Article Text

Download PDFPDF

Contribution of cerebral amyloid angiopathy to Alzheimer’s disease
  1. S Love
  1. Correspondence to:
 Dr S Love
 Department of Neuropathology, Institute of Clinical Neurosciences, Frenchay Hospital, Bristol BS16 1LE, UK;

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Contribution of CAA to Alzheimer’s disease

In patients with Alzheimer’s disease (AD), focal and diffuse ischaemic abnormalities of the cerebral white matter can be demonstrated neuropathologically1–3 and neuroradiologically.4–8 The focal lesions have been shown to contribute to motor and neuropsychiatric manifestations of AD,9–12 and the more widespread or diffuse abnormalities to impaired cognition.1314 In some series, ischaemic cerebral lesions in AD have been more frequent in patients homozygous or heterozygous for the epsilon 4 (e4) allele of the apolipoprotein E gene (APOE),15–17 but other studies have found no such association,18–21 Studies of the relation between white matter disease in patients with AD or probable AD, and the systemic manifestations of arteriosclerotic vascular disease, have yielded inconsistent findings.11121–24

Several observations implicate cerebral amyloid angiopathy (CAA) as the probable cause of much of the white matter damage in AD. The vascular deposition of amyloid β protein (Aβ) is much more frequent and tends to be much more severe in patients with AD than in age-matched controls.25–29 Furthermore, CAA is a well documented risk factor for cerebral infarction1630–32 and for focal and diffuse white matter ischaemic lesions.33–36 The mechanisms whereby CAA may cause ischaemic damage to the white matter probably include a combination of luminal stenosis, endothelial damage, basement membrane thickening, thrombosis, loss of autoregulation, and vasospasm.37–40 Because evidence of the involvement of CAA in AD is largely based on post-mortem studies, which are by their nature skewed towards end stage disease, it could be argued that any contribution of CAA may be confined to the terminal stages of disease. If this were true, it might be expected that an inverse relationship between the severity of CAA at autopsy and the duration of AD would be found. That this is not the case28 suggests that CAA may exacerbate AD even at an early stage.

The occurrence of CAA in AD is strongly associated with possession of the e4 allele of APOE.17283241 Indeed, possession of the e4 allele of APOE is much more strongly correlated with vascular than parenchymal deposition of Aβ.28 This may reflect on the pathogenesis of neurodegeneration in AD in patients with e4. AD patients with severe CAA, almost all of whom possess at least one e4 allele, have significantly less parenchymal Aβ than do patients with lesser degrees of CAA.284243 This is further evidence against the argument that CAA is simply a late manifestation of AD, and raises the possibility that CAA and parenchymal Aβ have additive effects on the progression of clinical disease. Although the Aβ within vessel walls could, in theory, also cause local neurotoxicity in the cerebral cortex, we have found no evidence of any associated reduction in the density of immunostaining of synaptophysin,44 a sensitive marker of presynaptic integrity.

Observations of Weller and colleagues suggest that the involvement of CAA in AD may go beyond a contribution of ischaemia to the clinical and pathological manifestations of the disease, and that CAA may be involved in the development of plaques and tangles, the histological hallmarks of AD. Weller et al4546 propose that soluble Aβ is normally eliminated from the brain within interstitial fluid pathways that reach the cerebrospinal fluid in the subarachnoid space by passing along perivascular spaces in the cerebral cortex. If this model is correct, the obstruction to drainage of interstitial fluid caused by perivascular accumulation of amyloid could contribute to the accumulation of Aβ within the parenchyma and the development of the plaques and tangles. Support for this comes from the occasional finding of tau immunopositive neurites clustered around larger arteries with dyshoric amyloid angiopathy (angiopathy in which amyloid extends from the affected blood vessels into the surrounding brain parenchyma) (fig 1). Although the flow of interstitial fluid within the perivascular space occurs in the opposite direction to that of the arterial blood flow, it may be enhanced by the pulsatile arterial distension. A failure of this propulsive mechanism has been proposed to explain the association of capillary CAA with thrombosis of overlying cortical arteries.47

However, several other observations indicate that the relationship between CAA, plaques, and tangles is more complicated than would be predicted by a simple model of obstruction to drainage. These include the mutually exclusive topographical relationship between capillary CAA and extensive diffuse plaques,47 and the inverse correlation between overall severity of amyloid angiopathy and parenchymal amyloid load in patients with moderate to severe CAA.284345 Further evidence suggests that soluble Aβ within the brain is largely cleared by lipoprotein receptor related protein-1 mediated transcytosis across the endothelial cells of the blood–brain barrier.48 Impaired clearance of Aβ across the blood–brain barrier is probably central to the development of CAA in hereditary cerebrovascular amyloidosis with Dutch type haemorrhage49 caused by a G→C transition at codon 693 of the β amyloid precursor protein gene.

Studies by Wyss-Coray et al435051 identify transforming growth factor β (TGFβ) as a key influence on the relationship between parenchymal and vascular Aβ in AD. TGFβ1 levels are significantly increased in patients with AD, not only in the cerebral cortex,52 but also in the serum and CSF.5354 Chao et al53 observed a significant in vivo correlation between the level of TGFβ1 in the serum and the severity of dementia. On the face of it, this might seem paradoxical, as TGFβ1 has been shown to promote the clearance of Aβ from the parenchyma of transgenic mice expressing human β amyloid precursor protein.43 However, in contrast to the beneficial effects of TGFβ1 on clearance of parenchymal amyloid, expression of TGFβ1 by astrocytes in transgenic mice actually induces deposition of amyloid in cerebral blood vessels, this being accelerated by co-expression of human β amyloid precursor protein.5051 A parallel can be drawn between the latter finding and the detection of severe CAA in regions of brain with markedly reduced parenchymal Aβ in a patient with AD who was immunized with Aβ (peptide fragment AN-1792).55 The relevance of the observations of Wyss-Coray et al to AD was strengthened by the authors’ demonstration of a strong correlation between TGFβ1 mRNA levels and the severity of CAA in post-mortem brain tissue from 15 patients with AD and 7 controls.51 However, while these post-mortem findings are of interest, it should be noted that the number of cases studied was small.

Many questions remain as to the relation between CAA and AD. Apart from e4, the putative genetic risk factors for CAA show relatively little overlap with those for AD,5657 and despite the fact that e4 is a major risk factor for CAA in AD and in patients presenting with cerebral haemorrhage, CAA is probably not associated with the APOE genotype if these conditions are excluded.29 Although CAA is present in over 90% of patients with AD, it is not present in all cases and is therefore clearly not necessary for the development of the disease. Indeed, it is becoming increasingly clear that what we refer to as AD is really a spectrum of disorders with different genetic (and probably environmental) risk factors but having overlapping pathological and clinical phenotypes. For example, e4 associated AD tends to be a disease with moderate to severe CAA, AD caused by some presenilin mutations is characterized by cotton wool plaques and pyramidal tract degeneration,58–63 and AD in patients with an e2 APOE allele and CAA carries an increased risk of parenchymal brain haemorrhage.6465

The accurate diagnosis of CAA is likely to become increasingly important as we evaluate and implement treatments such as immunization, which are aimed at clearing parenchymal Aβ in AD, particularly if these carry a risk of promoting vascular deposition of Aβ.55 However, the ante-mortem diagnosis of CAA in AD remains a challenge. Measurement of plasma levels of Aβ and TGFβ was found to be unhelpful in predicting CAA.66 For the time being, except in the relatively few patients who manifest with lobar cerebral haemorrhage67 or have a brain biopsy,6869 we will have to continue to rely on examination of the brain at autopsy to make a confident diagnosis of CAA.

Figure 1

Tau immunopositive neurites clustered around a cortical artery with dyshoric amyloid angiopathy.

Contribution of CAA to Alzheimer’s disease