Amyloid β peptide 1-42 highly correlates with capillary cerebral amyloid angiopathy and Alzheimer disease pathology

Recent studies reported both positive [Thal et al. (2003) J Neuropathol Exp Neurol 62:1287-1301] and negative [Tian et al. (2003) Neurosci Lett 352:137-140] correlations between cerebral amyloid angiopathy (CAA) and Alzheimer's disease (AD) pathology. We have recently shown high correlations between neuritic AD pathology and amyloid beta peptide (Abeta) deposits in the capillary/pericapillary compartment (CapCAA) with only low correlations to general CAA (non-capillary). We have now studied the relationship between CapCAA and AD pathology with respect to the distribution of Abeta40 and 42 in the frontal cortex of 100 human postmortem brains from both male and female, demented and non-demented patients (mean age +/- SD 84.3+/-9.3 years). Using polyclonal antibodies to Abeta40 and 42, capillary and plaques positivity were assessed semiquantiatively on a four-point scale. Abeta42 deposits in capillaries correlated highly with both Abeta42 deposits in plaques and morphological AD criteria (CERAD, Braak stages, and NIA-Reagan-Institute criteria), while only a low correlation with CAA was observed. Abeta40 deposits in capillaries differed morphologically from Abeta42 ones: they were limited to capillary walls, were significantly less frequent in both capillaries and plaques compared to Abeta42 (P<0.01), and showed a low correlation with morphological AD criteria (P<0.05) and general CAA (P<0.01). By contrast, Abeta42 deposits were seen in the glia limitans rather than in capillary walls themselves, and showed high correlation with morphological AD criteria (P<0.01). These data indicate that CapCAA is characterized by Abeta42 deposits in pericapillary spaces or in the glia limitans. A low correlation between CAA and CapCAA, but high correlations between morphological AD criteria and CapCAA suggest different pathomechanisms for both types of CAA, and a close relation between CapCAA and AD pathology (both neuritic and plaque type). These data support the concept of a neuronal origin of Abeta via drainage from interstitial fluid from the central nervous system along basement membranes to capillaries.