Clinical Manifestations of Cerebral Amyloid Angiopathy–Related Inflammation Jessica A. Eng, BA,1 Matthew P. Frosch, MD, PhD,2 Kyungchan Choi, MD,1 G. William Rebeck, PhD,1 and Steven M. Greenberg, MD, PhD1 To explore the clinical effects of inflammation associated with vascular deposits of the amyloid ␤ peptide (A␤), we analyzed 42 consecutive patients with pathologically diagnosed cerebral amyloid angiopathy (CAA) for evidence of an inflammatory response. Inflammation with giant-cell reaction surrounding amyloid-laden vessels was identified in 7 of the 42 cases. The clinical symptoms in each of the seven were subacute cognitive decline or seizure rather than hemorrhagic stroke, the primary clinical presentation in 33 of 35 patients with noninflammatory CAA (p < 0.001). Inflammatory CAA also was associated with radiographic white matter abnormalities, significantly younger age at presentation, and a marked overrepresentation of the apolipoprotein E 4/4 genotype (71% vs 4%, p < 0.001). Of the six inflammatory CAA patients with available follow-up information, five demonstrated clinical and radiographic improvement after immunosuppressive treatment. The syndrome of CAA-related perivascular inflammation appears to represent a subset of CAA with clinically distinct symptoms that may respond to immunosuppressive treatment. These data add to evidence that inflammation against A␤ can cause vascular dysfunction, a potential mechanism for the toxic response recently observed in clinical trials of A␤ immunization. Ann Neurol 2004;55:250 –256 The inflammatory response to deposits of the ␤-amyloid peptide (A␤) in the brain has been postulated to have both beneficial and deleterious consequences. Pathological1 and epidemiological2,3 studies of Alzheimer’s disease (AD) initially suggested that inflammation might contribute to disease pathogenesis. More recent findings of reduced A␤-containing plaque deposits in transgenic mice immunized with A␤4 or treated with anti–A␤ antibodies5,6 have instead suggested that an immune response to A␤ might decrease AD pathology. The possibility that anti-A␤ inflammation may be beneficial motivated a clinical trial of vaccination to A␤7 which subsequently was discontinued when a subset of patients developed a meningoencephalitis syndrome.8 In contrast with the spate of studies on inflammation and A␤ deposition in the brain parenchyma, relatively little information is available on the effect of the inflammatory response on vascular A␤ deposits or cerebral amyloid angiopathy (CAA). Both sporadic9,10 and some familial forms11–13 of CAA appear to associate with a modest increase in perivascular inflammatory cells. Individual cases have been reported of a more robust inflammatory response to CAA, often involving the appearance of multinucleated giant cells.9,14 –21 From the 1Department of Neurology and 2C. S. Kubik Laboratory of Neuropathology, Massachusetts General Hospital and Harvard Medical School, Boston, MA. Received Apr 25, 2003, and in revised form Jul 13 and Sep 15. Accepted for publication Sep 15, 2003. 250 Several of these case descriptions have suggested a severe clinical course associated with the CAA-related inflammation, although the clinical picture has not been characterized in detail. Immune-based treatments studied in transgenic mice have not been found to reduce the severity of CAA5,22 and in one model were associated with increased CAA-related intracerebral hemorrhage.22 In light of increasing interest in the benefits and possible risks of anti–A␤ immunotherapy, we analyzed the clinical, radiographic, pathological, and genetic findings of patients with and without CAA-related inflammation in a series of consecutive patients with pathologically diagnosed CAA. We sought to determine in particular whether the presence of inflammation was associated with specific clinical symptoms or risk factors. Our findings suggest that the inflammatory response to CAA can cause clinically significant vascular dysfunction. Subjects and Methods Identification of Subjects We reviewed all patients seen at the Massachusetts General Hospital (MGH) between July 1, 1994 and November 30, 2002 who were diagnosed pathologically with CAA as the Address correspondence to Dr Greenberg, Wang ACC 836, Massachusetts General Hospital, Boston, MA 02114. E-mail email@example.com © 2004 American Neurological Association Published by Wiley-Liss, Inc., through Wiley Subscription Services cause of their clinical syndrome. Forty-two eligible patients were identified from our prospective cohort of CAA patients23; systematic search of a database maintained by the MGH neuropathology laboratory yielded no additional cases. Brain samples were obtained by postmortem brain examination in 12 of the cases, hematoma evacuation in 19, and cortical biopsy in 11. The diagnostic brain samples were obtained at MGH for 37 of the 42 patients and from outside referring hospitals for 5. To identify specimens with evidence of CAA-related inflammation, a neuropathologist (M.P.F.) examined all available sections from these cases, stained by conventional methods (hematoxylin and eosin or Congo red), in random order and without knowledge of clinical information. Specimens were specifically examined by conventional staining methods for the presence of inflammatory cells associated with amyloid deposits, AD pathology, and white matter lesions. Data Collection and Analysis Hospital records and radiographic images from CAA patients with and without inflammation were reviewed for demographic information, vascular risk factors, clinical presentation, laboratory results, and radiographic findings. Thirtythree of the 42 patients (including all 7 with CAA-related inflammation) donated blood samples for determination of apolipoprotein E (APOE) genotype, performed as described.24 Clinical course was followed by telephone interviews with patients or caregivers at 6-month intervals.23 The study was performed in accordance with the guidelines of the institutional review board of MGH and with consent of subjects or family members. Comparison of patients with and without perivascular inflammation was performed by Student’s t test for continuous variables and by Fisher’s exact test for categorical variables. All statistical analyses were performed with Stata software (College Park, TX) and all significance tests were two-sided. Immunohistochemical Staining The following monoclonal antibodies (and concentrations) were used to immunostain sections prepared from paraffin blocks from patients with CAA-related inflammation: anti–A␤ antibody 6E10 (1:200; Signet Laboratories, Dedham, MA), CD68 (1:50; DAKO, Carpinteria, CA), CD4 (1:20; Novocastra, Newcastle, UK), CD3 (1:100; Novocastra, Newcastle, UK), CD8 (1:50; DAKO, Carpinteria, CA), and CD20 (1:200; DAKO, Carpinteria, CA). Incubations were at 4°C overnight except for CD4 immunostaining, which was performed at room temperature for 1 hour. Antibody staining was visualized using biotinylated goat anti– mouse IgG (1:200) and commercial ABC reagent (Vectastain-Rl Vector Laboratories, Burlingame, CA) according to the manufacturer’s recommendations Results Of 42 patients with pathologically diagnosed CAA seen over an 9-year period, a subset of 7 (17%) could be distinguished by pathology findings of perivascular inflammation with multinucleated giant cells. These seven patients demonstrated a characteristic set of clinical and laboratory findings (Table 1). The presenting symptom was cognitive or functional decline progressing rapidly over a period of months in three of the seven, new onset of seizures in one patient, and both progressive cognitive decline and seizure-like episodes in the remaining three patients. Headaches were prominent in four patients. The salient features on diagnostic testing were patchy or confluent white matter hyperintensities on T2-weighted magnetic resonance imaging (MRI) sequences (seven of seven patients; Fig 1), multiple microhemorrhages on gradient-echo sequences (six of seven), and mild to moderate elevations of cerebrospinal fluid (CSF) protein (four of five patients with CSF examination). Neuroimaging otherwise showed no evidence of vasculitis in the four patients who underwent angiography (one by MR and three by contrast angiography) and no definite areas of restricted diffusion in the six patients with diffusionweighted imaging. All seven patients received brief or prolonged treatment with corticosteroids or cyclophosphamide (Table 1). Five of six patients with follow-up information improved clinically, although none improved to premorbid functional baseline. Four of five with follow-up MRI scans also demonstrated clear improvement in extent of white matter changes (see Fig 1). Patient 2 did not show substantial improvement either clinically or radiographically. Two patients (including one otherwise lost to follow-up) died 2.7 and 3.2 years after presentation, whereas the remaining five have survived through a mean follow-up period of 22.4 ⫾ 18.9 months. Table 2 compares the seven cases of CAA-related perivascular inflammation with the other 35 cases of pathologically confirmed CAA without inflammation. The two groups differed significantly in presentation, with lobar hemorrhagic stroke as the presenting feature in 0 of 7 inflammatory cases compared with 33 of 35 noninflammatory CAA patients. Patients with CAArelated inflammation were also significantly younger at presentation by approximately 7 years. Genetic analysis showed a prominent association between perivascular inflammation and the APOE ε4 allele, particularly in its homozygous form. Five of the seven patients (71%) with CAA-related perivascular inflammation carried the APOE ε4/ε4 genotype compared with only one ε4/ε4 genotype among the 26 patients (4%) with noninflammatory CAA and available genetic data (odds ratio, 62.5; 95% confidence interval, 3.5–3062; p ⬍ 0.001). Pathological examination of the brain samples from these patients demonstrated moderate to severely advanced CAA17 together with accumulations of mononuclear and multinucleated white blood cells around amyloid-laden vessel in cerebral cortex and leptomeninges (Fig 2A–C). All vessel segments with surrounding inflammatory changes stained positive for A␤ (see Fig 2F). Inflammation within the vessel wall itself and Eng et al: CAA-Related Inflammation 251 Table 1. Clinical, Laboratory, and Radiographic Characteristics of Patients with CAA-Related Perivascular Inflammation MRI at Presentation Patient No./ Age/Sex White Matter Hyperintensity Microhemorrhages CSF WBC (/ml) Protein (mg/dl) Decreased CS ⫻ 5 days Increased 55 CP ⫻ 10 days CS ⫻ 6 days NA 164a CS ⫻ 16 days Decreased CS ⫻ 1 yr Decreased CP ⫻ 1 yr CS ⫻ 1 mo Decreased CP ⫻ 6 weeks NA 1 mo cognitive decline Patchy None 2/63/M 3 mo cognitive decline Patchy Multiple 25a 211a 3/69/F 3 weeks gait difficulty, few months cognitive decline 1 yr cognitive decline, sudden onset confusion 2 seizures over 1 mo Patchy Multiple 0 Confluent Multiple 1 Confluent Multiple Confluent Multiple 1 Confluent Multiple 2 5/71/F 6/73/M 7/79/M 3 seizures over 2 months, confusion, and personality changes over few months 4 months cognitive decline, seizure Immunosuppressive Treatment Follow-up MRI: White Matter Hyperintensities CS ⫻ 3 days 1/49/F 4/71/F a Clinical Symptoms at Presentation NA NA 50a 84* Clinical Course Improved, not to baseline Independent Minimal improvement Dependent Died 2.7 years after presentation Improved, not to baseline Recurrent seizure Independent Improved, not to baseline ⫾ independent Improved, not to baseline Recurrent seizures Died 3.2 yr after presentation Improved, not to baseline, independent Above upper limit of normal at laboratory where CSF obtained. CAA ⫽ cerebral amyloid angiopathy; MRI ⫽ magnetic resonance imaging; CSF ⫽ cerebrospinal fluid; WBC ⫽ white blood cell; CS ⫽ corticosteroids; CP ⫽ cyclophosphamide; NA ⫽ not available. granuloma formation were not observed. Immunostaining for CD68 (not shown) demonstrated many of the cells (including the multinucleated giant cells) to be of the monocyte/microglial line. Multinucleated cells could be seen in close association with A␤, suggesting that the cells may have internalized the amyloid deposit. T lymphocytes (staining for CD3) were also present in the perivascular infiltrate, including low numbers of both CD8-positive and CD4-positive T lymphocytes, whereas B lymphocytes, assessed by staining for CD20, were absent. Examination of the brain parenchyma demonstrated multiple microhemorrhages and small nonhemorrhagic infarctions in the cerebral cortex and rarefaction of the underlying white matter (see Fig 2D, E). Dense-cored plaques were present in cortical samples from two of the seven patients. One subject (Patient 6) had brain specimens available from two time points: a biopsy at presentation and an autopsy 3.2 years after diagnosis and immunosuppressive treatment. Comparison of the two specimens demonstrated persistence of severe CAA but resolution of most or all of the perivascular inflammatory changes noted at presentation. Discussion These data add a syndrome of perivascular inflammation, subacute cognitive decline, seizures, and leukoen- 252 Annals of Neurology Vol 55 No 2 February 2004 cephalopathy to the spectrum of presentations associated with CAA.25 The clinical manifestations of this syndrome potentially have important implications for the diagnosis and treatment of CAA. The apparent response of some patients to immunosuppressive therapy suggests that it may represent a treatable form of CAA, highlighting the importance of reaching this diagnosis in practice. CAA-related inflammation also may have a more general role in affecting function of amyloidcontaining vessels, a possibility suggested by the modest increases in perivascular monocytes observed even in apparently noninflammatory CAA.9 Radiographic white matter changes are common in CAA26; the possibility that CAA-related inflammation plays a part in this finding offers potential directions for future treatments. The mechanism for the cognitive decline, seizures, and white matter changes associated with this syndrome remains to be determined but appears likely to reflect immune-mediated effects on vascular function. CAA without inflammation has been reported to associate with impaired cognition,27 cerebral infarctions,28,29 and white matter lesions,26,30 but not with the aggressive clinical and radiographic picture observed in the current patients, suggesting that the inflammation may play a direct role. The localization of the inflammatory response to amyloid-laden vessel seg- Fig 1. Magnetic resonance imaging (MRI) appearance in patients with cerebral amyloid angiopathy (CAA)–related inflammation. The gradient-echo images from Patients 4 and 5 show multiple small hypointense lesions characteristic of CAA-related microhemorrhages. The fluid-attenuated inversion recovery (FLAIR) images performed at presentation and after follow-up intervals of 2 months (Patient 4) or 22 months (Patient 5) demonstrate confluent regions of white matter hyperintensity that largely resolve at follow-up. ments and the appearance of A␤ in close association with some of the inflammatory cells9,14,19,31 implicate A␤ as the likely trigger for the immune response. An inflammatory response to vascular amyloid also has been noted in a transgenic mouse model that develops prominent CAA.32 The currently described syndrome of perivascular inflammation appears distinct from granulomatous angiitis of the central nervous system,33 which characteristically includes both granuloma formation and inflammation and necrosis within the vessel wall itself, features absent from the current cases. True granulomatous angiitis has been reported in the setting of CAA,14,21 although both most of the cases in the literature and our own experience suggest that this is a less common response to CAA than the type of perivascular inflammatory response described here. Table 2. Comparison of Clinical and Genetic Characteristics of Patients with Severe CAA with or without Perivascular Inflammatory Changes Characteristic Age at presentation (yr ⫾ SD) Sex (M/F) Primary clinical presentation, n (%) Intracerebral hemorrhage Cognitive decline Seizure APOE genotype, n (%)b ε4/ε4 ε4/ε(2 or 3) ε(2 or 3)/ε(2 or 3) a CAA with Inflammation (n ⫽ 7) CAA without Inflammation (n ⫽ 35) 68.3 ⫾ 9.6a 3/4 75.8 ⫾ 8.3 10/25 0 (0)b 3 (43) 4 (57) n ⫽ 7 patients genotyped 5 (71)b 1 (14) 1 (14) 33 (94) 1 (3) 1 (3) n ⫽ 26 patients genotyped 1 (4) 12 (46) 13 (50) p ⫽ 0.04. p ⬍ 0.001 for comparison of distributions of clinical presentation or APOE genotype. b CAA ⫽ cerebral amyloid angiopathy; APOE ⫽ apolipoprotein E. Eng et al: CAA-Related Inflammation 253 Fig 2. Pathological findings in cases of cerebral amyloid angiopathy (CAA) with inflammatory changes. (A) Multifocal inflammatory perivascular infiltrate is present involving leptomeningeal and parenchymal vessels, in association with amyloid deposition (Congo Red stain; scale bar ⫽ 100m). (B) Involved parenchymal blood vessel with inflammatory cell infiltrate present but without evidence of associated damage to the vessel (H&E; scale bar ⫽ 25m). (C) Inflammatory perivascular infiltrate including a multinucleated giant cell (arrows; H&E; scale bar ⫽25um). (D) Small ischemic lesion in subcortical white matter is evident as an area of tissue rarefaction and vacuolation (H&E stain; scale bar ⫽ 100m). (E) Evidence of a prior microhemorrhage demonstrated by hemosiderin in the wall of an involved blood vessel (arrowheads) and in the adjacent parenchyma (arrows) (H&E stain; scale bar ⫽ 50m). (F) Immunostaining for A␤ demonstrates the spatial relationship between the amyloid deposit and the inflammatory cell infiltrate (primary antibody, 6E10; scale bar ⫽ 25m). The patients with CAA-related perivascular inflammation demonstrated a striking overrepresentation of the APOE ε4 allele and ε4/ε4 genotype. Although based on only a few cases, these data raise the possibility that the E4 isoform of apolipoprotein E may play a specific role in promoting an inflammatory response to CAA. Previous studies have shown that APOE ε4 is associated with increased fibrillar A␤ deposition24,34 –36 (with substantially increased vascular amyloid24,34,37) and that the apoE protein38 (including residues specific to the E4 isoform39) is present in the deposits themselves. Other studies have suggested that apolipoprotein E4 may enhance activation of complement or microglia,40,41 offering potential mechanisms for the association observed in the current cases. Regardless of the mechanism, this association if confirmed might offer a potentially useful tool for supporting the diagnosis of this syndrome. 254 Annals of Neurology Vol 55 No 2 February 2004 The major limitation of our study was its restriction to cases of CAA with available neuropathology samples. This requirement limited the overall number of patients who could be studied, making it particularly difficult to form conclusions about response to therapy and optimal treatment for this syndrome. The restriction to cases with pathology samples also introduces potential biases toward patients with atypical clinical presentations, who may be more likely to undergo biopsy or postmortem evaluation. The figure of 17% (7 of 42) for the frequency of CAA-related inflammation observed in our series therefore may represent an overestimate of its true frequency. Conversely, the patchy nature of the inflammatory response raises the possibility that this pathology might be missed in a small biopsy specimen, thereby causing an underestimate in its frequency. Although further studies clearly will be required to determine the full extent of this syndrome, the current data are sufficient to suggest that it is not rare. We also note that although the surgical specimens derived from hematoma evacuation were not obtained specifically for the evaluation of vessels, the samples of leptomeningeal vessels and fragments of superficial cortical vessels present in the specimens were adequate to evaluate for the perivascular inflammatory changes that characterized the novel subset. The clinical effects of CAA-related inflammation may have important implications for immune-based strategies for the treatment of the cerebral amyloidoses, particularly AD. A clinical trial of A␤ vaccination for AD was discontinued when same subjects (ultimately 18 of 298 treated with active vaccine) developed a syndrome described as meningoencephalitis.8 A recent report of the first autopsy of a patient with this syndrome42 described rapidly progressive cognitive decline, prominent white matter abnormalities on MRI, and the pathological appearance of both advanced CAA and inflammatory cells (primarily macrophages and T lymphocytes) surrounding some amyloid-laden vessels of the leptomeninges and cerebral cortex. The clinical and pathological features encountered in this patient show intriguing similarities to the currently described patients, pointing to the important possibility that the inflammatory response to CAA might account partially for the vaccine-associated meningoencephalitis syndrome. Determining the mechanism for this toxic response to the A␤ vaccine would be a major advance for immune-based treatment strategies,43 particularly if it ultimately allows a subgroup of patients at high risk for this complication to be identified before treatment. This research was supported by the National Institutes of Health (R01 AG021084, S.M.G.) and Massachusetts Alzheimer’s Disease Research Center Neuropathology Core (P50 AG05134). We are grateful to Dr B. T. Hyman for reviewing the manuscript and to Dr J.-P. G. Vonsattel for many very helpful discussions. References 1. McGeer PL, McGeer EG. The inflammatory response system of brain: implications for therapy of Alzheimer and other neurodegenerative diseases. Brain Res Brain Res Rev 1995;21: 195–218. 2. In’t Veld BA, Ruitenberg A, Hofman A, et al. Nonsteroidal antiinflammatory drugs and the risk of Alzheimer’s disease. N Engl J Med 2001;345:1515–1521. 3. Stewart WF, Kawas C, Corrada M, Metter EJ. Risk of Alzheimer’s disease and duration of NSAID use. Neurology 1997;48: 626 – 632. 4. Schenk D, Barbour R, Dunn W, et al. Immunization with amyloid-beta attenuates Alzheimer-disease-like pathology in the PDAPP mouse. Nature 1999;400:173–177. 5. Bacskai BJ, Kajdasz ST, Christie RH, et al. Imaging of amyloid-beta deposits in brains of living mice permits direct observation of clearance of plaques with immunotherapy. Nat Med 2001;7:369 –372. 6. McLaurin J, Cecal R, Kierstead ME, et al. Therapeutically effective antibodies against amyloid-beta peptide target amyloidbeta residues 4-10 and inhibit cytotoxicity and fibrillogenesis. Nat Med 2002;8:1263–1269. 7. Schenk D. Opinion: amyloid-beta immunotherapy for Alzheimer’s disease: the end of the beginning. Nat Rev Neurosci 2002;3:824 – 828. 8. Orgogozo JM, Gilman S, Dartigues JF, et al. Subacute meningoencephalitis in a subset of patients with AD after Abeta42 immunization. Neurology 2003;61:46 –54. 9. Yamada M, Itoh Y, Shintaku M, et al. Immune reactions associated with cerebral amyloid angiopathy. Stroke 1996;27: 1155–1162. 10. Uchihara T, Akiyama H, Kondo H, Ikeda K. Activated microglial cells are colocalized with perivascular deposits of amyloidbeta protein in Alzheimer’s disease brain. Stroke 1997;28: 1948 –1950. 11. Maat-Schieman ML, van Duinen SG, Rozemuller AJ, et al. Association of vascular amyloid beta and cells of the mononuclear phagocyte system in hereditary cerebral hemorrhage with amyloidosis (Dutch) and Alzheimer disease. J Neuropathol Exp Neurol 1997;56:273–284. 12. Vinters HV, Natte R, Maat-Schieman ML, et al. Secondary microvascular degeneration in amyloid angiopathy of patients with hereditary cerebral hemorrhage with amyloidosis, Dutch type (HCHWA-D). Acta Neuropathol 1998;95:235–244. 13. Shin Y, Cho HS, Rebeck GW, Greenberg SM. Vascular changes in Iowa-type hereditary cerebral amyloid angiopathy. Ann NY Acad Sci 2002;977:245–251. 14. Probst A, Ulrich J. Amyloid angiopathy combined with granulomatous angiitis of the central nervous system: report on two patients. Clin Neuropathol 1985;4:250 –259. 15. Le Coz P, Mikol J, Ferrand J, et al. Granulomatous angiitis and cerebral amyloid angiopathy presenting as a mass lesion. Neuropathol Appl Neurobiol 1991;17:149 –155. 16. Powers JM, Stein BM, Torres RA. Sporadic cerebral amyloid angiopathy with giant cell reaction. Acta Neuropathol 1990;81: 95–98. 17. Vonsattel JP, Myers RH, Hedley-Whyte ET, et al. Cerebral amyloid angiopathy without and with cerebral hemorrhages: a comparative histological study. Ann Neurol 1991;30:637– 649. 18. Mandybur TI, Balko G. Cerebral amyloid angiopathy with granulomatous angiitis ameliorated by steroid-cytoxan treatment. Clin Neuropharm 1992;15:241–247. 19. Anders KH, Wang ZZ, Kornfeld M, et al. Giant cell arteritis in association with cerebral amyloid angiopathy: immunohistochemical and molecular studies. Hum Pathol 1997;28: 1237–1246. 20. Fountain NB, Eberhard DA. Primary angiitis of the central nervous system associated with cerebral amyloid angiopathy: report of two cases and review of the literature. Neurology 1996;46: 190 –197. 21. Shintaku M, Osawa K, Toki J, et al. A case of granulomatous angiitis of the central nervous system associated with amyloid angiopathy. Acta Neuropathol 1986;70:340 –342. 22. Pfeifer M, Boncristiano S, Bondolfi L, et al. Cerebral hemorrhage after passive anti-Abeta immunotherapy. Science 2002; 298:1379. 23. O’Donnell HC, Rosand J, Knudsen KA, et al. Apolipoprotein E genotype and the risk of recurrent lobar intracerebral hemorrhage. N Engl J Med 2000;342:240 –245. 24. Greenberg SM, Rebeck GW, Vonsattel JPV, et al. Apolipoprotein E e4 and cerebral hemorrhage associated with amyloid angiopathy. Ann Neurol 1995;38:254 –259. 25. Greenberg SM, Vonsattel JP, Stakes JW, et al. The clinical spectrum of cerebral amyloid angiopathy: presentations without lobar hemorrhage. Neurology 1993;43:2073–2079. Eng et al: CAA-Related Inflammation 255 26. Smith EE, Eng JA, Rosand J, Greenberg SM. Presence of leukoaraiosis predicts recurrent lobar hemorrhage. Stroke (abs) 2003;34:242. 27. Pfeifer LA, White LR, Ross GW, et al. Cerebral amyloid angiopathy and cognitive function: the HAAS autopsy study. Neurology 2002;58:1629 –1634.. 28. Olichney JM, Hansen LA, Hofstetter CR, et al. Cerebral infarction in Alzheimer’s disease is associated with severe amyloid angiopathy and hypertension. Arch Neurol 1995;52:702–708. 29. Cadavid D, Mena H, Koeller K, Frommelt RA. Cerebral beta amyloid angiopathy is a risk factor for cerebral ischemic infarction. A case control study in human brain biopsies. J Neuropathol Exp Neurol 2000;59:768 –773. 30. Haglund M, Englund E. Cerebral amyloid angiopathy, white matter lesions and Alzheimer encephalopathy: a histopathological assessment. Dement Geriatr Cogn Disord 2002;14: 161–166. 31. Gray F, Vinters HV, Le Noan H, et al. Cerebral amyloid angiopathy and granulomatous angiitis: immunohistochemical study using antibodies to the Alzheimer A4 peptide. Hum Pathol 1990;21:1290 –1293. 32. Winkler DT, Bondolfi L, Herzig MC, et al. Spontaneous hemorrhagic stroke in a mouse model of cerebral amyloid angiopathy. J Neurosci 2001;21:1619 –1627. 33. Kolodny EH, Rebeiz JJ, Caviness VSJ, Richardson EPJ. Granulomatous angiitis of the central nervous system. Arch Neurol 1968;19:510 –524. 34. Schmechel DE, Saunders AM, Strittmatter WJ, et al. Increased amyloid beta-peptide deposition in cerebral cortex as a consequence of apolipoprotein E genotype in late-onset Alzheimer disease. Proc Natl Acad Sci USA 1993;90:9649 –9653. 256 Annals of Neurology Vol 55 No 2 February 2004 35. Rebeck GW, Reiter JS, Strickland DK, Hyman BT. Apolipoprotein E in sporadic Alzheimer’s disease: allelic variation and receptor interactions. Neuron 1993;11:575–580. 36. Holtzman DM, Bales KR, Tenkova T, et al. Apolipoprotein E isoform-dependent amyloid deposition and neuritic degeneration in a mouse model of Alzheimer’s disease. Proc Natl Acad Sci USA 2000;97:2892–2897. 37. Premkumar DR, Cohen DL, Hedera P, et al. Apolipoprotein E-epsilon4 alleles in cerebral amyloid angiopathy and cerebrovascular pathology associated with Alzheimer’s disease. Am J Pathol 1996;148:2083–2095. 38. Wisniewski T, Frangione B. Apolipoprotein E: a pathological chaperone protein in patients with cerebral and systemic amyloid. Neurosci Lett 1992;135:235–238. 39. Cho HS, Hyman BT, Greenberg SM, Rebeck GW. Quantitation of apoE domains in Alzheimer disease brain suggests a role for apoE in Abeta aggregation. J Neuropathol Exp. Neurol 2001;60:342–349.. 40. Barger SW, Harmon AD. Microglial activation by Alzheimer amyloid precursor protein and modulation by apolipoprotein E. Nature 1997;388:878 – 881. 41. McGeer PL, Walker DG, Pitas RE, et al. Apolipoprotein E4 (ApoE4) but not ApoE3 or ApoE2 potentiates beta-amyloid protein activation of complement in vitro. Brain Res 1997;749: 135–138. 42. Nicoll JA, Wilkinson D, Holmes C, et al. Neuropathology of human Alzheimer disease after immunization with amyloid-beta peptide: a case report. Nat Med 2003;9:448 – 452. 43. Greenberg SM, Bacskai BJ, Hyman BT. Alzheimer disease’s double-edged vaccine. Nat Med 2003;9:389 –390.