This concludes our 3-part series. Also see part 1 and part 2 or download PDF.

Scientists continue to debate the relative merit of using N-terminal versus mid-section or C-terminal antibodies, and of using antibodies against soluble versus against fibrillar Aβ. The discussion turns, in part, on the issue of microhemorrhages. Concern about small bleeds inside the brain arose when several groups began reporting them in the brains of immunized mice, especially mice who carry a heavy load of amyloid deposits in the blood vessel walls of their brains, not just in the parenchymal spaces. Called cerebral amyloid angiopathy (CAA), this pathology exists in the majority of AD patients. CAA is estimated to occur in up to 30 percent of elderly people, though that number varies. CAA can cause hemorrhagic strokes and cognitive impairment, and by itself is thought to cause numerous small bleeds that often go unnoticed.

The condition is attracting increasing attention from researchers. At the 10th ICAD meeting, held July 15 to 20 in Madrid, 23 presentations dealt with the topic of CAA and microbleeds. Presentations ranged from one showing that CAA pathology intensifies along with AD pathology, albeit in somewhat different brain regions, to another suggesting that the amyloid imaging agent PIB-PET can detect the region-specific CAA pattern in non-demented, living people. At the epidemiological level, scientists led by Meike Vernooij and colleagues at Erasmus University Medical School in Rotterdam used MRI to measure the frequency of microbleeds in 491 community-based participants of the Rotterdam Scan study. They found that 17 percent had such bleeds in their brains. Older people were at higher risk than younger people, and high blood pressure and use of thrombolytic agents upped the risk some more.

The worry with regard to CAA and AD immunotherapy is that clearance of blood vessel amyloid by anti-Aβ antibodies might lead to ruptures of the already dysfunctional vessel wall (Pfeifer et al., 2002; Burbach et al., 2006). A related worry is that overly fast removal of parenchymal amyloid would overwhelm the clearance capacity at nearby blood vessels and lead to renewed deposition of the amyloid there (Wilcock et al., 2004). Intriguingly, the sugar groups decorating a given antibody appear to influence this phenomenon (Wilcock et al., 2006). RN1219, a C-terminal Aβ antibody being developed by the San Francisco biotechnology company Rinat Neuroscience, a Genentech spinoff, is said to have an edge on that score.

This issue is in flux, and in Madrid, Sally Schroeter of Elan Pharmaceuticals addressed it with data from a 6-month study comparing the effects of several Aβ antibodies on CAA in the PDAPP mouse model. In short, Schroeter reported that, to her surprise, the 3D6 antibody, which recognizes fibrillar Aβ, not only did not worsen CAA, but instead cleared it in a dose-dependent fashion. The m266 capture antibody predictably had no effect on CAA. A multiphoton imaging study by Claudia Prada and colleagues at Massachusetts General Hospital in Charlestown paralleled this data by observing in live mice that a different antibody given to Tg2576 mice caused CAA to regress.

Furthermore, in Schroeter’s study, microhemorrhages did occur in conjunction with the CAA clearance, she acknowledged, but could be limited by reducing the antibody dose, essentially tuning down amyloid clearance to a manageable rate. David Morgan, of University of Southern Florida, who has collaborated with Rinat Neuroscience but has no financial interest in the company, pointed out that at the age at which Schroeter and colleagues began treating the PDAPP mice—12 months—the mice’s brains do not yet have a full load of parenchymal amyloid and also have less CAA than many AD patients. Repeating the study in older mice might more closely mimic the amyloid and CAA burden of AD patients and would test the findings raised by Wilcock et al. and Pfeifer et al. under more comparable conditions. For a prior comparison of 3D6 and m266 in older PDAPP mice, led by Ron DeMattos at Lilly, see Racke et al., 2005.

Yet another factor that young PDAPP mice represent poorly is the inflammation in blood vessels laden with CAA. In Madrid, Manuel Buttini and colleagues of Elan Pharmaceuticals in South San Francisco characterized inflammatory infiltrates in brain blood vessels of elderly people with AD and young normal controls, and found that more than twice the percentage of CAA vessels than CAA-free vessels had monocytes and T cells near them. “Age still is the most important risk factor in AD,” Morgan summed up.

But Morgan also emphasized that it’s all but clear how important the microbleeds will prove to be clinically in people. Clot busters such as the stroke treatment tissue plasminogen activator are known to cause microbleeds and, at least in acute situations, the risk is tolerated. Morgan’s antibody-treated mice, for all they are worth, retained the behavioral improvement despite having more of these bleeds. “Minute hemorrhages are not the big worry. Large ones are,” Morgan said. In summary, numerous scientists asked about this issue tended to agree that they expected some form of immunotherapy to slow progression of AD. But they also worry that some patients who receive these therapies for extended periods of time might develop larger hemorrhages, forcing the FDA’s hand against the approach.

Little phage to the rescue?

If reading all this leaves you scratching your head about how the pitfalls of AD immunotherapy can possibly be avoided, you are not alone. Some scientists are also looking for alternatives. One of them, Beka Solomon of Tel Aviv University, who has worked on AD immunotherapy for a decade (Solomon et al., 1997) happily announced a fortuitous observation that led her to begin exploring such an alternative. Solomon has long studied the potential of bacteriophages for ferrying either Aβ antigens or antibodies into the brain. It was during one of those studies that a control—naked phage—surprised her when it performed just as well as the study drug—phage studded with single-chain Aβ antibody—on her measures of reduction in amyloid burden. It seemed to do so more safely, too. Perhaps this humble life form could make for a treatment?

Bacteriophages come in two basic varieties. The better-known ones that lyse bacteria have been used as an alternative to antibiotics in the former Soviet Union, but Solomon uses the non-lytic phages. At almost a micrometer in length but only nine nanometers in width, they look like microscopic filaments of DNA packaged into a narrow protein sheath. Their physicochemical properties enable them to slip through membranes easily. In prior years, Solomon has explored the phages’ potential to carry foreign peptides into the brain of animals (e.g., Lavie et al., 2004), to serve as immunogens (Frenkel et al., 2000), and to exert effects in the brain when sprayed into the nose (Frenkel and Solomon, 2002). (Incidentally, the scent of intranasal delivery wafted through the ICAD conference, with some presentations extolling the benefits of this route and others describing its use in the delivery of specific experimental drugs. Examples include a talk by Suzanne Craft at the University of Washington, Seattle, on intranasal insulin and verbal memory in early AD, and another by Illana Gozes of Tel Aviv University on the neuroprotective peptide NAP, which acts to stabilize microtubules and is in early clinical trials with an intranasal formulation.)

In her studies with the phages, Solomon discovered that they alone, even without sporting Aβ or an antibody at their tips, had useful effects in AD models. Apparently, their size and structure allow them not only to penetrate the brain when given through the nose but also to intercalate into the β-sheet structure of amyloid and disrupt it. Electron microscopy images showed immuno-gold labeled Aβ fibrils alone, and amorphous Aβ aggregates in the presence of filamentous phage. The phages’ threadlike shape did the trick, because when hooked into spheres, the phages no longer busted amyloid fibrils.

Solomon then showed in-vitro data on the phage’s disaggregating properties, and on their ability to stain amyloid plaques. Injection of filamentous phage alone into the brains of Tg2576 mice reduced the mice’s amyloid load over the course of three days. A subsequent one-year study of biweekly, then monthly, administration of phage sprayed up the noses of PDAPP mice improved the mice’s memory performance in an object recognition test and a smell test. No water maze data were given. The phage also reduced the amyloid burden in the mice’s brain and increased their synaptophysin levels, Solomon said. Notably, unlike the untreated transgenic mice, the phage-treated mice had no astrocytosis in their hippocampuses. Microgliosis showed no difference between the groups. Microhemorrhages were undetectable in the phage-treated mice, Solomon stressed. Peripheral organs also suffered no ill effects, in keeping with the established safety profile of these organisms. The phages leave the body within 3 weeks, Solomon said. They exit the brain with the help of microglia and are concomitantly eliminated from the body by urine and feces.

Solomon did not show data about the detailed in vitro-in vivo correlations, pharmacokinetic data, and dose-effect curves that would become necessary once drug developers became interested in this approach. Likewise, other scientists were curious to see short-term in-vivo studies that track how Aβ levels change in brain and CSF soon after phage delivery. Presumably, levels of Aβ would initially go up as it gets freed from fibrils, before eventually being cleared. Solomon said this was her initial presentation on the approach and much remains to be done. But already, it is clear that the phages are safe and ubiquitous in the environment, she said. “You and I can pick them up by swallowing a bit of water while swimming outdoors, and we won’t even know it,” Solomon added. They are dirt cheap, too.—Gabrielle Strobel.


  1. It is true that at 12 months of age, the PDAPP mice have not yet developed their maximum extent of plaque burden and CAA pathology. However, we have identified "hot spots" that reliably develop dense vascular deposits that are clearly thioflavin S-positive and intensely immunoreactive with Aβ antibodies. Regions of concentrated amyloid deposits in the meningeal vasculature are commonly present in 12-month-old PDAPP mice and were identified in a detailed study that goes beyond previously published studies. Furthermore, our studies were conducted over a time of escalating pathology, during which amyloid pathology increases at least 20-fold, resulting in the very abundant plaque burden that Peter Seubert showed in his presentation.

    We do agree that our initial findings would be interesting to test in older mice, and these studies are currently being considered. We continue to be encouraged by our findings that no microhemorrhage was associated with CAA removal in the majority of treated mice, and that the observed instances could be significantly mitigated by modulation of dose.

    With regard to the coverage of Manuel Buttini's presentation: The relevance of the statement about PDAPP mice poorly representing blood vessel inflammation is unclear given that his study was conducted using human AD tissues. We have not yet performed a similar analysis in PDAPP mice, so we believe it is premature to say that the pathology is not reflected in PDAPP mice.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Madrid: News from the Vaccine Front—Phase 2 Postmortem, Part 1
  2. Madrid: News from the Vaccine Front—Phase 1 Hopefuls

Paper Citations

  1. . Cerebral hemorrhage after passive anti-Abeta immunotherapy. Science. 2002 Nov 15;298(5597):1379. PubMed.
  2. . Vessel ultrastructure in APP23 transgenic mice after passive anti-Abeta immunotherapy and subsequent intracerebral hemorrhage. Neurobiol Aging. 2007 Feb;28(2):202-12. PubMed.
  3. . Passive immunotherapy against Abeta in aged APP-transgenic mice reverses cognitive deficits and depletes parenchymal amyloid deposits in spite of increased vascular amyloid and microhemorrhage. J Neuroinflammation. 2004 Dec 8;1(1):24. PubMed.
  4. . Deglycosylated anti-amyloid-beta antibodies eliminate cognitive deficits and reduce parenchymal amyloid with minimal vascular consequences in aged amyloid precursor protein transgenic mice. J Neurosci. 2006 May 17;26(20):5340-6. PubMed.
  5. . Exacerbation of cerebral amyloid angiopathy-associated microhemorrhage in amyloid precursor protein transgenic mice by immunotherapy is dependent on antibody recognition of deposited forms of amyloid beta. J Neurosci. 2005 Jan 19;25(3):629-36. PubMed.
  6. . Disaggregation of Alzheimer beta-amyloid by site-directed mAb. Proc Natl Acad Sci U S A. 1997 Apr 15;94(8):4109-12. PubMed.
  7. . EFRH-phage immunization of Alzheimer's disease animal model improves behavioral performance in Morris water maze trials. J Mol Neurosci. 2004;24(1):105-13. PubMed.
  8. . Immunization against Alzheimer's beta -amyloid plaques via EFRH phage administration. Proc Natl Acad Sci U S A. 2000 Oct 10;97(21):11455-9. PubMed.
  9. . Filamentous phage as vector-mediated antibody delivery to the brain. Proc Natl Acad Sci U S A. 2002 Apr 16;99(8):5675-9. PubMed.

Other Citations

  1. PDF

Further Reading