Oligomer-only Alzheimer Disease?
19 November 2007. Of all other presentations on Aβ oligomers, a talk on Wednesday morning by Takami Tomiyama, who works with Hiroshi Mori at Osaka City University, Japan, created the greatest stir in the audience. Tomiyama told of a new variant of AD that appears to be due predominantly to oligomers. As such, it would offer the first direct evidence in humans that Aβ oligomers are sufficient to bring on AD symptoms in the absence of fibrils. Different aggregation forms of Aβ coexist in the human AD brain and until now, scientists have been unable to tease apart their relative contributions, except as reported above. That’s because methods to identify one over the other in live people, for example, oligomer-specific brain imaging ligands, do not exist as yet. Tomiyama presented a possible genetic separation of clinical AD from plaque formation. He told the audience that their clinic had examined a 59-year-old Japanese woman with a pedigree of early-onset familial AD. She had a classic clinical presentation. The scientists found a novel mutation within the Aβ sequence of her APP gene. The patient and her symptomatic sister were homozygous for this mutation, while other, unaffected siblings were heterozygous for it, raising the possibility that the mode of inheritance might be recessive. This would be highly unusual, as pathogenic APP mutations are typically autosomal-dominant. The scientists next screened a sample of more than 5,300 Japanese people gathered by the Japanese Genetic Study Consortium for Alzheimer's Disease for the new mutation, and they found three additional families carrying it.

When Tomiyama and colleagues next characterized the mutation in terms of Aβ production and aggregation, they came in for a surprise. Not only did the mutation reduce total Aβ secretion without affecting the Aβ42/40 ratio, but the mutant peptide also failed to aggregate into fibrils. Thioflavin T fluorescence and electron microscopy suggested as much, and on subsequent Western blotting, too, the mutant Aβ peptide displayed a unique aggregation behavior. It formed oligomers more abundantly than do wild-type peptides but did not form fibrils; instead, it seemed to get firmly stuck in the oligomer conformation. Finally, the scientists assessed how the mutant Aβ would affect synaptic plasticity. They injected it into rat lateral ventricles and soon after stimulated the Shaffer collateral pathway. When recording EPSPs from the hippocampus, the scientists noted that wild-type Aβ slightly reduced LTP in this experiment, as others have found, but the new Aβ mutation did so more potently. In total, these families offer a human genetic validation for the hypothesis that synaptic and cognitive impairment in AD may be more closely tied to oligomers than fibrils and plaques, Tomiyama said.

Several scientists in this session told this reporter that this was the most exciting talk they had heard at the conference so far, and that they were eager to see the paper published. They noted that it will be interesting to follow the heterozygous relatives, one of whom reportedly has MCI, to see if they, too, in time develop AD. Independent investigators in the field said that this work may have profound implications for the underlying biology of AD. One question the data reported at the conference raised in their minds is what kind of pathology this mutation might cause in the brains of affected patients. No tissue is available for autopsy to date, but Tomiyama said in the question-and-answer period following his talk that a PIB PET scan in the proband was negative (PIB binds primarily fibrillar plaques). Other investigators emphasized that it will be important to perform careful imaging studies on these families to ascertain that their disease truly is AD. In a conversation after Tomiyama’s talk, Mori agreed with this point. He noted that the patients were diagnosed using DSM-IIIR and NINCDS-ADRDA criteria, and said that though they were clinically typical for AD, they indeed may not have AD as traditionally defined. “To have AD, do you have to have senile plaques? This is exactly my question. Even once we have autopsy material and do not find plaques, we cannot answer that question,” he said. Mori emphasized, however, that the disease represents a rare variant of AD. To his mind, it presents an opportunity to test the hypothesis that AD is a disease of synaptic failure.

In the same session, Anna Lord, in Lars Lannfelt’s group at Uppsala University in Stockholm, Sweden, presented new data on that group’s exploration of the Arctic mutation in APP, which, like the still-undisclosed new Japanese mutation, falls into the Aβ sequence of APP. Discovered by the Swedish group in 2001, the Arctic mutation has done much to separate effects of oligomeric Aβ from those of fibrillar forms, and in San Diego Lord continued the story. Previously, Lord and colleagues had characterized a transgenic mouse line made to express this mutation plus the Swedish APP mutation. They had found that pathology in these mice began with early intraneuronal Aβ accumulation and then continued with plaque pathology later (Lord et al., 2005).

At the time, the scientists had no means to measure Aβ oligomers in vivo. Now they do. With Hillevi Englund, Dag Sehlin, and Frida Ekholm Pettersson in the same group, the Swedish scientists used their monoclonal IgG mAb158 to develop a sandwich ELISA. This new assay is specific for Aβ oligomers/protofibrils, and it detects large oligomers down to a concentration of 1 picomolar, about 5,000-fold better than it detects low-molecular-weight Aβ (monomers to tetramers). In a poster presented at the AD/PD conference last March in Salzburg, Austria, Pettersson showed that unlike 6E10, a widely used anti-Aβ antibody in the field, Ab158 does not bind APP in its native state. In the hands of this group, 6E10 also did not distinguish between low-n oligomers or monomer and larger protofibrils, see also Englund et al., 2007. (First, a word about nomenclature: these Swedish investigators refer to Aβ oligomers as “protofibrils,” partly because they follow an early naming by David Teplow, and partly because they focus on species larger than dimers or trimers. The protofibrils in their hands are above 60 kDa in molecular weight but fall apart upon SDS treatment, as did the ones found by Shankar and colleagues. Theoretically, the correlates to the Swedes’ “protofibrils” in the human AD extracts would have eluted first, in the solvent of the size exclusion column used by the Boston/Dublin team.)

In San Diego, Lord reported that the mAb158 sandwich ELISA detects synthetic protofibrils, as well as protofibrils in biological samples from a variety of mouse models. The scientists detected and quantified protofibril levels in their Arctic mouse, in the APP+PS1 model, and in the Tg2576 model. In all of these mouse strains, protofibrils predated the appearance of plaques. The Arctic strain contained by far the highest levels, and protofibril levels correlated with Aβ burden but not with amyloid. This finding led some scientists attending the session to speculate that the Japanese Aβ mutation eventually upon autopsy might yield a similar picture of diffuse deposits but no neuritic plaques.

Lord also presented initial behavioral data on the Arctic mice, reporting that the mice did poorly in finding the platform in the Morris water maze at 4 to 8 months of age, prior to plaque deposition. She also reported that levels of the protofibrils rise in the mice between 8 and 14 months of age but then plateau, whereas formic-acid extracted insoluble Aβ (the stuff in neuritic plaques) keeps increasing steeply with age as the mice continue to deposit more plaques. (This last finding jibes with findings from Karen Ashe’s group, who found the Aβ*56 species by looking for forms of Aβ that stay stable over certain periods of age.) In summary, Lord said that oligomers/protofibrils represent an early, neurotoxic stage of amyloid pathology and could serve as a marker of cognitive impairment. (Lord disclosed an ownership interest in BioArctic Neuroscience, a company co-founded by Lannfelt that attempts to develop immunotherapies for AD.)

Adding to the field’s toolbox, another group also reported their generation of new sandwich ELISAs selective for Aβ oligomers. This reporter missed this presentation, but according to the abstract (search SfN abstracts for No. 444.13), these assays use the oligomer-selective Nab61 monoclonal antibody (Lee et al., 2006) , as well as others. M. Kim, Virginia Lee, and colleagues at the University of Pennsylvania in Philadelphia write in the abstract that it can measure oligomers in human brain extracts and other biological fluids, as well as the response to anti-amyloid treatment in a mouse model of amyloidosis. Along the same line of anti-amyloid treatment, Igor Klyubin at Trinity College, Dublin, and collaborators there and at four other institutions, presented a poster (Abstract 485.5) also suggesting that systemic passive immunization might work. These investigators intravenously injected anti-Aβ antibodies into rats. They found that this rescued not only the inhibition of LTP that these authors had previously reported with cell-secreted Aβ oligomers (Walsh et al., 2002), but also a similar LTP inhibition that was caused by Aβ oligomers obtained from CSF of patients with AD.—Gabrielle Strobel.

This is Part 2 of a three-part summary of presentations about Aβ oligomers at the Society for Neuroscience conference held 3-7 November 2007. See Parts 1 and 3.