Cerebrovascular pathology in Down syndrome and Alzheimer disease

People with Down syndrome (DS) are at higher risk for developing Alzheimer disease (AD), which is thought to be primarily due to the overexpression of amyloid precursor protein [19, 46]. Beta-amyloid (Aβ) plaques and neurofibrillary tangles are typically observed by 40 years of age (reviewed in [27], with dementia onset most typically occurring almost a decade later [26, 35, 36, 52]. Up to 55% of people with DS between 40 and 49 years of age develop dementia and the numbers rise to 77% in people 60–69 years of age (reviewed in [6]).

There is increasing recognition of the vascular contribution to cognitive impairment and dementia (VCID; [33, 37]). The presence of cerebrovascular pathology may be a critical comorbidity that accelerates the age of onset of dementia and also leads to faster disease progression. In the general population, ~6%–45% of autopsy cases show mixed AD neuropathology with cerebrovascular pathology [32]. Cerebrovascular pathology resulting from atherosclerosis and arteriolosclerosis may serve as a “second hit” necessary for conversion to dementia particularly when significant Aβ is present in the brain [33, 47]. Atherosclerosis is thought to be a major contributor to VCID as is arteriolosclerosis [1, 29, 33]. Cerebral amyloid angiopathy (CAA) is also observed frequently in AD [3, 15, 25, 38] and may lead to microhemorrhages and infarcts [57].

VCID in DS has been less well studied [60] but is thought to be rare based upon fewer vascular risk factors being present in DS (e.g. hypertension, atherosclerosis, smoking) [45]. In a recent neuroimaging study using T2*and susceptibility weighted imaging, the location and number of microbleeds were evaluated in 91 nondemented and 26 individuals with DS [14]. CAA in people with DS was present in 31% of symptomatic DS participants, which was similar to that observed in sporadic AD (38%). Microbleeds observed by neuroimaging may be larger than the smaller bleeds that can be observed at autopsy and may underrepresent the extent of this vascular pathology in DS. At autopsy, CAA has been reported in small autopsy studies of people with DS over the age of 55 years [30, 35] or in case reports [8, 18, 40] with CAA also containing post-translationally modified Aβ [23]. However, the accumulation of CAA as a function of age in DS has yet to be explored. Extensive cerebrovascular hemorrhages and stroke are associated with CAA in DS [8, 18, 31, 39, 40, 43] in most studies but not in all [30, 35]; the majority of these studies are based on small autopsy series or case reports. Despite increased CAA in DS with age, VCID (or multi-infarct dementia as it was termed then [16]) is rare with only one case report in the literature of a 55-year old woman with DS.

It is important to note that people with DS show virtually no evidence for several of the risk factors for cerebrovascular pathology particularly atherosclerosis and hypertension [12, 20, 21, 41, 42, 61]. Thus, the purpose of our study was two-fold. First, we sought to characterize the frequency of cerebrovascular pathology (defined as atherosclerosis, arteriolosclerosis and CAA) in a series of autopsy cases with DS and AD with direct comparison to sporadic AD and nondemented controls. Second, we hypothesized that cerebrovascular pathology associated with atherosclerosis would be less frequent in DS relative to non-DS AD cases. It is interesting to note that these vascular pathologies have not been directly compared in DS and AD cases.

This study examined prevalence of cerebrovascular pathology in three University of California at Irvine Alzheimer Disease Research Center (UCI ADRC) cohorts including DS cases, sporadic AD cases and nondemented control cases. All AD cases, sporadic AD and DS with AD (Braak neurofibrillary tangle (NFT) stage VI and amyloid plaque stage C), were selected based upon neuropathological criteria collected using the recently revised National Alzheimer Disease Coordinating Center (NACC) guidelines [28]. DS cases were all demented at the time of death based upon DSM-IV criteria [56] Atherosclerosis was evaluated using procedures described by Beach and colleagues (2007) [7] with the Circle of Willis being graded by gross visual inspection and rated as none, mild (0–25% obstruction), moderate (26–60% obstruction) or severe (> 60% obstruction). Arteriolosclerosis was determined by H&E staining of the thalamus and subthalamic nuclei, basal ganglia, middle frontal gyrus, superior & middle temporal gyrus, inferior parietal lobule, and occipital cortex [28]. Arteriolosclerosis, as described by Grinberg and colleagues (2010) included concentric hyaline thickening of small arteries (40–150 μm in diameter) associated with a concentric stenosis of the vessel lumen [25]. Arteriolosclerosis was noted as being mild, moderate or severe based upon the severity of hyalinosis of the media and adventitia of small parenchymal and/or leptomeningeal vessels. CAA was scored based upon Aβ immunostaining as none, mild, moderate or severe. Nondemented control cases were selected to be Braak NFT stage 0, I or II and amyloid plaque stage 0 or A [11]. Data collected using the NACC vascular pathology forms at the UCI ADRC describing atherosclerosis, arteriolosclerosis and amyloid angiopathy (or CAA) were used [28]. A total of 149 cases were evaluated, including 32 from the DS cohort, 80 from the AD cohort, and 37 control cases. Demographic characteristics are shown by cohort in Table 1. Similar numbers of males and females were observed in the DS and AD cohorts, but more males were represented in the control cohort. The average age at death was lowest in the DS cohort, and highest in the control cohort (Table 1).

Observations had been recorded using the NACC data forms [28] by a neuropathologist (RK). Atherosclerosis and arteriolosclerosis were frequently observed in the AD cohort and less so in the DS and control cohorts (Table 2, Fig 1). In contrast, CAA was more common in the DS cohort (Fig. 1). The relative risks of each of these vascular outcomes are provided in Table 3 and suggest that the DS cohort was 1.21 times more likely to have CAA relative to AD cases, and 4.6 times more likely to have CAA compared to control cases.

To test the hypothesis that the DS cohort had a higher frequency of more severe CAA and a lower frequency of severe atherosclerosis or arteriolosclerosis, we used a cumulative logistic regression analysis. Based on a χ² test with 2 degrees of freedom, there was a significant association of CAA with the type of autopsy cohort (Likelihood ratio statistic 47.96, p < .0001). Cumulative logistic regression may be used to compare the odds of a higher-versus-lower severity finding between cohorts based on the odds ratio. According to Table 4, the odds of severe CAA in the DS cohort was approximately three times the odds of a higher severity finding in the AD cohort (odds ratio = 3.05 (95% CI 1.4, 6.67), p = 0.005). Arteriolosclerosis, in contrast, was absent in the DS cohort, and there were no significant differences between the AD and control cohorts. Atherosclerosis was significantly less severe in the DS cases relative to the AD (odds ratio = 0.23, (95% CI 0.09, 0.59), p = 0.0026) and control cases (odds ratio = 0.11, (95% CI 0.04, 0.30), p < .0001).

We hypothesized that people with DS would have significantly more frequent and more severe CAA associated with overexpression of APP relative to sporadic AD and control cases but less cerebrovascular pathology typically associated with cardiovascular risk factors including atherosclerotic lesions and arteriolosclerosis. In the current study, as expected, we found: (1) the presence of CAA in DS was more frequent than in the AD and control cohorts; (2) that the DS cohort showed an increased probability of more severe CAA with age; and (3) atherosclerosis and arteriolosclerosis vascular pathology was uncommon in DS.