In Alzheimer disease immunotherapy research, the peripheral sink hypothesis claims that by soaking up the amyloid-β (Aβ) in plasma, certain antibodies help drain the peptide from the central nervous system. Originating from experiments in mice using the m266 Aβ monoclonal antibody (see ARF related news story on DeMattos et al., 2001), the hypothesis suggests that passive immunotherapy may work without the antibodies ever getting into the brain. Results from a Phase 2 clinical trial of Solanezumab (also known as LY2062430), a humanized version of m266 created by Eli Lilly and Company, supports this idea. The trial showed that a dramatic increase in plasma Aβ occurs despite only a small percentage of the administered antibody turning up in the cerebrospinal fluid (see ARF related news story). But some new mouse data question the sink hypothesis, instead suggesting that antibodies can keep Aβ afloat inside the brain.

In the September 9 Journal of Neuroscience, researchers led by Takeshi Iwatsubo at the University of Tokyo, Japan, report that m266, which recognizes the central part of Aβ and binds primarily soluble forms of the peptide, helps to retain Aβ in the brain. “We set out to definitively prove the sink hypothesis because we like the idea, but to our surprise we found very clear retention [of brain Aβ] and retardation of clearance in the presence of 266,” Iwatsubo told ARF. The work was done in collaboration with researchers from Elan, which is pursuing its own passive (see ARF related news story) and active (see related ARF news) immunotherapy programs for AD.

What the finding means for passive immunotherapy is unclear, and given that much is at stake in ongoing human trials, none of the interviewed scientists were willing to speculate. Lilly is currently enrolling up to 2,000 patients in two Phase 3 clinical trials of the antibody (see EXPEDITION and EXPEDITION2 on

It is possible that we don’t have a complete understanding of the mechanism of passive immunotherapy with antibodies such as Solanezumab, Iwatsubo said. He suggested that the antibodies may have dual roles. In the periphery, they promote efflux of Aβ from the brain, while in the parenchyma a small fraction of antibody stabilizes soluble forms of Aβ. “That will not necessarily be bad for the brain, because it should sequester monomers and prevent them from forming more toxic oligomers or fibrils,” he said. An industry researcher, who declined to speak on the record citing company policy, noted that this new work was very well done and in concert with what Lilly scientists would likely expect themselves, since it is known that a small fraction of these antibodies get across the blood-brain barrier. Lilly researchers have already reported indications of a central effect of this antibody (see ARF related ICAD story).

To attempt to prove the sink hypothesis, first author Kaoru Yamada and colleagues quantified clearance of radiolabeled Aβ that they injected into the cortex of normal C57BL/6J mice who had previously been given m266 intraperitoneally. Yamada and colleagues expected to see efflux of the I125-Aβ accelerate in animals with m266, but instead found that the peptides stayed in the brain longer than in animals with labeled Aβ but no 266. In contrast, injecting the mice with 10D5, an antibody that recognizes the Aβ N-terminal and binds preferentially to Aβ fibrils, had no effect on the efflux of the radiolabeled Aβ. The Japanese researchers reasoned that the retention is probably due to m266 entering the brain and forming a complex with Aβ there. They used protein G, which binds immunoglobulins, to confirm their hunch. The scientists found that protein G captured about 45 percent of the I125-Aβ in m266-treated mice but only about 10 percent of Aβ from the brains of animals treated with 10D5 or no antibody. These results indicate that Aβ forms a complex with m266 in the brain.

Similarly, the authors found evidence for m266-Aβ complexes in the brains of four-month-old A7 transgenic mice. These animals express mutant human amyloid precursor protein, but their brains contain primarily soluble Aβ and have not yet begun to develop amyloid deposits. Using brain extracts prepared with or without guanidine hydrochloride to measure total and soluble Aβ, respectively, the authors found that soluble monomeric forms of the peptide are retained in the brain when the animals receive m266. Monomeric Aβ40 and 42 are elevated two- and threefold, respectively, 120 hours after injection of the antibody.

How long antigen-antibody complexes stay in the brain is not clear. Iwatsubo stressed that these findings do not exclude a sink effect in a therapy setting, noting that his group has measured only an acute effect, being limited to measuring Aβ efflux within two hours of injecting the radiolabeled peptide. The scientists also confirmed that in the absence of m266, the half-life of Aβ in the brain is about 30 minutes, a number that is in keeping with other reports. “So there may be some specific mechanisms related to rapid clearance of Aβ from brain,” he said.

Overall, Iwatsubo noted that the findings modify the current understanding of the mechanism of Aβ immunotherapy, raising the possibility that antibody-antigen reactions within the CNS play an important part. This idea is echoed by Yona Levites, Mayo Clinic, Jacksonville, Florida (see comment below). Levites writes that one of the unanswered questions now is what happens to the antibody-Aβ complex in the brain.—Tom Fagan


Make a Comment

Comments on this content

  1. Although a number of mechanisms have been proposed for how Aβ immunotherapy might work to prevent deposition, or clear, Aβ from the brain, no definitive answer has emerged. Some active and passive immunization studies show correlation between the efficacy of immunization and the ability of anti-Aβ monoclonal antibodies (mAbs) to recognize amyloid. Others show quite the opposite—certain mAbs that effectively reduce Aβ loads bind preferentially to monomeric Aβ. Such data raise the possibility that there may be multiple ways in which anti-Aβ antibodies influence amyloid deposition and other AD-like pathologies. However, important as it is, few studies were actually dedicated to the question of, How does anti-Aβ immunization work?

    In our mechanism of immunization study (Levites et al., 2006) we have shown that, at least in mice, the following statements are true:

    • Binding of mAbs to Aβ significantly prolongs the half-life of plasma Aβ.
    • Very little free anti-Aβ mAb actually enters the brain.
    • Anti-Aβ mAb:Aβ complexes are rapidly cleared from the brain.
    • Passive administration of anti-Aβ mAbs has little effect on total steady-state pre-deposition brain Aβ levels.

    Now, a very elegant study by Takeshi Iwatsubo and colleagues proposes a novel mechanism of anti-Aβ immunotherapy, whereby an antibody that binds soluble Aβ sequesters the peptide in the CNS. The researchers used radiolabeled Aβ, which enabled them to extract clean data and to show the dynamics of Aβ depletion from the brain with high accuracy.

    The questions that remain to be asked, and are crucial for predicting whether immunotherapy with this antibody will be efficient in humans, are the following:

    • How much of the antibody gets into the brain, and is it enough to sequester soluble Aβ, taking into account the Aβ accumulation rate in humans?
    • What exactly happens to antibody-Aβ complex in the CNS?
    • Could the latter promote unwanted inflammatory reactions?

    Additionally, the fact that only the anti-soluble Aβ antibody, not anti-fibrillar one, prolonged the half-life of radioactive Aβ in the brain might suggest that this antibody has better blood-brain barrier penetration qualities and presents at a higher level in the CNS, thus being more efficient.

    As we proposed in our studies, there might be great potential to an antibody that binds soluble oligomers with high affinity, and may be efficient in a therapeutic and not only prophylactic setting.


    . Insights into the mechanisms of action of anti-Abeta antibodies in Alzheimer's disease mouse models. FASEB J. 2006 Dec;20(14):2576-8. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Two Ways to Attack Amyloid: Metal Chelator and Antibody
  2. Chicago: Lilly’s Antibody Appears to Do No Harm, But Will It Help?
  3. Chicago: Bapineuzumab’s Phase 2—Was the Data Better Than the Spin?
  4. Research Brief: Elan/Wyeth Vaccine Back on Track

Paper Citations

  1. . Peripheral anti-A beta antibody alters CNS and plasma A beta clearance and decreases brain A beta burden in a mouse model of Alzheimer's disease. Proc Natl Acad Sci U S A. 2001 Jul 17;98(15):8850-5. PubMed.

External Citations


Further Reading

Primary Papers

  1. . Abeta immunotherapy: intracerebral sequestration of Abeta by an anti-Abeta monoclonal antibody 266 with high affinity to soluble Abeta. J Neurosci. 2009 Sep 9;29(36):11393-8. PubMed.