29 March 2012. Amyloid imaging with radiolabeled compounds has become a research staple for detecting and tracking amyloid plaques in humans, but the technique has not yet proven routinely feasible in mice. In a March 9 PLoS ONE paper, researchers led by Alexander Drzezga, Technische Universität München, Germany, outline a protocol with a new mouse model to image and validate plaques in living transgenic mice. The technique may enhance the efficiency of preclinical research and aid in the search for new amyloid imaging agents. "People were skeptical about whether animal imaging would be possible," said first author André Manook, who contends that the study's success stems from combining a research model that generates robust amyloid plaques early in life with recently available, higher-resolution small animal PET scanners and analysis software. "We are presenting a working and feasible protocol that we validated by pathological examination," Manook told Alzforum.
In people, one agent currently used for plaque detection in positron emission tomography (PET) is 11C-labeled Pittsburgh compound B, or PIB (see ARF related news story). While the compound binds well to plaques in human brains, in previous studies, it whizzed through transgenic mouse brains as quickly as it did through controls (see ARF related news story and Toyama et al., 2005). Bill Klunk, University of Pittsburgh, Pennsylvania, who co-developed PIB, proposed that there are far fewer high-affinity binding sites for PIB in mouse plaques (see Klunk et al., 2005). In another study, researchers got around that problem by using a higher radioactive tracer to make the PIB that did bind more visible (see ARF related news story on Maeda et al., 2007), but routine use of PIB-PET in animal models has never caught on.
Manook and colleagues took a slightly different tack to improve imaging with regular old PIB. He and his colleagues capitalized on a recently characterized mutant APP/PS1 transgenic mouse model called ARTE10. The homozygous mutant mouse breeds well, consistently accumulates amyloid in the brain, and shows episodic memory impairments at 12 months of age (see Willuweit et al., 2009). It starts accumulating Aβ early—around three months of age (five months if hemizygous)—and lives upwards of two years, which allows plenty of time—at least a year and a half—for observers to watch plaques develop. "The mouse model we used shows a huge amount of amyloid early in life, which may increase the measurable tracer signal," said Drzezga. He added that Manook's protocol likely contributed to the study's success. "It's similar to what is performed in human imaging."
Manook looked cross-sectionally at accumulated Aβ in 70 mice, which were divided into five groups. He examined aged hemizygous transgenic mice when they were 23 months old, young (nine months) and old (21 months) homozygotes, as well as young and old C57BL/6J controls. After injecting them with PIB, the researchers scanned representatives from the five groups—47 in all—in a mouse-sized PET scanner. Most also had a structural magnetic resonance imaging scan to locate important brain structures as a reference for PET analysis. In contrast to control animals, both homozygous and hemizygous transgenic mouse cortices bound PIB in amounts that corresponded to their respective pathologies. PIB signals in aged heterozygotes—with an intermediate number of plaques—fell between that of young homozygotes, which had the fewest plaques, and aged homozygotes, with the most. In these transgenic animals, the cerebellum had previously been shown to be free of plaques, so the researchers used it as a reference region for PIB binding.
To be sure that the PIB uptake reflected amyloid plaques, the researchers then validated their PET findings with a series of in-vitro and ex-vivo tests. Autoradiographs of brain slices after the animals had been injected with tritium-labeled PIB revealed that plaque-laden areas in the autoradiographs corresponded to areas of PET detection. Likewise, marking plaques with thioflavin S (which stains amyloid), and antibodies for Aβ40 and Aβ42, showed that plaque area and size matched the PET data. In addition, ELISA tests showed that higher amounts of insoluble Aβ40 and Aβ42 in the forebrain reflected elevated PET ligand binding. Further, the ex-vivo tests showed that plaques did not accumulate in the cerebellum, meaning it could be used as a reference region in this model. Together, the results suggest PET imaging works well in this animal model, wrote the authors. The group is trying the technique with other animal models, but is not yet ready to publish results, said Manook.
Until now, researchers tracked plaque deposition in mice by sacrificing a certain number of animals at set time periods to peer inside their brains and quantify plaque development over time. Not only does this method require large numbers of animals, but since animals often vary in their amyloid loads across time, the measurements are not always reliable, Drzezga said. "If you can image plaques in animals, then the overall number used in experiments could be reduced,” said Drzezga. "That is an ethical argument for these kinds of animal experiments," he told Alzforum. A reliable animal model for amyloid imaging could help screen potential agents for effectiveness before trying them in humans, he added.
PET imaging with animals will be a valuable tool, agreed Klunk. "There's really no substitute for longitudinal data in a single subject," he said. "You're going to get a lot less noise by following one mouse over time." For now, however, amyloid imaging may only work in this mouse model, which he thinks probably overcame the problem of few PIB binding sites by overproducing Aβ even more than other models. As with any model, but especially this one because it produces so much amyloid, he wonders how well findings in this model will translate to humans. Pending the resolution of some third-party intellectual property right issues, Taconic, headquartered in Hudson, New York, aims to make the ARTE10 model available for distribution some time this year.
In the near future, Drzezga and colleagues plan to further characterize the plaques that bind PIB in this mouse model to see if they differ qualitatively from those in other mice. Though they have not yet partnered with any pharmaceutical company, they are also thinking about using PET imaging to monitor the effectiveness of Aβ therapies in mice, and to evaluate potential new tracers—not just for Aβ, but for detection of tau and cerebral amyloid angiopathy as well.
Researchers in David Holtzman’s lab at Washington University, St. Louis, Missouri, have now, for mice, capitalized on a newly developed functional imaging technique to illuminate links between Aβ deposition and functional connectivity in animal models of AD (see ARF related news story). The combination of functional and PIB-PET imaging has the potential to bolster longitudinal studies in mice, said Holtzman. However, the resolution of PIB-PET imaging needs to improve before it is useful for this research, he said. Such PET improvements are in the works, said Drzezga. Meanwhile, "it might be interesting, even if we could not resolve small anatomical aspects, to see how amyloid levels overall increase in the brain in these animals, and how this affects the functional connectivity," he added.—Gwyneth Dickey Zakaib.
Manook A, Yousefi BH, Willuweit A, Platzer S, Reder S, Voss A, Huisman M, Settles M, Neff F, Velden J, Schoor M, von der Kammer, H, Wester H, Schwaiger M, Henriksen G, Drzezga A. Small-Animal PET Imaging of Amyloid-Beta Plaques with [11C]PIB and Its Multi-Modal Validation in an APP/PS1 Mouse Model of Alzheimer’s Disease. PLoS ONE 2012 March 9; 7(3): e31310. Abstract