Aβ vaccination effects on plaque pathology in the absence of encephalitis in Alzheimer disease

Patients and methods.

The patient was a 71-year-old man with a 10-year history of moderate to severe cognitive impairment (Mini-Mental State Examination score of 12 1 month prior to death) consistent with AD according to National Institute on Aging/Consortium to Establish a Registry for Alzheimer’s Disease criteria with a Braak stage of VI and an APOE genotype of E3E4. The patient was enrolled in a double blind randomized multicenter clinical trial and received a total of three injections (225 μg/each) of AN-1792 (Elan Pharmaceuticals, San Francisco, CA). The first immunization was given in October 2001, followed by two other injections 4 and 12 weeks afterwards with no noticeable adverse effects. The patient did not receive any anti-inflammatory therapy. Maximum serum anti-Aβ titers reached were 1:2,771 at 6 months after the first injection of AN-1792. The patient died 1 year later with the cause of death as “failure to thrive.” The brain was removed within 6 hours of death and weighed 1,100 g. The right hemibrain was rapidly frozen for biochemical analysis while the left hemibrain was placed in formalin for neuropathologic and immunocytochemical analysis. For Braak staging and analysis of plaque and tangle density in 10 random fields, thioflavin-S stained sections were imaged by fluorescence microscopy as previously described.8 Further analysis of amyloid deposition, neuritic plaques, and tangle formation was performed in vibratome-cut sections immunostained with antibodies against Aβ (1:500, 3D6, Elan Pharmaceuticals) and phosphorylated tau (1:500, AT8, Innogenetics). Studies of Aβ colocalization within macrophages/microglia were performed by double labeling immunocytochemistry utilizing the Tyramide Signal Amplification system (NEN, Boston, MA) with antibodies against Aβ and CD68 (1:1,000, DakoCytomation, Carpinteria, CA) and imaged with the laser scanning confocal microscope. Leukocytes were evaluated with an antibody against CD45 (1:1,000, DakoCytomation) and astrocytes were detected by an anti-GFAP antibody (1:500, Sigma Chemicals, St. Louis, MO). Mononuclear cell phenotyping was performed in paraffin sections with antibodies against CD3 (T cell marker, 1:1,000, DakoCytomation) and CD20 (B cell marker, 1:1,000, DakoCytomation). Levels of Aβ-40/42 were assessed by ELISA as previously described.9 Neuropathologic findings were compared to those in nonimmunized AD cases (n = 10, Braak stages of V and VI) from the AD Research Center (ADRC) at UCSD.

Results.

The brain was grossly normal, with only mild atrophy of the temporal cortex and hippocampus (figure, A). The lateral ventricles were normal in size and the underlying white matter was normal (see the figure, A). The table summarizes the neuropathologic differences between the immunized case and the nonimmunized AD cases. Neuropathologic and immunocytochemical analysis in the AN-1792 treated case demonstrated an almost complete absence of extracellular amyloid deposits and neuritic plaques in the frontal cortex. Consistent with the apparent reduction in plaques, very low levels of Aβ-42 (14.57 ng/mL) were found by ELISA in the frontal cortex of the treated case when compared to nonimmunized AD cases. In this region levels of Aβ-40 were 31.32 ng/mL. High levels of Aβ were detected by ELISA in other brain regions in the immunized case such as the parietal cortex (Aβ-40 = 361.84 ng/mL, Aβ-42 = 960.15 ng/mL), the temporal cortex (Aβ-40 = 29.89 ng/mL, Aβ-42 = 226.47 ng/mL), and hippocampus (Aβ-40 = 14.96 ng/mL, Aβ-42 = 353.68 ng/mL). Aβ-immunoreactivity was only detected as aggregate structures within cells (figure, B). Further double labeling analysis showed that these aggregates were within CD68-immunoreactive macrophages/microglia (figure, C through E). Similar amyloid containing cells were observed in the hippocampus and temporal cortex, to a lesser extent. Such CD68-positive cells containing amyloid aggregates were not detected in nonimmunized AD cases. In the leptomeninges, minimal perivascular lymphocytic infiltration was observed, consisting of a mixture of T and B cells (figure, F and G). In the parietal cortex, abundant amyloid plaques were observed (figure, H). Macrophages/microglia were identified among these plaques (figure, I through K). These changes were accompanied by mild astrogliosis and microgliosis. No multinucleated giant cells were observed. Neurofibrillary tangles and neuropil threads were abundant throughout the neocortex and hippocampus, consistent with a Braak stage VI. Luxol fast blue analysis showed that the white matter was well myelinated and no macrophage infiltration was observed (figure, L and M).

• Download figure
• Open in new tab
• Download powerpoint

Figure. Neuropathologic and immunocytochemical analysis. (A) Gross appearance at the level of the anterior hippocampus. (B) Aβ immunoreactive cells in the frontal cortex. (C, D) Confocal microscopy image of CD68 (C) and Aβ (D) immunoreactive cells in the frontal cortex. (E) Merged image shows colocalization of Aβ in CD68 positive macrophages. (F) B cells (CD20 positive lymphocytes) around a vessel in the leptomeninges. (G) T cells (CD3 positive lymphocytes) in the leptomeninges. (H) Aβ immunoreactive plaques in the parietal cortex. (I) Confocal microscopy image of CD68 immunoreactive cells. (J) Aβ positive plaques in the parietal cortex. (K) Merged image of I and J. (L, M) Hematoxylin/eosin and Luxol fast blue staining of white matter.

Discussion.

This study shows that vaccination with Aβ-42 resulted in a considerable reduction of plaque burden and promoted amyloid phagocytosis in the frontal cortex and to a lesser extent in the temporal lobe. These Aβ immunoreactive macrophages/microglia are similar to those previously observed in Aβ-immunized APP tg mice2 and in two AN-1792 immunized human cases.6,7⇓ In addition, plaque-associated neuritic dystrophy in the frontal cortex was undetectable, suggesting that, as in animal models of AD, these neurites might recover a nonpathologic phenotype following plaque clearance. However, this case differs from previous autopsy case reports because only minimal lymphocytic infiltration was observed in the leptomeninges, the white matter was unaffected, and the most prominent effect was in the frontal rather than in the parietal cortex. Consistent with two previous studies, we observed neurofibrillary pathology and CAA remaining after treatment.6,7⇓ The comparative selectivity for the Aβ immunization effects in the frontal, and to a lesser extent, the temporal lobe in the present case is unclear. Possible mechanisms include the degree of amyloid compaction, region specific post-translational modification of Aβ, and relative proportion of Aβ species in different brain regions. In nonimmunized AD patients, Aβ-40 is the predominant species in CAA and Aβ-42 is more abundant in the plaques. Since in the vaccinated patients only amyloid in the plaques was cleared but not in the vessels, the vaccination approach might favor the clearance of Aβ-42 over Aβ-40. This is consistent with studies in immunized APP tg mice where the significant reduction in the levels of Aβ-42 rather than Aβ-4010 was accompanied by a reduction in plaque formation but no effect on CAA was observed and these mice exhibited an increase in CAA-associated hemorrhages.10

Another potential explanation for the difference in findings might be related to the number and dose of immunizations received. In this respect, while in our case the patient received three injections over a span of 12 weeks, in other case reports the patients received five injections over a span of 36 weeks6 and two injections over a span of 1 month.7 Although demise in all three cases occurred approximately 1 year after the first injection, the patient in the first report was treated with a dose of 50 μg per injection6 compared to 225 μg injections in our case and in the second report.7