Subacute meningoencephalitis in a subset of patients with AD after Abeta42 immunization
Orgogozo JM, Gilman S, Dartigues JF, Laurent B, Puel M, Kirby LC, Jouanny P, et al.;
Commented by , 22 Aug 2003
Aim of the study
To investigate the safety of active immunization against Abeta42 in patients with AD in a Phase II study.
Method
372 patients with mild to moderate AD age 50 to 85, MMSE 15 to 26, were randomized to receive intramuscular injections of 225 micrograms of aggregated Abeta42 (AN1792) with 50 micrograms of QS-21 as adjuvant or saline placebo (4:1) at baseline, months 1, 3, 6, 9 and 12.
Outcome measures included ADAS-cog and whole-brain volumetry using MRI. Serum samples were analyzed for anti-Abeta42 antibodies using ELISA. CSF levels of anti-AN1792, tau and Abeta were to be done at baseline and end of the trial.
Results
Dosing was terminated after four early reports of meningoencephalitis, but blinding of treatment allocation (except for serious adverse effects cases) and long term follow-up over two years is continuing.
A total of 18 of 298 (6%) of patients treated with AN1792 and none on placebo developed symptoms (predominantly confusion, headache, lethargy) and laboratory findings (mononuclear pleocytosis in CSF, 71 to 100% lymphocytes) consistent with meningoencephalitis; 16 patients had received two doses, 1 had received one dose, 1 had received three doses. The mean latency from the first and last injection to symptoms was 75 and 40 days. No case occurred later than six months after the first treatment.
Twelve patients have recovered within weeks, six remained with cognitive or neurologic sequelae. Anti-Abeta42 antibody titers were not correlated with the occurrence or severity of symptoms.
Discussion
There was high hope based on research performed in transgenic mice expressing the mutations implicated in dominantly inherited AD that active immunization against Abeta42 would remove beta-amyloid plaques and improve cognitive function.
A phase I study in 104 patients with AD had demonstrated good safety and tolerability of single or multiple injections of AN1792, with a detectable rise of Abeta42 antibodies in 25% of patients. It was thus considered appropriate to initiate a multicentre randomized study using state-of-the art outcomes to establish is disease progression over one year was modified by active immunization.
The meningoencephalitis associated with the active anti-Abeta immunization was unexpected, and resembles the delayed postvaccination meningoencephalitis after measles vaccinations and the postinfectious acute disseminated encephalomyelitis after measles.
The authors suggest that T-cell and microglial activation may be responsible. Future immunotherapeutic strategies may thus exclude components of the Abeta molecule that provoke abnormal T-cell reactions. Passive immunization strategies with humanized anti-Abeta antibodies are currently being developed.
A note of caution against immunotherapy has been put forward by McGeer and McGeer (Neurobiology of Aging, 2003; 24: 391-395) since they have established that complement is already overactivated in the brain of patients with AD, and there is a risk that immunotherapy will increase production of the membrane attack complex adding to the autodestruction of neurons.
The excellent review on immunotherapy for AD by Dodel RC, Hampel H and Du Y (Lancet Neurology 2003; 2: 215-220) offer cautious optimism that it is still possible to test this therapeutic approach.