After a tough year, during which believers in immunotherapy for Alzheimer’s disease have had to fend off vociferous criticism, the news now appears to pick up. In tomorrow’s Neuron, Swiss researchers announce that among the AD patients who received the stalled Elan/Wyeth-Ayerst AN1792 vaccine, those whose immune systems made antibodies against the injected Aβ preparation indeed enjoyed a clinical benefit. They held steady, or slowed their decline, on several different measures of cognitive function and daily living, report Christoph Hock, Roger Nitsch, and colleagues at the University of Zurich.
This phase 2A trial was suspended last January after 17 of the 300 study participants developed meningoencephalitis (see ARF live discussion), which has since been successfully treated in most, but not all cases. In an accompanying preview article, Bengt Winblad of Stockholm’s Karolinska Institute and Neuron editor Kenneth Blum note that this potentially fatal side effect remains an "overriding concern." Even so, they also write "this article shows that the concept of vaccination is alive."
In the present study, Hock et al. followed up their prior analysis of antibodies generated in the 30 members of the Zurich cohort of this multicenter trial (see ARF related news story) to find out if mounting an antibody response did the patients any good. They still do not know which study participant received a primer and a booster shot of vaccine (24) or of placebo (6). Instead, they studied the generation of antibodies against β-amyloid plaques on brain tissue sections by using their newly developed TAPIR assay, and correlated these data with their cohort’s clinical performance at baseline, and then eight and 12 months after immunization.
Twenty patients developed β-amyloid antibodies and 19 of those were analyzed; the other 10 were regarded as controls for this study. The 19 antibody generators remained stable on the Mini Mental State Exam (MMSE), whereas the controls declined. Hock et al. write that the clinical stabilization in the 19 responders differs markedly from the published natural history of AD. On this result, Winblad and Blum caution that the rate of decline of the controls was steeper than normal, and that this could prove to be a confounder when the data from the other study cohorts is analyzed.
The patients’ caregivers also appeared to notice an effect. Hock et al. interviewed them in a double-blind setup using the Disability Assessment for Dementia (DAD) rating scale, which measures the patient’s ability to perform activities of daily living independently. Again, those making amyloid-β antibodies scored better, hinting that the cognitive effects detected with the MMSE translated into a practical benefit. A test of hippocampal function produced a statistically significant improvement, while other neuropsychological instruments showed only trends.
To measure the antibody response, Hock et al. used an assay they had developed earlier, which measures the patient’s serum antibodies directed against β-amyloid in brain slices of AβPP-transgenic mice. Called tissue amyloid plaque immunoreactivity (TAPIR), this assay correlated better with clinical benefit than did more customary ELISA tests measuring the binding of antibodies against synthetic Aβ. Those with the highest TAPIR scores also enjoyed the greatest protection from disease progression. This could mean that functional improvement depends on conformation-specific antibodies that TAPIR is better suited to picking up than ELISA, the authors suggest.
Finally, the paper contains data suggesting that serum antibodies remained high in the patients for the entire year. The researchers did not, however, note changes in plasma and CSF levels of Aβ, indicating that the present data does not support the peripheral sink hypothesis (see ARF related news story).
Do the responders have less Aβ in their brains? While an autopsy case suggests as much (see ARF related news story), correlations with pathology in these living patients would require a live imaging technique (see ARF related news story). And how about neurofibrillary pathology, synaptic dysfunction, and neuronal degeneration? It remains unclear whether Aβ immunization can affect the damage wrought by these factors.
In their discussion, the authors draw an analogy to infectious disease, where the nature of an infectious agent (or in this case β-amyloid as the pathogenic agent) is proven by way of transmission and vaccination. Similarly, Hock et al. write, the vaccine trial tests the amyloid cascade hypothesis, and the present study supports this hypothesis by providing "the first successful clinical evidence for a central role of β-amyloid in causing cognitive decline and dementia in AD patients." In summary, Winblad and Blum call the present results impressive, but caution that they should be regarded as preliminary until data from larger patient numbers become available.—Gabrielle Strobel
Q&A with Roger Nitsch-Posted 26 May 2003.
Q: The meningoencephalitis was successfully treated in most patients. What happened to the others?
A: A complete account of all clinical details of every single patient with meningoencephalitis is in press and will be published separately in one of the upcoming issues of Neurology (Orgogozo et al., 2003 in our reference list). In our study, we had three patients who suffered from meningoencephalitis. Two of them who had developed antibodies improved rapidly upon treatment with cortisone, and subsequently enjoyed the full beneficial clinical effects. They were among the best responders. One patient in our group with meningoencephalitis but without antibodies against β amyloid also improved upon cortisone treatment, but unfortunately continued to decline cognitively.
Q: Are preparations underway for a clinical trial for second-generation immunization protocols, in which the Zurich site will participate?
A: Yes. Several companies are preparing for second-generation trials. We are interested to participate in these, and we are also pursuing independent avenues.
Q: Can amyloid load in these patients still be analyzed with emerging live-imaging techniques, for example, the Pittsburgh compounds?
A: Yes. The problem is, though, that there are no baseline data to compare with. But the sooner in-vivo analyses are started, the more likely it is that meaningful data can be obtained.
Q: How long do you expect the antibody titers to stay high without further booster shots?
A: To date, there are still high levels of antibodies. I hope that they will remain high in the future-we are closely following up on our patients to obtain these data.
Q: Will you continue following these patients?
Q: How far has analysis of the other 90 percent of participants progressed, and are you collaborating with those other groups for the final analysis?
A: This is internal Elan information. Elan is currently using our TAPIR protocol to analyze the multicenter trial.
Q: Alzforum recently covered the paper by Rakez Kayed et al. A common structure of soluble amyloid oligomers implies a common mechanism of pathogenesis (see ARF related news story). Does this finding dovetail with your working hypothesis about conformation-specific antibodies?
A: We are currently collaborating with Charlie Glabe to find out.
Q&A with Steven Paul, Lilly Research Laboratories, Indianapolis-Posted 23 May 2003.
Q: What do you think of this study?
A: This is potentially an important paper. I am quite excited by the findings. It is clearly a positive sign that 20 of the 30 subjects in this cohort generated reasonably high circulating antibody titers to this Aβ species. At the same time, the clinical data is preliminary. Methodological issues related to the small sample size prevent one from being definitive at this point. Because the Ns are so small, there is the possibility of a false-positive result. Having said that, the indications that these patients develop very little further cognitive impairment compared to the controls is interesting and encouraging.
I am pleased to see the Zurich group report this. In a recent article (Dodart et al., 2003), we speculated that the participating sites in this large trial would eventually publish data not only on safety and adverse events, but also on efficacy, which is clearly the most important aspect. I would encourage Elan and Wyeth-Ayerst to gather the same data from the other sites. All told, there must have been well over 100 AD patients who developed antibodies. The sponsors could make the analysis much more powerful if they analyzed the data from all patients. A more definitive result would be quite important.
Q: Anything noteworthy about the meningoencephalitis?
A: The authors found no relationship between the three cases of meningoencephalitis in their cohort and circulating antibody titer as measured in their TAPIR assay; again, this is very preliminary due to the small numbers. This finding suggests, however, that this adverse event was most likely due to a cell-mediated response, and is not part and parcel to the underlying mechanism responsible for the observed therapeutic effect. It seems at least hopeful from this study that one could craft a better antigen for active immunization or a much better-defined antibody for passive immunization that would circumvent the emergence of this adverse event.
Q: Do we know how safe an AD immunotherapy must be?
A: The question will always be what is the acceptable margin of safety for such a treatment. This issue has perplexed me from the beginning: Alzheimer’s disease is fundamentally fatal, not unlike many cancers. The acceptable serious adverse event rate for an effective anticancer medication is much higher than five percent. We must do everything we can to eliminate serious adverse events, particularly if they are unrelated to the therapeutic mechanism. But in the event that we still have an adverse event rate of five percent for an effective treatment, we must have a discussion about whether it is acceptable, given the alternative. Academic and private investigators, working with the FDA, need to think through what is an acceptable margin of safety for Alzheimer’s disease. If you had a loved one who had early stages of AD and you knew a given treatment really worked, would you accept a one-in-20 serious adverse event rate?
Also, it is important to recognize that none of the current mouse models of AD develop profound neurodegeneration; there is almost no loss of gray matter. By the time you can diagnose AD, the person has already lost considerable gray matter, especially in certain brain regions. So if these treatments are to be completely effective, they will need to be started earlier. Preferably you may even want to treat those individuals who are at high genetic risk for developing AD, but while they are still asymptomatic. That means we need agreement on safety standards, better biomarkers, better diagnostic tests and imaging technology to diagnose and monitor treatment. In this study, while the patients were not deteriorating much, they still had cognitive impairment. This is not surprising because they have lost synapses and neurons already, and this treatment will not repair that.
Q: How about fibrillized vs. soluble Aβ?
A: Hock et al have some suggestion that the vaccine-generated antibodies are specific to a deposited, "pathological" epitope that forms as the peptide aggregates and deposits in the brain as amyloid. It seems these antibodies did not recognize soluble Aβ. Hock et al. cite several of our papers dealing with the antibody M266, which recognizes soluble Aβ (see, for example, ARF related news story). They claim that the antibodies generated in the patients don’t recognize soluble Aβ and, therefore, may not recognize the small, oligomeric Aβ that Bill Klein, Charlie Glabe, Dennis Selkoe, our group, and others have worked on. I think that it’s fair, but not definitive. The antibodies need to be characterized further.
Q: How about peripheral Aβ?
A: Hock et al. claim that the levels of Aβ in the plasma and the CSF do not change. As I inspect the scatter grams in Figure 4, panel b looks to me as if there is a trend toward an increase in Aβ42, the more pathological species, in the CSF of immunized patients. That would be interesting, given some of the work our group and collaborators have done. Theoretically, it could mean that the antibody, which Hock et al. demonstrate is in the CSF, is attached to Aβ and is preventing its clearance. Or it could be that the antibody in the blood is causing CSF Aβ to go up because it is coming out of the brain. But again, one needs to establish definitively if CSF Aβ levels are elevated or not.
- The Alzheimer's Vaccination Story, Continued
- Early Diagnosis of Alzheimer's—Making Use of the Blood-Brain Barrier
- Trials and Tribulations—Autopsy Reveals Pros and Cons of AD Vaccine
- Bill Klunk Reports from Paris on The Living Brain and Alzheimer’s Disease
- Amyloid Oligomer Antibody—One Size Fits All?
- One-Shot Deal? Mice Regain Memory Day After Vaccination, Plaques Stay Put
- Dodart JC, Bales KR, Paul SM. Immunotherapy for Alzheimer's disease: will vaccination work?. Trends Mol Med. 2003 Mar;9(3):85-7. PubMed.
- Following Footsteps of AD Vaccination: Passive Shots Against Prions Protect Mice
- Pertussis Toxin Stokes Autoimmune Reaction in Aβ-Vaccinated Mice
- Mini-strokes from Passive Immunization?
- Plaque Clearance, Antibody Isotype Are Key for Passive Aβ Immunization
- Stockholm: Core Resistance
- Autoimmunity: Good or Bad for Neuroprotection?
- Conference Coverage: IPSEN Foundation
- Trials and Tribulations—Autopsy Reveals Pros and Cons of AD Vaccine
- The Alzheimer's Vaccination Story, Continued
- Hock C, Konietzko U, Streffer JR, Tracy J, Signorell A, Müller-Tillmanns B, Lemke U, Henke K, Moritz E, Garcia E, Wollmer MA, Umbricht D, de Quervain DJ, Hofmann M, Maddalena A, Papassotiropoulos A, Nitsch RM. Antibodies against beta-amyloid slow cognitive decline in Alzheimer's disease. Neuron. 2003 May 22;38(4):547-54. PubMed.
- Winblad B, Blum KI. Hints of a therapeutic vaccine for Alzheimer's?. Neuron. 2003 May 22;38(4):517-8. PubMed.