21 Mar 2003

In an advanced online publication in this week's Nature Medicine, doctors report the first autopsy of a patient enrolled in trials for the ill-fated Alzheimer's disease (AD) vaccine (see ARF related news story; also ARF news story).

James Nicoll, Roy Weller, and colleagues from the University of Southampton, and Moorgreen and Southampton General hospitals, carried out the autopsy on a woman who had received a series of five injections of Elan's AN-1792 Aβ42 peptide vaccine. The patient had no apparent side effects after the first four injections, but six weeks after the fifth, 42 weeks after treatment started, she suffered dizzy spells, drowsiness, unstable gait, and fever. Two weeks later doctors took an MRI scan, which suggested inflammation in the brain, and started her on the antiinflammatory dexamethasone. Her condition remained relatively unchanged until she died about 10 months later.

Remarkably, when Nicoll and colleagues examined the brain of this patient, they found that there were very few amyloid plaques in her neocortex. The temporal lobe, for example, contained almost no plaques; this is in stark contrast to the 100 or so plaques/mm² seen in seven nonimmunized AD control subjects. Parietal, frontal, and occipital lobes were similarly devoid of plaques, though Nicoll et al. did find substantial numbers of them in the cerebellum and basal ganglia, regions which usually become affected in later stages of the disease. "These data suggest an astonishingly powerful effect of the vaccination-clearance of Aβ from much of the cerebral cortex," write Steven Greenberg, Brian Backsai, and Bradley Hyman from Massachusetts General Hospital in an accompanying News & Views article. That’s the good news.

On the downside, the autopsy revealed that other pathological features, such as neurofibrillary tangles, were unaffected by the vaccination. This appears to dash prior hopes that reducing plaques would lead to an ensuing reduction in neurofibrillary pathology. Moreover, the accumulation of Aβ in blood vessels, called cerebral amyloid angiopathy, or CAA (see recent live discussion), was also unchanged. Confounding this, lymphocytes-which normally do not pass the blood-brain barrier-had infiltrated the membrane that covers the brain, and this invasion of the meninges was most pronounced in areas where CAA was the worst. More than a dozen other patients who volunteered for the vaccine also showed symptoms of meningoencephalitis, or inflammation of the meninges, and it was this fact that led to termination of the clinical trial. Greenberg and colleagues suggest that the vaccination may have triggered an inflammatory response against amyloid in the blood vessels, resulting in abnormal cerebral blood flow that was responsible for clinical decline.

Where to go from here? The failure to clear neurofibrillary tangles, which are tightly associated with cognitive impairment, suggests that Aβ-specific therapy may not remedy much of the damage that already exists, according to Greenberg and colleagues. But recent advances in neuroimaging of Aβ raise the possibility of treating those likely to develop the disease before cognitive decline begins, they write. Another approach may be to use vaccines that do not elicit a cellular immune response, and so prevent the deleterious inflammatory side effects. The latter successfully clears Aβ in mouse models of AD (Bacskai et al., 2002).—Tom Fagan