7 May 2010. As the field tries to pin down the molecular culprit behind Alzheimer disease, the finger pointing has shifted in recent years from amyloid plaques to its building blocks—free-floating Aβ peptides. Now, two small human studies solidify this move by correlating levels of soluble Aβ, particularly dimers, directly with dementia. Reporting in this month’s issue of Brain, researchers led by Dominic Walsh of University College Dublin, Ireland, use biochemical methods to measure Aβ load in the brains of recently deceased elderly. They show that soluble forms of Aβ, but not plaques, associate with AD dementia. Using mass spectrometry to analyze blood samples in living people, Kevin Barnham, University of Melbourne, and colleagues report in this week’s Journal of Neuroscience that AD patients have elevated levels of Aβ monomers and dimers, and that these measures track with cognitive decline and brain atrophy.
For much of the last few decades, Aβ plaques have been the big guns in AD research—the prime pathological hallmark, a yardstick for measuring efficacy of AD drug candidates, the stuff detected by in vivo amyloid tracers. But then came a trickle of data—which now gushes forth—showing that a fair number of seniors have plaque-laden brains yet seem cognitively fine (e.g., Aizenstein et al., 2008; see ARF related news story). These data again threw into question the centrality of plaques and shifted attention toward soluble forms of Aβ as being the true troublemakers in AD. The new premise finds support in recent work by Dennis Selkoe’s lab at Brigham and Women’s Hospital, Boston. He and colleagues isolated Aβ from postmortem AD brains and confirmed that the peptides—especially dimers—are, in fact, toxic (Shankar et al., 2008 and ARF related news story), presumably by disrupting glutamate uptake at synapses (Li et al., 2009 and ARF related news story). Those studies, and many others that examined the neurotoxicity of Aβ oligomers, were done in rodent hippocampal slices.
The present studies correlate soluble Aβ with cognition in people. As reported in Brain, first author Jessica McDonald and colleagues measured soluble and insoluble Aβ in postmortem brain tissue from 43 seniors in a UK population-based longitudinal study, and determined whether Aβ load correlated with dementia status at the time of death. From the brain samples, the researchers prepared buffer-soluble, detergent-soluble, and formic-soluble extracts, immunoprecipitated them with a polyclonal anti-Aβ antibody (AW8), and analyzed the captured material by Western blot using C-terminal monoclonal Aβ antibodies (2G3 and 21F12). The scientists used N-terminal Aβ antibodies (3D6 and 6E10) to confirm that the bands appearing on the blot as monomer (4.1 kDa), dimer (7.5 kDa)—and in a few cases, trimer (12.1 kDa) and tetramer (17.8 kDa)—were, in fact, full-length Aβ species.
Unlike other studies that have used ELISA to measure Aβ from human samples (Wang et al., 1999; Lue et al., 1999), “we use a method where we capture things in their native state,” Walsh told ARF. “The polyclonal antibody recognizes multiple different assemblies of Aβ, which we then denature so we can see the total Aβ present in the material being captured.” Most ELISA methods only detect monomers, Walsh said. Selkoe, who is a coauthor on the Brain paper, and colleagues recently developed an ELISA for low-molecular-weight Aβ oligomers, but by the time the assay was published (Xia et al., 2009 and ARF related news story), the UK brain study was already underway. Furthermore, the immunoprecipitation/Western method is more sensitive because it allows researchers to probe samples with volumes two to four times greater than what fits onto an ELISA plate, Walsh said.
Another distinguishing feature of the current study is its patient population. Whereas prior studies of soluble Aβ and dementia have restricted their analysis to tissue from nursing home residents (e.g., Näslund et al., 2000 and ARF related news story) or clinically predefined patient groups, “we went after a population base,” Walsh said. “Though relatively small, this study has the type of distribution of pathology and cognitive impairment that you see in a whole population. We didn't know anything about the participants’ clinical status or pathology until we completed the study.”
After doing the extractions and IP/Westerns, the researchers grouped their 43 samples into three groups: non-demented (n = 19), demented without AD pathology (n = 10), demented with AD pathology (n = 14). The AD-type dementia samples were readily distinguished from the first two groups based on Aβ concentrations (both monomer and dimer) in the buffer-soluble and detergent-soluble extracts, most clearly with the SDS-soluble fractions. Dimers were better than monomers at differentiating the patient sets.
When demented and non-demented patients were further stratified based on the extent of AD pathology (diffuse plaques, neuritic plaques, or tangles rated “none,” “low,” “intermediate,” or “high”), the researchers found that the demented group had more soluble Aβ than cognitively intact people with severe pathology.
This suggests there may be further nuances within existing categories of soluble and insoluble Aβ. It may be that “if you have certain types of plaques—i.e., things that are free-floating and in equilibrium with the soluble phase—then you’re demented. But if you have plaques that are essentially a tombstone—i.e., they’re a dead end, and don’t exchange with the soluble phase—then you’re not demented,” Walsh said. “This is the type of thing you'd miss if you did a straight immunohistochemical analysis. You'd treat one plaque as being the same as the other.”
One caveat with the IP/Western approach used in this study is that it cannot detect heavier (>30 kDa) Aβ oligomers because they appear on the blot in the same molecular weight range as the non-specific bands from control IP samples. The technique also may not distinguish true dimers from higher-molecular-weight Aβ oligomers that fell apart during sample processing. We boil samples in SDS buffer,” Walsh said. “You wouldn’t expect non-covalent higher-molecular-weight assemblies to be resistant to SDS.” Last month, Japanese researchers published a sensitive ELISA that picks up large (40-200 kDa) Aβ oligomers—but not the much more prevalent monomers, dimers, or small oligomers (Fukumoto et al., 2010 and ARF related news story).
In the JNS paper, the Australian researchers steer clear of ELISAs and Westerns. They don’t even study brain tissue. Instead, lead investigator Barnham, first author Victor Villemagne, and colleagues used a technique known as SELDI-TOF (surface enhanced laser desorption ionization time of flight) mass spectrometry to detect Aβ monomers and dimers in the blood of living patients. What’s more, the researchers analyzed the part of the blood sample that is usually discarded, i.e., the pellet that contains erythrocytes, leukocytes, and platelets. They chose to look at this fraction based on studies suggesting that the toxic effects of oligomeric Aβ are mediated by its interaction with cell membranes (Glabe and Kayed, 2006; Hung et al., 2008).
Though more costly and lower throughput, mass spectrometry has several advantages over biochemical approaches—the primary one being specificity. “When you look on a Western blot, you see Aβ as single band,” Barnham said. That same sample appears as “a half dozen different species on mass spectrometry.” Most likely this is because APP gets cut by a range of other proteases besides the α-, β- and γ-secretases that AD researchers care about, he said. “There’s an absolute soup of APP species of various lengths, including Aβ, floating around in the tissue. Mass spectrometry allows you to pick out each of the individual species,” he said. It is also limited to smaller oligomers. “I wouldn’t trust anything above 20 kDa (tetramer). As molecular weight goes up, sensitivity goes down. Peaks broaden as well,” Barnham said.
Using SELDI-TOF to analyze blood from 118 seniors (43 AD, 23 mild cognitive impairment, 52 healthy controls), Barnham and colleagues found overall higher levels of Aβ dimers in AD patients compared with healthy controls. However, there was considerable overlap between the clinical groups. Dimer levels did correlate modestly with cognitive (e.g., Mini-Mental State Exam score) and neuroimaging (e.g., gray matter volume, fibrillar amyloid) measures.
Nevertheless, “it's really the first time we've been able to see oligomers in living subjects,” Barnham said, noting the technique might help identify other AD biomarkers. Another potential application may be in smaller clinical trials of anti-Aβ therapies. “The ability to look at what happens in blood is quite powerful, particular in small Phase 2 trials where you’re looking for a biomarker move,” he said.
For now, the researchers are looking to try their method on a new cohort to see if the current findings hold up. They also plan to look at how the mass spectrometry profiles change as people progress from normal to mild dementia to AD.
For their part, Walsh’s team has received government funding for a larger-scale analysis of 245 brains from three different centers. The researchers plan to relate soluble Aβ levels not only to the presence but also the severity of AD pathology. “That's really a big void that needs to be filled. Does soluble Aβ correlate with the degree of cognitive impairment? We need to hold soluble Aβ to the same criteria to which we hold plaques,” Walsh told ARF.
Such efforts will need to overcome daunting technical hurdles stemming from huge variations in Aβ concentrations, which can differ by up to three orders of magnitude depending on the protocol and/or cohort used. “If there’s a relationship between soluble Aβ and certain types of plaques, and the distribution of plaques is heterogenous throughout certain regions of brain, then depending on which piece of tissue you take, you may not find a direct correlation with clinical outcome,” Walsh said.
Perhaps Delphine Boche and James Nicoll of the University of Southampton, UK, were writing only partially tongue-in-cheek when they noted in an accompanying editorial that to non-biochemists, “the Aβ oligomer field can seem to border at times on the mystical.” Boche and Nicoll note that more precise refinement of Aβ detection methods will be important “not only to the AD field in general, but also specifically to the Aβ immunotherapy field. For example, it is not yet understood if Aβ immunization in AD lowers soluble/oligomeric Aβ as well as plaque Aβ.”
Questions about reliability and reproducibility of data across research centers also spawned rich discussion at the Human Amyloid Imaging Conference held in Toronto last month (see ARF related news story). On the CSF biomarker front, worldwide quality control initiatives have already sprung into action to deal with variable results (see ARF related news story).
In the meantime, the field’s intensified focus on oligomers is becoming more grounded. In two recently published oligomeric Aβ mouse models, soluble Aβ peptide levels correlated with cognitive impairment, even the absence of plaques (Tomiyama et al., 2010 and ARF related news story; Gandy et al., 2010).—Esther Landhuis.
McDonald JM, Savva GM, Brayne C, Welzel AT, Forster G, Shankar GM, Selkoe DJ, Ince PG, Walsh DM et al. The presence of sodium dodecyl sulphate-stable Abeta dimers is strongly associated with Alzheimer-type dementia. Brain. 133(5):1328-1241. May 2010. Advance Access published online 19 Apr 2010. Abstract
Boche D and Nicoll JA. Are we getting to grips with Alzheimer’s disease at last? Brain. 133(5):1297-1299. May 2010. Abstract
Villemagne VL, Perez KA, Pike KE, Kok WM, Rowe CC, White AR, Bourgeat P, Salvado O, Bedo J, Hutton CA, Faux NG, Masters CL, Barnham KJ. Blood-borne amyloid-beta dimer correlates with clinical markers of Alzheimer’s disease. J. Neurosci. 30(18):6315-6322. 5 May 2010. Abstract