Abstract
Background:
Capillary amyloid-β (capAβ) deposition in the cerebral cortex is the neuropathological feature providing the basis for categorizing cerebral amyloid angiopathy (CAA) into two distinct types, CAA-Type1 with capAβ and CAA-Type2 without capAβ.
Objective:
We investigated the neuropathological and clinical characteristics of capAβ deposition in a prospective population-based study.
Methods:
Vantaa 85+ includes 601 individuals aged ≥85 years, of which 300 were studied clinically and neuropathologically. 278 subjects were analyzed for the apolipoprotein E (APOE) genotype. The diagnosis of capAβ was determined using immunohistochemistry against Aβ, and of CAA using Congo red confirmed by Aβ immunohistochemistry, both analyzed in six brain areas. The severity of capAβ was graded semi-quantitatively, and the severity of CAA was based on the percentage of affected vessels. Alzheimer’s disease (AD)-type neuropathology (CERAD score and Braak stage) was analyzed using standard methods.
Results:
CAA-Type1 was present in 86/300, CAA-Type2 in 135/300, and 79/300 had no CAA. CapAβ was most frequent in the occipital lobe (79/86). CAA-Type1 was associated with the severity of CAA (p < 0.001), dementia (p < 0.001), severe AD-type neuropathology (p-value 0.09 for CERAD C and 0.017 for Braak stages V-VI), and APOE ɛ4 allele carrier status (p < 0.001).
Conclusions:
This population-based study confirmed the presence of distinct CAA-Type1 and its association with the severity of CAA, severe AD-type neuropathology, and the APOE ɛ4 carrier status. Both the severity and localization of the deposition seemed to determine the clinical significance of capAβ.
INTRODUCTION
Cerebral intra cortical amyloid β-protein (Aβ) deposition is assumed to be essential in the pathogenesis of Alzheimer’s disease (AD) [1]. Aβ is produced by proteolytic cleavage of the amyloid-β protein precursor (AβPP) by β- and γ-secretases, and accumulates in the brain parenchyma as senile plaques and in the walls of meningeal and cortical small and middle-sized blood vessels as cerebral amyloid angiopathy (CAA) [2]. The prevalence of CAA increases with age [1, 3] and severe CAA occurs more frequently with AD [4, 5]. Population-based studies have shown higher prevalence of CAA in the demented than in non-demented individuals [6]. A correlation, although relatively low, between the clinical diagnosis of dementia and the presence and severity of CAA has been found in hospital-basedstudies [7].
In addition to accumulating in small and middle-sized blood vessels to create CAA, Aβ may accumulate in the capillary basement membranes as small bumps, referred to as capillary Aβ (capAβ) [8, 9]. Similarly to CAA, capAβ mainly consists of Aβ42 and Aβ40 [10]. Like the CAA affecting the small and middle-sized blood vessels, capAβ has been observed in most parts of the brain [11, 12]. Aβ deposition at vessel walls has been thought to interfere the metabolism of brain tissue, leading to degenerative changes by two mechanisms: (1) blocking the perivascular lymphatic drainage pathway [13] and thereby impairing the clearance of Aβ at the blood-brain barrier (BBB) [14], or (2) impairing the transendothelial clearance ofAβ [15].
Some hospital-based studies have found that the presence of capAβ is associated with the severity of CAA [7, 16] and consequently, capAβ has been regarded as the end stage of the most severe forms of CAA [17]. However, in other studies, correlation between the severity of CAA and occurrence of capAβ [9] has been low, leading to an alternative view, i.e., that capAβ and CAA would represent distinct entities. This view is supported by a recent finding that capAβ can occur in the context of relatively mild CAA in non-selected material [16] and in ADpatients [18].
It has been suggested that CAA occurs in two morphologically distinct types: CAA-Type1 with capAβ and CAA-Type2 without capAβ [9]. Hospital-based studies have observed CAA-Type1 to be more frequently associated with (1) clinical AD [10], (2) presence of severe AD type pathology [7, 22], and (3) AD-related genetic risk factor apolipoprotein E ɛ4 allele (APOE ɛ4) [9, 16]. Subjects with CAA-Type2 have been noticed to have higher APOE ɛ2 frequencies compared to CAA-Type1 or without CAA [9] but other clinical characteristics for Caa-Type2 have not been identified.
To our knowledge, no population-based studies have been published reporting the frequency of different CAA-types. Here we report the frequency and severity of CAA-Type1 and its relation to AD-type neuropathology, dementia, and APOE genotype in a prospective population based cohort of very elderly subjects (Vantaa 85+).
MATERIALS AND METHODS
Subjects
Vantaa 85+ Study includes 601 individuals, aged at least 85 years, living in the city of Vantaa on April 1, 1991. Autopsy with neuropathological examination was performed on 300 subjects (mean age 92.4±SD 3.7 years, ranges 85–105). The baseline clinical examination was performed in 1991 with follow-up studies in 1994, 1996, 1999, and 2001 when possible. The clinical diagnosis of dementia and hypertension was performed as previously described [19–21]. Cholesterol and triglyceride levels were analyzed using standard biochemical methods [21]. APOE genotyping was performed on DNA extracted from peripheral blood cells from 278 subjects [21]. The clinical characteristics of the whole study population (n = 552) and of the neuropathologically examined subpopulation (n = 300) are shown in Supplementary Table 1. The populations were similar except for the slightly higher frequency of demented subjects in the neuropathologically examined subpopulation.
Neuropathological examination
AD- and Lewy body-related pathologies
The brains were fixed in 4% formaldehyde for at least two weeks and embedded in paraffin. AD-related neuropathology (Braak stages and CERAD scores) was evaluated as previously described [19, 20] based upon the guidelines originally published by Braak and Braak [22] and Mirra et al. [23]. In this study we compared CERAD neuritic plaque scores “no” (0), “sparse” (A), and “moderate” (B) against the score “frequent” (C) and Braak (neurofibrillary tangle pathology) stages 0-IV versus stages V and VI (extensive cortical neurofibrillary pathology). Lewy-related pathology was determined as previously described [24]. Here the subjects having neocortical Lewy-related pathology were compared with controls.
Cerebral amyloid angiopathy
The prevalence and severity of CAA in the middle-sized and large meningeal and cortical blood vessels were assessed on 6μm thick paraffin sections stained with Aβ antibodies (clone 4G8, residues 17–24) in the six brain regions (frontal, parietal, temporal, occipital cortex, hippocampus, and cerebellum). For the immunohistochemistry, the sections were deparaffinized and pre-treated with 0.5% H2O2 for 30 min and then 100% formic acid for 5 min, followed by incubation with a primary antibody overnight (Mouse anti-beta amyloid clone 4G8, residues 17–24). The immuno reactivity was detected using the avidin-biotinylated HRP complex (ABC) system (Vector Lab, CA, USA). The findings were confirmed by histological Congo red staining in eight μm-thick formalin-fixed paraffin-embedded tissue sections analyzed under polarized light. The severity of CAA was determined [25] by counting the percentage of the Congo red and Aβ immunoreactive positive small and middle-sized vessel profiles of all vessel profiles seen in each section separately for each region. The counts were combined to create the median value for all six regions.
Capillary Aβ
The presence of capAβ was analyzed in the same six brain regions as CAA by using immunohistochemistry as described above independently from clinical or other neuropathological data. In 16 samples (from nine subjects) with weak Aβ-staining result, the diagnosis was confirmed with Congo red. In the hippocampal area, the data includes the findings in both the hippocampus proper (Ammon’s horn) and subiculum. Presence of capAβ was determined as described previously [9, 10]: only clear and obvious bumpy globular capillary wall depositions were included, whereas pericapillary parenchymal Aβ deposition was excluded. The presence (yes/no) of the capillaries with Aβ deposition was analyzed on whole tissue slices using x400 magnification. The severity of capAβ was graded as previously described [7]; 0 = no affected capillary; 1 = less than one affected capillary per high power field (HPF); 2 = one to two affected capillaries per HPF; 3 = more than two affected capillaries per HPF. Grade 1 capAβ in 1HPF was defined as mild, grade 2-3 capAβ/ 1/HPF as severe. Multiple capAβ was defined as capAβ deposition in more than one brain region. Subjects with both severe (>1/HPF, grade 2-3) and multiple (with more than just one brain region) capillary Aβ deposition, are described here as severe-multiple-capAβ.
The term CAA-Type1 was defined as subjects having at least one Aβ positive capillary in any brain region [26] with or without CAA. Subjects with CAA-Type2 were defined as those having CAA without capAβ in any brain region investigated. Subjects without positivity neither in the Aβ immunohistochemistry nor Congo red in blood vessels of any size were defined as non-CAA.
Cerebral infarcts and hemorrhages
The presence of large (>15 mm) or small (5–15 mm) cortical infarctions was determined macroscopically at autopsy [21]. Cortical micro-infarcts (MI) were defined as a focal star-like lesions <2 mm with neuronal loss and cystic tissue necrosis with surrounding foamy macrophages and glial cell reaction, detected in the H&E stained tissue sections [24]. The presence of old micro hemorrhages (MH) was based on the microscopic analysis of Prussian blue staining [25]. The MI and MH were analyzed in the same six brain regions as described above.
Statistical analyses
The statistical analyses were performed using SPSS for Windows version 18 and 19 software. Comparison of dichotomous variables (gender, dementia, severe AD-type neuropathological variables or hypertension) and the distributions of the APOE genotype across the CAA-types were performed by Chi-square (χ2-test). Binary logistic regression analysis (univariate and multivariate) was used to estimate the association of the CAA-types with neuropathological variables controlling age and gender. Odds ratios (OR) were obtained with 95% confidence intervals (CI).
Approval for the study
The Vantaa 85+ study was approved by the Ethics Committee of the Health Centre of the City of Vantaa and by the Medical Ethics committee of the Helsinki University Central Hospital. The Finnish Health and Social Ministry approved the use of the health and social work records, and death certificates. Blood samples were collected only after the subjects or their relatives gave informed consent. The National Authority for Medico legal Affairs (VALVIRA) has approved the collection of the tissue samples at autopsy as well as their use for research. A written consent for autopsy was obtained from the nearest relatives.
RESULTS
Frequency and topography of capillary Aβ and cerebral amyloid angiopathy
Capillary Aβ deposition (CAA-Type1) was found in 86/300 (28.6%) individuals. CapAβ was present in the layers II-V of the cerebral cortex, most frequent in the occipital cortex, hippocampus, and in the temporal lobe (79/48/28 respectively, Fig. 1a). 26 subjects (30.2% of those with CAA-Type1) had capAβ deposition solely in the occipital region. 48 subjects had capAβ in the hippocampal region (Ammon’s horn and subiculum), of which ten had capAβ in the CA4-CA1 regions of the Ammon’s horn. The frequency of capAβ in different brain regions was similar compared with the frequency of CAA in the same brain regions (Fig. 1b).
Approximately one third (32/86) of the CAA-Type1 cases had severe capAβ. That was most frequently found in the occipital and temporal lobes and hippocampi (Fig. 1a). 14/86 (16%) of those with CAA-Type1 had severe-multiple-capAβ (severe, >1/HPF, grade 2-3 capAβ deposition) in multiple brain regions.
Comparison between subjects with CAA-Type1 and 2 and those without CAA
CAA was observed in 221 (74%) of all 300 autopsied subjects. 86 of those had CAA-Type1 and 135 CAA-Type2. 79 (26%) did not show CAA (Table 1).
Clinical variables
There were slightly more men in the CAA-Type1 group and more females in the non-CAA group compared to the other groups (χ2-tests, p = 0.004, Table 1). 195 (65%) of all 300 neuropathologically examined subjects had dementia. Of these, a higher proportion belonged to the CAA-Type1 group than to the other groups (p < 0.001 for the association between dementia and CAA-Type1, χ2-test, Table 1, Supplementary Table 1).
Alzheimer-type pathology and apolipoprotein E genotypes
Severe AD-type pathology was more frequent in the subjects with CAA-Type1 than in subjects in the other groups (Table 1, Fig. 2). CAA-Type1 was associated with severe AD-type neuropathology in multivariate logistic regression analysis (OR 2.5, OR 3.6; Table 2). There was no association between Lewy-related pathology and CAA-Type1 (Table 1). The severity of CAA was significantly higher in the subjects with CAA-Type1 when compared with those having CAA-Type2 (median 7.58% versus 1.33% , p < 0.001, Table 1). The severity of CAA was even higher in those with severe-multiple-capAβ (median value 15.5% , 9.04–47.71 IQR) and in those with capAβ in the hippocampus proper (17.25% 7.92–47.58 IQR). The occipital lobe was the most common brain region to show capAβ and its severity was also highest in that location (Fig. 1b).
All ten individuals with capAβ (CAA-Type1) in the hippocampus proper were demented. In addition, 6/7 of those with even mild (grade 1) capAβ-deposition in the subiculum as the only brain manifestation of CAA-Type1 had dementia. All subjects (14/14) with severe-multiple-capAβ had dementia, CERAD score C, and a higher median value of general CAA (15.5% versus 7.58% in CAA-Type1). There was no association between severe-multiple-capAβ and severe neurofibrillary pathology.
APOE genotyping was performed in 278 of the 300 autopsied subjects. The frequency of the different APOE ɛ4 alleles in the three CAA groups is shown in Fig. 3. Two individuals were homozygous for the ɛ4 (0.7%), both representing CAA-Type1. 62.9% of those with CAA-Type1 and 26.9% with CAA-Type2, but only 5.3% of the non-CAA controls, carried at least one APOE ɛ4 allele (p < 0.001 for the association between APOE ɛ4 an CAA-Type1, χ2-tests) (Table 1). Possession of APOE ɛ2 allele was not associated with the CAA types, not even if the demented and non-demented subjects were analyzed separately.
Cerebral infarcts and hemorrhages
The presence of neither the MI nor MH associated with any of the CAA-Types (Table 2). The presence of any cortical infarction (large = >15 mm, small = 5–15 mm, and MI = <2 mm) showed a marginal association with the CAA-Type1 in the univariate (p = 0.03) but not in the multivariate analysis (p = 0.086) to CAA-Type1 (Table 2).
DISCUSSION
In this first population-based study we confirmed the morphological division of CAA into two distinct types, and the association between CAA-Type1 and clinical dementia, AD type neuropathology, and carrier status of the APOE ɛ4 allele. We also present the prevalence and severity for the CAA types 1 and 2 and show the clinical significance of hippocampal area capAβ.
Topography
Previously, the presence of capAβ has been observed in all neocortical areas, hippocampus and cerebellum [11]. To our knowledge, this is the first population-based study reporting the topography of capAβ systematically in several brain regions (frontal, parietal, temporal, occipital cortex, hippocampus, and cerebellum). In the previous hospital-based series, variable cortical and hippocampal areas have been used in the measurements of capillary Aβ deposition [7, 26].
We confirmed the previous finding [11] that capillary Aβ deposition follows the topography of CAA, with the occipital region being the most frequently affected. Interestingly, some studies have found different occipital capAβ-scores between CAA-Types [18]. In our cohort, one third of CAA-Type1 had capAβ deposition only in the occipital area. We have also previously observed [25] that the severity of CAA varies from one brain area to another. Remarkably, this tendency was found here to be even more obvious for capAβ (Fig. 1b). Interestingly, even mild capAβ deposition in the hippocampus proper or subiculum seemed to indicate dementia. Our findings are in accord with the important role of the hippocampus in AD pathology.
Frequency and severity of cap Aβ deposition
Here, capAβ was mostly mild, as had been observed before in a hospital-based material [27]. Previously, some studies have excluded mild capAβ findings because of the common occurrence of capAβ in an aged population [16]. However, as explained above, even mild capAβ in a certain location (hippocampus) may be clinically significant, although, naturally not only the localization but also the severity of capAβ influences the clinical outcome (dementia, AD-type dementia).
Demographic and clinical variables across the different CAA-types
Most previous hospital-based series have not highlighted any effect of gender on CAA [9, 10]. However, some studies including ours [18, 25] have found a tendency of under-representation of women in CAA groups, suggesting perhaps higher earlier mortality in women with CAA and capAβ vessel wall deposition or alternatively more efficient transfer of Aβ across the endothelial lining in women. In contrast, we did not find any association between gender and severe AD-type pathology. Similar to our results (Table 1), previous studies [9] found no age differences between individuals with CAA-Type1 andCAA-Type2.
CAA-Type1 was associated with clinical dementia, even more strongly than in some previous studies [11], perhaps because of the higher mean age of our study participants when compared with other studies (92.4±3.7 versus 84.3±9.3 years) [7].
Vascular lesions
CAA has been shown to associate with MI [28]. We did not find any association between CAA-Types and single or multiple MI nor MH in any brain region, in contrast to what has been described previously [9, 26]. It is possible that within this very elderly unselected population, which probably contains various cardiovascular and cerebrovascular pathologies, the additive effect of CAA in the risk of cortical infarcts or hemorrhages is too small to be detected even in 300 brains. Presumably, an ischemic lesion caused by CAA is in general much larger than small MI or MH next to cortical capillary. As for ischemic lesions, another explanation may relate to the variety of backgrounds for these changes, e.g., atherosclerosis, of larger arteries and might explain the lack of association to the CAA-type.
Capillary Aβ deposition, CAA, and AD
It is known that capAβ occurs only with CAA [9, 16]. We confirmed that capAβ deposition was exclusively found in subjects having CAA in the medium-sized and large cortical and leptomeningeal vessels. Furthermore, we confirmed the significant association between capAβ and severe AD-type pathology [7, 27]. Interestingly, the presence of both severe and multiple capAβ in the brains, and the presence of capAβ in the hippocampal region seemed to associated to AD-type dementia, possibly due to the impairment of cerebral blood flow in this region due to the accumulation of Aβ in the capillary walls [29].
The present findings suggest that the presence of both severe and multiple capAβ taking into account hippocampal capAβ might be a useful neuropathological tool for the diagnosis of AD-type dementia. It is of interest that a severe CERAD score was more frequent in the subjects with capAβ deposition than the presence of severe neurofibrillary tangle (Braak stages 5-6) pathology. This may support the novel “two-hit” theory on the development of AD [25], which suggests that Aβ accumulation in the BBB and brain hypoperfusion occur first, followed by the neurofibrillary tangles in the development of AD pathology. Thus, this theory emphasizes the role of capillaries and capillary metabolic events in the development of AD type neuropathology and its clinical consequences.
Comparing CAA-Type1 and CAA-Type2, we did not find any association between CAA-Type2 and the severity of CAA, severe AD-type neuropathology, or APOE ɛ4 allele carrier status. Our results did not support the previous finding showing an association between CAA-Type2 and possession of APOE ɛ2 allele [9].
As might be expected, we found an association between CAA-Type1 and the carrier status of at least one ɛ4 allele, emphasizing the significance of the APOE ɛ4 in AD pathology. The proportion of the ɛ4 allele carriers was even slightly higher among subjects with CAA-Type1 (62.9%) compared with previous studies in hospital-based series (46.7% and 54%) [9, 16], presumably due to the higher mean age and higher number of demented subjects in our cohort.
CONCLUSION
CapAβ deposition in the brains of very elderly people associates with AD-type genetic and neuropathological findings, but not with Lewy-related pathology, cortical infarcts, or hemorrhages, thus favoring the view that capAβ is strongly related with AD-type dementia.
Footnotes
ACKNOWLEDGMENTS
Professor Raimo Sulkava and Dr. Leena Niinistö are acknowledged for their essential contributions in the Vantaa 85+ study. We thank Tuija Järvinen and Merja Haukka for skillful technical assistance, M.Sc. Antti Nevanlinna for statistical advisement, and PhD Christopher Carroll for language revision. Drs. Minna Oinas, Sari Rastas, and Olli Tynninen are acknowledged for their contributions in collection of the clinical or neuropathological data.
