Ever since encephalitis halted development of Elan’s active vaccine for Alzheimer disease (see ARF related news story), researchers have been racing to find a better alternative. Passive immunotherapy, the administration of antibodies rather than antigens, is considered safer, and there is ample evidence to suggest that it could help clear amyloid-β (Aβ) peptides from the brain (see ARF related news story, Wilcock et al., 2004, and ARF news story). But two recent reports indicate that antibody-mediated clearance of Aβ may be heavily influenced by age, at least in mice. The findings suggest that passive immunotherapy should be started early to be most effective, and that some antibodies might only work as prophylactics.
In the first paper, which appeared in the December 8 Journal of Clinical Investigation online, Todd Golde and colleagues at the Mayo Clinic College of Medicine, Jacksonville, Florida, report that some monoclonal antibodies to Aβ fail to attenuate deposition of the peptide in older mice. First author Yona Levites and colleagues tested three different monoclonal antibodies in prevention and therapeutic studies. One of the monoclonals, recognizing Aβ amino acids 1-16 as an epitope, bound Aβ1-40, Aβ1-42, and amyloid plaques. In addition to this “pan-Aβ” antibody, the authors tested two monoclonals that fail to bind plaques; these specifically recognize Aβ1-40 and Aβ1-42, respectively. Levites and colleagues found that the three monoclonals worked equally well when given to seven-month-old Tg2576 transgenic mice for four months. These animals express mutant human Aβ precursor protein (AβPP) harboring the Swedish mutations and begin to show behavioral deficits and plaque accumulation by about nine months. The authors found that at 11 months, mice treated with any of the antibodies had half the amount of SDS-soluble Aβ found in control animals. Plaque numbers were decreased by a similar amount.
But in mice already burdened with Aβ plaques, the antibodies did not work as well. When 11-month-old Tg2576 or three-month-old CRND8 mice were started on the Aβ1-40 or Aβ1-42 monoclonals, then the amount of Aβ in the brain four months later was not significantly different from that in controls. (CRND8 mice express human AβPP with Swedish and Indiana mutations, and already have detectable plaques by three months.) The pan-Aβ antibody fared slightly better, reducing SDS-soluble Aβ by about 40 percent in CRND8 animals and by about 60 percent when injected directly into the brains of 18-month-old Tg2576 mice. The Aβ1-40 and Aβ1-42 antibodies had no effect in this last experiment.
“We interpret our peripheral passive immunization data as being more consistent with preventing additional plaque deposition than with clearing existing plaques,” write the authors. What that means for human passive immunotherapy is a little hazy right now, especially given that other groups have found the technique does clear existing plaques. Dave Morgan and colleagues, for example, found that monoclonal antibodies against an Aβ28-40 antigen clears parenchymal plaques in mice that are even older (19-month-old Tg2576 animals), and it improves learning and memory (see Wilcock et al., 2004). Levites and colleagues suggest that the reason for the different findings might be related to the preponderance of diffuse plaques in some experimental models. In contrast to dense core plaques, diffuse plaques contain Aβ that many monoclonals recognize and perhaps clear. Whatever the reason, it is obvious from these experiments that not all antibodies are created equal. It will be interesting to see the outcome of the first human clinical trial AD passive immunotherapy (see Clinical Trials database).
In the meantime, the second paper, from Berislav Zlokovic and colleagues at the University of Rochester Medical Center in New York might shed some light on why the older mice do not respond as well to passive immunotherapy. Writing in the December 14 Journal of Neuroscience, they reveal that the mechanism for antibody-mediated efflux of Aβ across the blood-brain barrier (BBB) changes as mice get older.
First author Rashid Deane and colleagues report that in young mice, efflux of Aβ in complex with a monoclonal antibody (4G8, which recognizes an Aβ17-24 epitope) is mediated by either lipoprotein receptor-related protein (LRP) or the neonatal Fc receptor (FcRn). In aged mice, however, the LRP contribution is minimal, while FcRn-mediated efflux is increased. The authors found, for example, that 4G8-mediated clearance of endogenous Aβ from the brain of nine-month-old FcRn-negative mice was negligible. In normal mice, on the other hand, the passive antibody treatment halved the amount of Aβ in the cerebral cortex.
The loss of LRP-mediated efflux might explain why the authors detected increasing influx of Aβ into the brain of transgenic mice once they reached about 15-20 months. They also reported that this influx can be attenuated by passive immunotherapy with the 4G8 antibody. When APPSwe+/- mice were started on the treatment at 2-3 months, they had significantly lower influx of Aβ into the brain later in life.
Based on these findings, Deane and colleagues propose two ways to improve clearance of Aβ across the BBB—maintain or even increase the activity of the LRP pathway, and optimize the FcRn pathway. The latter may, in fact, happen naturally because the authors found that FcRn is elevated almost fourfold in brain samples from AD patients compared to controls. “Increased FcRn expression at the BBB in AD may likely be of a therapeutic value for Aβ immunotherapy,” the authors write.
Both papers also appear to agree that early is better for passive immunotherapy. That’s assuming, of course, that the passive approach is truly safer than other forms of immunization. Recent work from John Trojanowski’s lab challenges that view and shows that passive Aβ immunotherapy can, in fact, lead to encephalitis in mice (see Lee et al., 2005). If that finding repeats itself in the human trial, it would send efforts back to square one.—Tom Fagan