23 July 2012. The going hypothesis on amyloid-β forms states that oligomers are the most toxic species, and yet no one can robustly measure them in human CSF, where oligomers are rare clumps floating in a sea of monomer abundance. If researchers could quantify the oligomer population in a person, then they might get a solid clue as to the person’s Alzheimer’s or cognitive impairment status, as well as a measure of whether anti-amyloid therapies are doing any good. At the Alzheimer’s Association International Conference, held 14-19 July 2012 in Vancouver, Canada, two companies presented their progress toward oligomer-measuring assays. Merck’s West Point branch in Pennsylvania and the biotech company Amorfix, of Mississauga, Ontario, are taking distinct approaches to pick up femtogram amounts of oligomers. Both are observing similar trends in human cerebrospinal fluid samples. The assays are in early stages, and the designers are planning more work to validate them.
The field has been chasing the Aβ oligomer concensus for several years. In fact, a consensus on oligomers has been so elusive that it has been derided as the emperor without clothes (Benilova et al., 2012). One approach is to perform an enzyme-linked immunosorbent assay (ELISA) using the same Aβ antibody as both the capture and detection reagent. If the antibody can only bind a monomer at one site, then any conglomerate detected in this manner must be a dimer or more (see ARF related news story on Xia et al., 2009). For example, IBL International, based in Hamburg, Germany, sells such a single-antibody kit. However, this approach is not very sensitive, commented Kevin Felsenstein of the University of Florida, Gainesville, who was not involved in the current AAIC presentations but has collaborated with Amorfix in the past. IBL’s kit detects oligomers in the picomolar range, but CSF samples—particularly those from healthy people—may contain far less than that. With CSF containing 1,000-fold more monomer than oligomer, noted William Goure of Acumen Pharmaceuticals in Livermore, California, oligomer detection must be exquisitely sensitive and selective.
Merck’s approach, presented on a poster by Mary Savage, is to apply an ELISA-like technology using an antibody specific for oligomeric Aβ. The company’s 19.3 antibody, developed in collaboration with Acumen, preferably binds to synthetic oligomers over monomers. It was originally made in mice, but the researchers have modified its sequence to make it more human-like. Its epitope is unknown, but Savage suspects it binds to a three-dimensional conformation only present in oligomers. No other humanized amyloid antibody bests 19.3 for oligomer selectivity, Goure said.
Savage and colleagues first tried using 19.3 to capture oligomers in a typical ELISA platform, with a pan-amyloid antibody called 82E1 as the detecting antibody. That worked, but could only pick up oligomers reliably at 4.2 picograms/milliliter or more—fine for brain tissue but not good enough for CSF samples. The researchers turned to a more sensitive assay developed by Singulex of Alameda, California. Called Erenna, this ELISA sandwiches the antigen and antibodies between a magnetic bead and a fluorescent tag, then quantifies the fluorescence pulled down by the magnetic beads. With this ELISA, the team was confident they could detect as little as 420 femtograms of oligomers per milliliter of CSF. The assay’s sensitivity exceeds that of previously reported CSF oligomer assays, the scientists said.
In preliminary tests with a handful of purchased CSF samples, Savage was excited to see that people with AD had more oligomers than did young control donors. A further 72 samples confirmed the assay was able distinguish the two populations, with less overlap between the groups than the field is used to seeing with CSF monomer ELISAs. “You can start to draw a line” differentiating the populations, Savage told Alzforum. “I think this might have some utility for diagnosis.” In collaboration with groups in Norway and in the U.K., Savage will next test the assay in longitudinal cohorts that also collect other fluid, imaging, and cognitive markers.
However, there is still fine-tuning to be done. Many control samples had fewer oligomers than the researchers could confidently measure. Savage hopes to improve the ELISA by treating the CSF samples with a dash of detergent; without it, oligomers stick to the walls of the tubes in which they are stored (Pica-Mendez et al., 2010). In addition, the oligomer standards she used likely contained some monomer, so purer standards might improve the assay, she said (Hepler et al., 2006).
The assay is not currently available to researchers outside Merck. Merck does not typically produce diagnostics, and this assay was originally intended as an internal companion biomarker for a therapeutic antibody development program that has since been discontinued. Savage said that if further validation proves the assay’s worth, she would like to find a way for others to use it.
Amorfix took a different tack, measuring oligomers indirectly, as presented in a July 16 talk by Marni Uger. They have already commercialized a size-filtering method to quantify mouse amyloid oligomers, which is available from JSW Life Sciences. To this technique, they added a monomer-masking method developed in the laboratory of Amorfix founder Neil Cashman of the University of British Columbia in Vancouver. The researchers treat CSF samples with a molecule that modifies epitopes exposed in monomers, preventing detection by their amyloid antibody. The idea is that only epitopes buried inside of oligomers escape the modification, while monomers are tagged and become invisible to the antibody. Then, the researchers dissociate the oligomers and measure unmodified monomer concentration with an ELISA method that, like Merck’s, measures the fluorescence signal from antigens pulled down by magnetic beads.
Amorfix claims it can detect as little as 44 femtograms of oligomeric amyloid-β. In nearly 200 purchased CSF samples, the assay distinguished people with mild cognitive impairment (MCI) from those with AD, and each disease state from age-matched healthy control samples. MCI samples reportedly had the highest oligomer signal, with AD falling between MCI and control levels. “The strength of our assay is the ability to distinguish the MCI individuals,” Uger said, since those are the people who might benefit most from early treatment. Cashman and Uger speculate that oligomers might be most prevalent in the early stage of neurodegeneration because they are produced by neurons under attack from the disease.
However, biphasic assays can be difficult to interpret. For example, a mid-range oligomer value alone might indicate a person is either in the early stages of MCI or on the way to AD dementia. Clinicians would have to use the oligomer measurements in conjunction with clinical symptoms to assign the right stage of disease, Cashman said.
Taken together, preliminary data from the two teams indicate that people with AD had approximately two to five times the amount of CSF oligomer as controls; this stands in contrast with monomer Aβ, which falls as one approaches AD. However, the new techniques require extensive validation. Meanwhile, Felsenstein sounded a note of caution about oligomer measurements. “It may be artifact,” he said. One problem is that storing CSF might alter the proportion of monomers and oligomers; another is that detergents and salts added in processing can skew oligomer counts. The best approach, Felsenstein suggested, is to assay fresh CSF and limit manipulation to simple dilution of the sample.—Amber Dance.