8 December 2012. Most Aβ antibodies in development can prevent plaque buildup, but few eliminate existing plaques. Enter a new monoclonal antibody developed by scientists at Eli Lilly and Company, Indianapolis, Indiana. Dubbed mE8, it seems to target plaques directly by binding the rare pyroglutamate form of Aβ (pGluAβ). As reported in the December 6 Neuron, mE8 bound the modified peptide in amyloid deposits in an AD mouse model, activated microglia to clear them, and did it all without causing microhemorrhages, which has been a problem in AD immunotherapy. Results line up with data published earlier this year by Cynthia Lemere and colleagues at Brigham and Women’s Hospital, Boston, and could have implications for Alzheimer’s treatment.
Developing "therapies that can effectively remove and decrease toxicity of pre-existing plaques and CAA [cerebral amyloid angiopathy] without causing side effects is a major goal," wrote David Holtzman and Jee Hoon Roh of Washington University School of Medicine, St. Louis, Missouri, in an accompanying editorial.
Anti-Aβ antibodies tested in clinical trials bind both soluble and insoluble forms of Aβ, but have not been shown to clear larger deposits. Ronald DeMattos and colleagues at Lilly wanted to raise an antibody that would target and clear plaques directly. For an antigen, the researchers zeroed in on pGluAβ, which is found almost exclusively in amyloid deposits but not in blood or cerebrospinal fluid. The peptide forms when two amino acids are clipped from the N-terminus of Aβ, leaving a glutamate that is enzymatically cyclized. The resulting pGluAβ resists degradation, is extremely hydrophobic, and aggregates quickly.
To create a pGluAβ antibody, the researchers adopted a similar monoclonal strategy to that used by Lemere’s group (see Frost et al., 2012). DeMattos and colleagues immunized mice with pGluAβ and, after cloning B cells, found an antibody—mE8—that recognized only that form of Aβ. Though it bound both soluble and insoluble pGluAβ, the team hypothesized the antibody would be plaque specific, since pGluAβ exists almost entirely in those deposits.
Engineered on an IgG1 backbone, mE8 weakly activated microglia, while on an IgG2a backbone the antibody elicited a stronger microglial response. IgG2 antibodies typically have better effector function, meaning they robustly stimulate immune cells. To test these antibodies in vivo, the scientists intraperitoneally injected aged, plaque-laden PDAPP mice (Johnson-Wood et al., 1997) (23 to 24 months old) weekly with mE8-IgG1, mE8-IgG2a, or 3D6, an antibody that binds to soluble and insoluble Aβ42. 3D6 is the mouse forerunner of the human bapineuzumab antibody (see ARF related news story).
After three months, animals treated with mE8-IgG1 and mE8-IgG2a contained 38 and 53 percent less Aβ, respectively, than controls. The difference between the two IgG versions suggests that stronger microglial activation translates to greater Aβ clearance. “This confirms that microglia are likely one of the key players in plaque clearance,” DeMattos told Alzforum. These mice showed no signs of microhemorrhage, which has plagued development of anti-Aβ immunotherapy in humans (see ARF related news story and ARF news story).
While the mE8 antibody helped clear plaques, it did not prevent their formation as effectively. Though young mice (5.5-months-old) treated with mE8-IgG2a for seven months accumulated 30 percent less plaque relative to controls, the difference was not significant. In contrast, mice treated with 3D6 made 70 percent less plaque during the same period. The mE8 antibody may not have worked as well in younger animals because of a dearth of pGluAβ early in disease, the authors wrote. However, Lemere found that mAb07/1 lowered plaque burden in both treatment and prevention paradigms. She is unsure why mAb07/1 and mE8 would differ in this respect. “We believe one of the reasons we see prevention is because pGluAβ can act as a seed for general Aβ deposition,” she told Alzforum. Altogether, DeMattos’ results suggest that antibodies that bind soluble Aβ help prevent plaque deposition, while those that mostly bind plaques, such as mE8, primarily clear existing amyloid.
What mechanism underlies the difference between antibody types? DeMattos suggested that antibodies that bind soluble Aβ, such as 3D6, do not clear plaques well because they become mired in the halo of free Aβ surrounding them. Saturated with these smaller Aβ species, the antibodies never reach the plaques. However, since mE8 specifically binds pGluAβ, it slips past the cloud of Aβ and goes directly to the plaques. That may also explain why mE8 does not cause microhemorrhages. DeMattos and colleagues claim that mE8 is trapped by plaques until cleared by microglia. Antibodies that do not bind plaques, on the other hand, are free to diffuse to the vasculature where they release their Aβ cargo, which then builds up along vessels, causing microbleeds (see Winkler et al., 2001). “It suggests that you can clear existing plaques by phagocytosis and not have this [microhemorrhage] liability,” said DeMattos.
How can an antibody that recognizes such an uncommon Aβ peptide induce such substantial plaque reduction? Lemere suggested in her paper that once pGluAβ antibodies bind plaques, they activate microglia, which then clear Aβ indiscriminately. "That's quite an interesting concept," agreed Dennis Selkoe, Brigham and Women’s Hospital. "It suggests that activating microglia via just a subset of Aβ peptides allows a more generalized phagocytic clearing of plaque amyloid."
Does this ultimately help the mice? “This study lacks any information about beneficial effects, such as rescue of synaptotoxicity, neurotoxicity, or behavioral deficits,” said Thomas Bayer, University of Göttingen, Germany. He previously reported that a pGluAβ antibody specific for non-plaque oligomers reduced plaque deposition and alleviated anxiety in an AD mouse model (see Wirths et al., 2010). He believes that pGluAβ oligomers, while absent from CSF and blood, exist between or within neurons, where they exert their toxicity—thus representing the ideal target. Lemere’s group plans to present behavioral, pathological, and biochemical characterization of their pGluAβ antibody this coming March at the AD/PD meeting in Florence, Italy.
DeMattos’ results may have implications for human AD treatment, speculated some. For example, recent results on solanezumab—Eli Lilly's humanized monoclonal antibody that binds soluble Aβ—indicate a modest cognitive benefit and trend toward reduced amyloid burden in a mild AD patient subgroup (see ARF related news story). “This paper leads me to suggest combining solanezumab with a humanized form of a plaque-directed antibody like mE8,” said Selkoe. “The two antibodies might complement one another, with solanezumab targeting soluble monomers and an mE8-like antibody targeting fibrillar plaques.”
The mE8 antibody is not suitable for human use in its murine form, but Lilly has humanized it and plans to begin a Phase 1 clinical trial by 2014, a Lilly representative told Alzforum via e-mail. Whether pGluAβ antibodies will behave the same way in humans as they do in mice is unclear, however. Some evidence suggests much more pGluAβ exists in the AD brain than in transgenic mouse models (see ARF related news story), though DeMattos says his data consistently indicate similar levels in AD brain as in the PDAPP mice. If pGluAβ was more prominent, it might make up more of the peri-plaque halo of Aβ and contribute to microhemorrhage. “A real test of this concern would be to administer chronic passive immunization against pGluAβ in a mouse that overexpresses and deposits it,” said Lemere.—Gwyneth Dickey Zakaib.
DeMattos R, Lu J, Tang Y, Racke MM, DeLong CA, Tzaferis JA, Hole JT, Forster BM, McDonnell PC, Liu F, Kinely RD, Jordan WH, Hutton ML. A Plaque-Specific Antibody Clears Existing β-amyloid Plaques in Alzheimer’s Disease Mice. Neuron 2012 Dec 6;76: 908-920. Abstract
Roh JH, Holtzman DM. Stealth Attack: Plaque-Specific Antibody Allows for Efficient Aβ Removal without Side Effects. Neuron 2012 Dec 6;76: 859-861. Abstract