Persistent neuropathological effects 14 years following amyloid-β immunization in Alzheimer's disease

Abstract

We performed a 15-year post-mortem neuropathological follow-up of patients in the first trial of amyloid-β immunotherapy for Alzheimer's disease. Twenty-two participants of a clinical trial of active amyloid-β42 immunization (AN1792, Elan Pharmaceuticals) or placebo were studied. Comprehensive post-mortem neuropathological assessments were performed from 4 months to 15 years after the trial. We analysed the relationships between the topographical distribution of amyloid-β removal from the cerebral cortex and tau pathology, cerebrovascular territories, plasma anti-AN1792 antibody titres and late cognitive status. Seventeen of 22 (77%) participants had Alzheimer's neuropathological change, whereas 5 of 22 (23%) had alternative causes for dementia (progressive supranuclear palsy = 1, Lewy body disease = 1, vascular brain injury = 1, and frontotemporal lobar degeneration = 2). Nineteen of the 22 participants had received the active agent, three the placebo. Fourteen of 16 (88%) patients with Alzheimer's disease receiving the active agent had evidence of plaque removal (very extensive removal = 5, intermediate = 4, very limited = 5, no removal = 2). Of particular note, two Alzheimer's patients who died 14 years after immunization had only very sparse or no detectable plaques in all regions examined. There was a significant inverse correlation between post-vaccination peripheral blood anti-AN1792 antibody titres and post-mortem plaque scores (ρ = - 0.664, P = 0.005). Cortical foci cleared of plaques contained less tau than did cortex with remaining plaques, but the overall distribution of tangles was extensive (Braak V/VI). In conclusion, patients with Alzheimer's disease actively immunized against amyloid-β can remain virtually plaque-free for 14 years. The extent of plaque removal is related to the immune response. This long duration of efficacy is important in support of active immunization protocols as therapy for, or potentially prevention of, neurodegeneration-associated protein accumulations. Inclusion of patients without Alzheimer's disease in Alzheimer's therapy trials is a problem for assessing the efficacy of treatment. Despite modification of Alzheimer's pathology, most patients had progressed to severe dementia, notably including the five with very extensive plaque removal, possibly due to continued tau propagation. Neuropathology follow-up of patients in therapeutic trials provides valuable information on the causes of dementia and effects of treatment.

Keywords: Alzheimer’s disease; amyloid-β; dementia; immunotherapy; neuropathology.

Figure 1

Illustrations of the morphological features of amyloid-β immunohistochemistry associated with plaque clearance. (A) Plaque-free areas of cerebral cortex containing scattered ‘naked’ residual plaque cores, lacking surrounding diffuse amyloid-β. (B) A ‘moth-eaten’ appearance of some remaining plaques. (C) A granular pattern of intracytoplasmic amyloid-β within microglia (arrows) with some ‘moth-eaten’ extracellular amyloid-β (right of image). (D) Triple label confocal image demonstrating microglia clustered around a remaining plaque (amyloid-β, cyan; HLA-DR, green; CD68, red; overlap, yellow). (E) A small cortical area lacking plaques, with remaining plaques on either side, associated with full-thickness, full-circumference amyloid-β staining of overlying leptomeningeal arteries indicating severe cerebral amyloid angiopathy. (F) Numerous capillaries with amyloid-β staining, indicating capillary angiopathy, in a plaque-free area of cerebral cortex. (A–C, E and F) Immunohistochemistry for pan-amyloid-β. (D) Triple label confocal microscopy for amyloid-β, HLA-DR and CD68. Scale bars = 100 μm in A–C and F; 50 μm in D; 1 mm in E.

Figure 2

Scans of coronal histological sections of cerebral hemisphere from participants in the AN1792 trial immunolabelled to demonstrate amyloid-β (appearing black). The cases are grouped according to the extent of positive histological evidence of plaque removal on examination of all brain regions (Method 1). (A) Placebo with neuropathologically confirmed Alzheimer’s disease demonstrating a continuous band of plaques throughout the cerebral neocortex (Case 15). (B) Alzheimer’s subjects with nearly complete plaque removal (Cases 7, 8, 16, 20 and 21). (C) Alzheimer’s subjects with intermediate plaque removal (Cases 1, 2, 6 and 9). (D) Alzheimer’s subjects with limited plaque removal (Cases 4, 10, 11, 17 and 22). (E) Alzheimer’s subjects with no evidence of plaque removal (Cases 3 and 19). (F) Non-Alzheimer’s subjects (Cases 5, 12, 13, 14 and 18). Amyloid-β plaque density was scored throughout the neocortex of each hemisphere section using a CERAD-adapted method (Method 2) with the results colour-coded for illustrative purposes as: frequent plaques = red, moderate = yellow, sparse = green, none = blue, expressed as % of each category.

Figure 3

Negative association between the mean anti-AN1792 antibody response during the trial period and post-mortem amyloid-β plaque score indicating that the magnitude of the immune response to AN1792 influenced the degree of amyloid-β removal from the brain..

Figure 4

The effect of amyloid-β removal was quantified for each of the different histological features of intraneuronal tau accumulation (tangles, dystrophic neurite clusters and neuropil threads) by comparing cortical regions with amyloid-β plaques remaining (Aβ+) to regions with amyloid-β plaques removed (Aβ−). There were significantly fewer tangles and dystrophic neurite clusters in the regions where amyloid-β plaques had been removed in response to the amyloid-β immunotherapy; not significant for neuropil threads. Scale bar = 50 μm.

Figure 5

Illustrations of the single case of post-immunization meningoencephalitis (Case 1). Numerous lymphocytes in the subarachnoid space clustered around arteries showing the ‘double-barrel’ morphology associated with severe cerebral amyloid angiopathy (A) haematoxylin and eosin; and (B) CD3 immunohistochemistry for T lymphocytes. (C) Lymphocytes were relatively sparse in the cerebral cortex. (D) Cerebral white matter showed focal rarefaction of myelinated fibres (Kluver Barrera). (E) Old cortical microbleed indicated by the presence of a cluster of haemosiderin pigment granules. (F) Severe cerebral amyloid angiopathy with full thickness, full circumference involvement of the walls of leptomeningeal and parenchymal arteries by amyloid-β (immunohistochemistry for pan-amyloid-β). Scale bars = 100 μm in A and E; 50 μm in B and D; 30 μm in C; and 500 μm in F.