013 May 1;70(5):594-9. PubMed.

    . Soluble Aβ oligomer production and toxicity. J Neurochem. 2012 Jan;120 Suppl 1:125-39. PubMed.

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

    . Plaque-bearing mice with reduced levels of oligomeric amyloid-beta assemblies have intact memory function. Neuroscience. 2008 Feb 6;151(3):745-9. PubMed.

    . Brain amyloid-β oligomers in ageing and Alzheimer's disease. Brain. 2013 May;136(Pt 5):1383-98. PubMed. Correction.

    . Aβ dimers mediate plaque-associated cytopathology without affecting cognition. Alzheimers Dement. 2011 Jul; 7(4 Suppl):e23.

    . Grape seed polyphenolic extract specifically decreases aβ*56 in the brains of Tg2576 mice. J Alzheimers Dis. 2011;26(4):657-66. PubMed.

    . Detecting aβ*56 oligomers in brain tissues. Methods Mol Biol. 2011;670:45-56. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Aβ Star is Born? Memory Loss in APP Mice Blamed on Oligomer
  2. New Assays for Aβ Oligomers—Spinal Fluid a Miss, Brain Awash
  3. Paper Alert: Patient Aβ Dimers Impair Plasticity, Memory
  4. Bad Guys—Aβ Oligomers Live Up to Reputation in Human Studies

Paper Citations

  1. . A specific amyloid-beta protein assembly in the brain impairs memory. Nature. 2006 Mar 16;440(7082):352-7. PubMed.
  2. . Amyloid-beta protein dimers isolated directly from Alzheimer's brains impair synaptic plasticity and memory. Nat Med. 2008 Aug;14(8):837-42. PubMed.
  3. . The presence of sodium dodecyl sulphate-stable Abeta dimers is strongly associated with Alzheimer-type dementia. Brain. 2010 May;133(Pt 5):1328-41. PubMed.
  4. . Blood-borne amyloid-beta dimer correlates with clinical markers of Alzheimer's disease. J Neurosci. 2010 May 5;30(18):6315-22. PubMed.
  5. . Cerebrospinal fluid levels of amyloid beta-protein in Alzheimer's disease: inverse correlation with severity of dementia and effect of apolipoprotein E genotype. Ann Neurol. 1995 Apr;37(4):512-8. PubMed.
  6. . Amyloid-β oligomerization in Alzheimer dementia versus high-pathology controls. Ann Neurol. 2013 Jan;73(1):104-19. PubMed.
  7. . Oligomers, fact or artefact? SDS-PAGE induces dimerization of β-amyloid in human brain samples. Acta Neuropathol. 2013 Apr;125(4):549-64. PubMed.

Other Citations

  1. Tg2576 transgenic mice

External Citations

  1. MCI study

Further Reading


  1. . A specific amyloid-beta protein assembly in the brain impairs memory. Nature. 2006 Mar 16;440(7082):352-7. PubMed.
  2. . New ELISAs with high specificity for soluble oligomers of amyloid β-protein detect natural Aβ oligomers in human brain but not CSF. Alzheimers Dement. 2013 Mar;9(2):99-112. PubMed.
  3. . Amyloid-β oligomerization in Alzheimer dementia versus high-pathology controls. Ann Neurol. 2013 Jan;73(1):104-19. PubMed.
  4. . Soluble Aβ oligomer production and toxicity. J Neurochem. 2012 Jan;120 Suppl 1:125-39. PubMed.

Primary Papers

  1. . Correlation of Specific Amyloid-β Oligomers With Tau in Cerebrospinal Fluid From Cognitively Normal Older Adults. JAMA Neurol. 2013 May 1;70(5):594-9. PubMed.