5 February 2013. Like a blue moon, could oligomeric amyloid-β in the cerebrospinal fluid (CSF) be too rare to be informative? In a paper in the February 1 Alzheimer's & Dementia online, researchers led by Dennis Selkoe at Brigham and Women's Hospital, Boston, report on a new assay to measure these fleeting fragments. Their enzyme-linked immunosorbent assay (ELISA) appears highly specific and detects gobs of oligomeric Aβ in the brain—but none in the CSF. "This was a disappointment," Selkoe told Alzforum, though he believes the finding may make sense. "Because these peptides are highly hydrophobic and sticky, they are unlikely to remain long in aqueous solution, and it is possible they are just not in the CSF," he said.
Researchers were banking that Aβ oligomers, widely believed to be the most toxic entity in the AD brain, could form the basis for CSF-based diagnostic and prognostic tests. Is that hope now dashed? Scientists contacted by Alzforum were reluctant to throw the baby out with the cerebrospinal fluid. "I think the primary reason for lack of signal here is that the assay is too insensitive," said Henrik Zetterberg, University of Gothenburg, Sweden. Zetterberg predicts that an assay needs to detect as little as 1 picogram of oligomer per mL, or even less, to have a chance of success. The new ELISA has a lower detection limit of about 39 pg/mL, which the scientists can get down to about 6 pg/mL by concentrating the CSF. Selkoe agreed that a more sensitive assay might work, and his group continues to make improvements. He told Alzforum that, in unpublished work, they now have a lower limit of about 0.02 pg/mL and will rescreen CSF in the coming months. In the meantime, he believes that using these assays to measure oligomeric Aβ in brain tissue itself will be immensely valuable.
Oligomer-Specific ELISAs and the CSF
Selkoe and colleagues were concerned that previous assays developed to measure oligomeric Aβ were not specific enough. "Many labs have tried to develop assays for oligomers, but we felt that they rarely, if ever, validated them against natural human oligomers or oligomer standards," said Selkoe. To test the specificity of their ELISA, first author Ting Yang and colleagues used a modified peptide that readily forms dimers. A cysteine substitution for serine at position 26 facilitates disulfide cross-linking under oxidizing conditions (see ARF related news story). Yang found that a combination of an NAB61 antibody for capture and a biotinylated 3D6 antibody for detection gave the highest sensitivity for these synthetic dimers. Using 3D6 for both capture and detection yielded a less sensitive assay with a lower detection limit of 197 pg/mL. NAB61 recognizes a conformation found in oligomers (see Lee et al., 2006), while 3D6 recognizes the first five amino acids of Aβ and also binds fibrillar forms of the peptide.
The ELISAs seem highly specific for oligomers. Neither gave a signal with full-length APP, or with its secreted extracellular domains called sAPPα and sAPPβ. When the researchers broke the S26C dimers using a sulfhydryl reagent, the ELISA signal disappeared.
While other ELISAs that detect monomeric Aβ turned up strong signals from 90 human CSF samples, neither of the oligomer-specific ELISAs did, indicating they have little affinity for monomers. Even after immune-precipitating Aβ from CSF to concentrate it, the assays detected nothing. The authors concluded that either there are no oligomers in the CSF, or they are below the assay’s detection limit.
That limit may turn out to be the crux of the issue. Other researchers have used similar strategies to develop oligomeric Aβ assays. In the January Annals of Neurology, researchers led by David Brody at Washington University, St. Louis, Missouri, reported an immunoassay that detects Aβ oligomers at concentrations as low as 6.25 pg/mL dimer equivalents. Like Selkoe's group, first author Thomas Esparza and colleagues failed to detect any signals in human CSF. Brody also validated their assay with the same S26C dimers. Zetterberg told ARF that his group now has a very similar assay as well. It, too, is validated against Aβ oligomers and has an even lower detection threshold.
Perhaps the most sensitive assay reported so far comes from researchers at Merck. Mary Savage and colleagues reported at the 2012 Alzheimer’s Association International Conference in Vancouver, Canada, that they were able to detect as little at 0.42 pg/mL of oligomeric Aβ. That assay did find oligomers in human CSF. Savage reported that AD patients have more of them than do young controls (see ARF related news story).
"We have a nice story emerging here from multiple groups," said Brody. "I think it is possible that all of our assays are similar qualitatively, but not quantitatively, and that the levels of oligomer in the CSF are very low." Brody suggested that researchers in the field swap CSF samples so that groups can directly compare their different assays in the same samples.
Some researchers noted that Selkoe's collaborators previously reported Aβ dimers in human CSF (see ARF related news story), and wondered why they are not picking those up with the ELISA. In that earlier report, the researchers had immune-precipitated the CSF and concentrated everything on a Western blot using denaturing electrophoresis. "That means that every larger species of Aβ in the sample may have ended up as monomers and dimers," said Selkoe. "I think there may be some cases where oligomers are detectable," said Selkoe, "but those cases may be very few."
A provocative explanation might be that the dimers were artifacts. Recently, researchers led by Kevin Barnham at the University of Melbourne, Australia, reported that denaturing electrophoresis induces formation of Aβ dimers (see Watt et al., 2013). Brody agreed that artifacts are a potential problem. In their assays, the WashU researchers spiked samples with Aβ monomer prior to manipulating them, and then checked to see that they got all those monomers back after the procedure. "This is a really important control to run to make sure Aβ does not artifactually oligomerize during the preparations," he said.
While Selkoe was disappointed that oligomers were undetectable in the CSF, he sees a silver lining. He believes oligomer-specific ELISAs will help researchers to finally pin down and characterize these enigmatic entities, and also that the assays will be valuable in postmortem pathology studies. Yang found, for example, that brain extracts from AD cases contained up to 1,000-fold more Aβ oligomer than monomer. Furthermore, looking at soluble brain extracts from 13 cases and nine age-matched controls, Yang found oligomer levels, on average, 50-fold higher in the former. The data support the idea that soluble oligomers, unlike Aβ plaques, track with disease.
Brody's data fit with this. His group looked at AD cases and high-pathology controls, i.e., people who appear normal but have a lot of amyloid plaques. Reporting in the January Annals of Neurology, first author Thomas Esparza and colleagues found that the ratio of oligomeric Aβ to plaque fully distinguished who had had dementia and who had not. "No other marker I'm aware of has been able to fully distinguish demented patients from high-plaque controls," said Brody, "so that was very exciting."
Selkoe and Brody both see their data as lending credibility to the idea that Aβ plaques could be protective. "It seems that we can populate plaques for years prior to first symptoms, but then a steady-state level builds outside of plaques. Oligomers will then circulate, contact membranes, and cause problems. That's bad news," said Selkoe. Brody agreed. "Plaques could buffer oligomers. Figuring out that buffering capacity will be crucial to understanding and treating the disease," he said. Artificially boosting the buffering capacity, for example, could prove beneficial.
Immunotherapies directed toward oligomers would fit that bill, said Selkoe. But first, he thinks that oligomers need to be better characterized. These ELISAs provide tools to do that, he said. Researchers can use them to purify the oligomers, size them accurately, and figure out their structures and properties—goals that have been a royal headache (see ARF Webinar).—Tom Fagan.
Yang T, Hong S, O'Malley T, Sperling RA, Walsh DM, Selkoe DJ. New ELISAs with high specificity for soluble oligomers of amyloid β-protein detect natural Aβ oligomers in human brain but not CSF. Alzheimer's Dement. 2013 Jan 30. Abstract
Esparza TJ, Zhao H, Cirrito JR, Cairns NJ, Bateman RJ, Holtzman DM, Brody DL. Amyloid-beta oligomerization in Alzheimer dementia versus high-pathology controls. Ann Neurol. 2013 Jan;73(1):104-19. Abstract