07 Dec 2007

Understanding just how anti-amyloid antibodies remove amyloid-β (Aβ) from the brain is a burning issue in the design and fine-tuning of immunological treatments for Alzheimer disease. Two distinct but not mutually exclusive mechanisms have been proposed. One (the “CNS clearance hypothesis”) requires the antibodies to enter the brain, where they mediate uptake of Aβ into local microglia. The other (the “peripheral sink hypothesis”) relies only on peripheral antibodies to sequester Aβ in the blood, lowering the level of free Aβ and inducing the brain to release its store of the peptide. A new study in a novel strain of APP mice validates the peripheral sink as the major mechanism of Aβ removal after immunization in this system. The study shows that, at least in this strain of mice, little or no antibody enters the brain. The results, from David Cribbs and colleagues at the University of California, Irvine, appear in the December 5 Journal of Neuroscience.

To study Aβ clearance, first author Vitaly Vasilevko and colleagues used the Tg-SwDI mouse, a strain created by coauthor William Van Nostrand of Stony Brook University in New York. That strain expresses low levels of human APP with both the Swedish mutations and the vasculotropic Dutch and Iowa mutations (Davis et al., 2004). The Dutch and Iowa mutations result in an Aβ peptide that is a poor substrate for the efflux transporter LRP-1, and so accumulates to high levels in the brain. In effect the mice have no peripheral sink, and despite a massive buildup of vascular amyloid and parenchymal plaque in brain, Aβ remains undetectable in their blood (Deane et al., 2004; Davis et al., 2006).

Immunizing either young or old Tg-SwDI mice with a potent Aβ epitope vaccine (see ARF related news story) resulted in extremely high titers of antibody. Affinity-purified antibody was fully capable of binding to Aβ-DI peptides or oligomers and blocked fibril assembly in vitro. However, the immunization procedure neither cleared established plaque when performed on older mice, nor prevented plaque buildup in younger animals. Passive immunization with purified antibodies also had no effect. In essence, the animals had high amounts of Aβ in the brain and high amounts of antibodies to this Aβ in the periphery, but the two never saw each other.

However, the antibodies were capable of clearing Aβ when delivered directly to the brain by intrahippocampal injection. In that case, the antibodies caused a marked reduction in parenchymal Aβ in the area around the site of injection. That did not help the immunized mice, however, as direct measurements of brain extracts revealed that little or no antibody was able to enter the brain from the periphery. From these results, the investigators conclude that, in this mouse model, not nearly enough antibody crosses the blood-brain barrier to allow effective clearance of Aβ despite blood levels of antibody that reach an estimated 50 times higher than that achieved in the trial of the Elan/Wyeth’s AN1792 vaccine in humans.

“Our results provide support for the peripheral sink hypothesis as a viable mechanism for anti-Aβ antibody-mediated clearance of Aβ from the CNS when the BBB remains intact,” the authors write. They go on to express their concern about relying on peripheral clearance across the blood-brain barrier in aged people, in whom the efficiency of Aβ export may be compromised. This, plus the risk of antibody-mediated hemorrhages at sites of cerebral amyloid angiopathy, reduces the authors’ enthusiasm for peripheral, passive or active, antibody-mediated reduction of CNS Aβ. These authors expect direct delivery of immunotherapy to the CNS to be more efficient.—Pat McCaffrey