Skip main navigation

White Matter Lesions in an Unselected Cohort of the Elderly

Molecular Pathology Suggests Origin From Chronic Hypoperfusion Injury
and on behalf of the
Originally publishedhttps://doi.org/10.1161/01.STR.0000221308.94473.14Stroke. 2006;37:1391–1398

Abstract

Background and Purpose— “Incidental” MRI white matter (WM) lesions, comprising periventricular lesions (PVLs) and deep subcortical lesions (DSCLs), are common in the aging brain. Direct evidence of ischemia associated with incidental WM lesions (WMLs) has been lacking, and their pathogenesis is unresolved.

Methods— A population-based, postmortem cohort (n=456) of donated brains was examined by MRI and pathology. In a subsample of the whole cohort, magnetic resonance images were used to sample and compare WMLs and nonlesional WM for molecular markers of hypoxic injury.

Results— PVL severity was associated with loss of ventricular ependyma (P=0.004). For DSCLs, there was arteriolar sclerosis compared with normal WM (vessel wall thickness and perivascular enlargement; both P<0.001). Capillary endothelial activation (ratio of intercellular adhesion molecule to basement membrane collagen IV; P<0.001) and microglial activation (CD68 expression; P=0.002) were elevated in WMLs. Immunoreactivity for hypoxia-inducible factors (HIFs) HIF1α and HIF2α was elevated in DSCLs (P=0.003 and P=0.005). Other hypoxia-regulated proteins were also increased in WMLs: matrix metalloproteinase-7 (PVLs P<0.001; DSCLs P=0.009) and the number of neuroglobin-positive cells (WMLs P=0.02) reaching statistical significance. The severity of congophilic amyloid angiopathy was associated with increased HIF1α expression in DSCLs (P=0.04).

Conclusion— The data support a hypoxic environment within MRI WMLs. Persistent HIF expression may result from failure of normal adaptive mechanisms. WM ischemia appears to be a common feature of the aging brain.

“Incidental” white matter (WM) lesions in the aging brain appear as hypodensities on computed tomography or as hyperintensities in T2-weighted and proton density MRI sequences, which are absent from T1-weighted images.1 WM lesions (WMLs) increase in frequency with age and are associated with lower cognitive performance.2,3 They resemble the severe lesions affecting the cerebral WM in Binswanger disease and are seen in Alzheimer disease (AD). Severity of WMLs ranges from ≥1 small isolated foci to extensive confluent areas. WMLs may not be detected in autopsy brain tissue from examination of fresh or fixed brain slices and are not systematically included in diagnostic pathology protocols for dementia.4

Histologic studies in Binswanger disease and AD show WM changes of myelin attenuation, axonal loss, oligodendrocyte loss, astrocytic gliosis, and arteriolar sclerosis,5 interpreted as evidence for an ischemic origin. In AD, other possibilities arise, including axonal depletion from Wallerian changes attributable to cortical neuronal attrition, or toxic effects of soluble amyloid on vascular permeability. WMLs in AD patients have been described as “incomplete infarction,” although this concept is not clearly defined.5–8 Congophilic amyloid angiopathy (CAA) is related to the presence of WMLs.9,10

The pathogenesis of “incidental” WMLs in elderly subjects is incompletely understood. Various theories have been proposed: degenerative changes of small vessels leading to chronic cerebral hypoperfusion, altered vascular permeability and blood–brain barrier dysfunction, and cerebrospinal fluid accumulation.11,12

We used MRI of fixed coronal brain slices to enable precise sampling of WMLs.13 In WM from brains donated to the population-based MRC Cognitive Function and Ageing Study (CFAS), we compared molecular markers of tissue ischemia and indices of vascular pathology in demented and nondemented individuals. The study was conducted in 2 phases. We initially assessed vascular morphology, ependymal integrity, vascular endothelial activation (intercellular adhesion molecule [ICAM]/collagen IV [Col IV] ratio), microglial activation (CD68), and molecular markers of hypoxia-inducible factors (HIFs) HIF1α and HIF2α. We found persistent HIF expression, indicating a hypoxic environment, and we therefore conducted further work to assess the expression of other markers that we (matrix metalloproteinase-7 [MMP7], neuromedin B receptor [NMBR], vascular endothelial growth factor receptor 2 [VEGFR2]) and others (neuroglobin [Ngb]) have shown to be hypoxia regulated.14,15 The findings support an ischemic contribution to the pathogenesis of these lesions, more marked in deep subcortical lesions (DSCLs) compared with periventricular lesions (PVLs).

Materials and Methods

Brain Tissue Donors

Respondents in CFAS donated the brain tissue used for the present study through a process of premortem liaison counseling approved by a research ethics committee. Dementia status was derived from longitudinal prospective collection of neuropsychological test data, including the mini mental state examination16 and Automated Geriatric Examination for Computer-Assisted Taxonomy organicity algorithm.17 Version 1 of the neuropathology data set, including 207 postmortem brain donations up to July 1998, was used for the assessment of ventricular morphology, vascular morphology, and the molecular markers HIF1α, HIF2α, ICAM1 versus Col IV, and CD68. HIF expression studies were performed in those donated brains with the shortest postmortem intervals because these markers are labile in paraffin processed tissues.14 Version 3.1 of the neuropathology data set includes 456 brain donations up to July 2004 and was used for the assessment of MMP7, Ngb, NMBR, and VEFGR2. All donated brains were characterized for neurodegenerative, vascular, and other pathologies.18 WMLs were identified by MRI of fixed coronal brain slices from 3 standardized levels (midfrontal, posterior frontal, and parietotemporal) using a method validated by direct comparison with histology.13 The number of cases and tissue blocks sampled for each phase of the study varied for the different markers (Table 1). The cohort included 48% demented versus 52% nondemented respondents; >77% were ≥80 years of age at death, and women were in a majority (males/females, 88/119).18 CFAS respondents are not selected on the basis of interaction with medical services, but the Version 1 data set is weighted toward respondents with cognitive decline and physical frailty.

TABLE 1. No. of Brains and Readings Used in Each Investigation

DementedNot Demented
Brains (Readings)Brains (Readings)
†Thirty-six individuals were not demented when last seen but had no information close to death. Two are included in the analysis of hypoxia-regulated proteins; they are classified in the table as nondemented but are excluded from the analysis of dementia status at death.
CFANS indicates Cognitive Function and Ageing Neuropathology Study.
Total (CFANS version 1)99108
Vascular morphology11 (201)8 (189)
Ependymal morphology
    PV42 (125)56 (104)
    DSC43 (130)55 (118)
ICAM/Col IV
    PV12 (16)26 (39)
    DSC14 (21)30 (56)
    Diffuse13 (16)18 (22)
Microglial activation, HIF1α and HIF2α
    PV2 (3)12 (16)
    DSC4 (5)10 (19)
Total (CFANS version 3.1)238182
Hypoxia-regulated proteins (MMP7, NMBR, VEGFR2)
    Control3 (5)9 (10)
    PV6 (7)5 (8)
    DSC5 (7)6 (8)

Magnetic Resonance Imaging

Fixed coronal brain slices, sealed in polythene, were presented for MRI in a custom-built Perspex stack. MRI was performed at 1.0 T (Siemens) using pulse sequences optimized for postmortem WMLs: T2-weighted spin echo (2500/98 ms; TR/TE excitations, where TR=repetition time and TE=echo time); proton density (2500/25 ms); and T1-weighted inversion recovery image (TR 6838 ms, inversion time of 600 ms, with a TE of 60 ms). A single MRI “cut” was obtained at the center of each brain slice using a 5-mm slice window with an in-plane resolution of 0.63×0.59 mm.

MRI scans were scored by 3 observers using a validated semiquantitative rating scale.19 Scores for periventricular WM (PVWM) and deep subcortical WM (DWM) were recorded separately.3 The severity of WMLs for a given case was summed from the scores for each slice.

Pathological Methods

Tissue blocks, representing WMLs and nonlesional WM, were selected using the MRI images to direct sampling. Each block was stained with hematoxylin and eosin and Luxol Fast Blue. Antibodies directed against CD68, Col IV, HIF1α, HIF2α, ICAM1, MMP7, Ngb, NMBR, and VEFGR2 were used (Table 2) using conventional immunocytochemical techniques. CAA in frontal and parietal regions overlying the WMLs studied was assessed with an antibody against βA4. Standard immunocytochemistry (ICC) methods using diaminobenzidine (DAB) as chromogen were used, with antigen retrieval as indicated in Table 2. Negative control staining omitted primary antibody, and antibody absorption experiments were undertaken when not reported previously. Positive control sections included tumors and inflammatory brain disease as appropriate. Dual labeling with immunofluoresence and conventional DAB was used to colocalize molecular markers of tissue hypoxia and markers of glial cell phenotype. Sections were first processed for molecular markers of hypoxia using the method above, followed by incubation with avidin-biotin blocking kit (Vector Laboratories). After incubation with the cell phenotype primary antibody, the sections were incubated with biotinylated anti-mouse IgG (for CD68) or anti-rabbit IgG (for glial fibrillary acidic protein) and visualized with streptavidin tetramethylrhodamine B isothiocyanate (Serotec; red). Image capture was performed using Cell software (Olympus Biosystems) and colocalization analyzed using JascPaint Shop Pro 9.

TABLE 2. Primary Antibodies: Source and Protocols

Antibody (clone)IsotypeDilutionAntigen RetrievalSource
MW indicates microwave; SBH, sodium borohydride in PBS; ARTS, antigen retrieval target solution.
CD68 (PG-M1)IgG31:200MW 2×5 min in citrate bufferDAKO
Col IV (COL-94)IgG11:500MW 15 min citrate bufferSigma
HIF1α (NB 100–105)IgG11:1000ARTS (DAKO) at 98°C 45 min(Gift from ICRC Labs Oxford)
HIF2αIgG11:20MW 20 min citrate buffer+incubate SBH w15 min(Gift from ICRC Labs Oxford)
ICAM1/CD54 (AQX01/8041)Poly1:1000MW 10 min citrate bufferR&D Systems
MMP7IgG2b1:100MW 10 min citrate bufferCalbiochem
NgbPoly1:1000MW 10 min citrate bufferAbCam
NMBRPoly1:400MW 10 min citrate bufferAbCam
VEGFR2 (Flk-1)IgG11:400MW 15 min citrate bufferSanta Cruz
βA4 (3D/6F)IgG11:2006 h 80% formic acidDAKO

Details of the approach used to define and derive quantitative and semiquantitative measures of the molecular biological and morphological variables included in the study are included in the online data supplement, available at http://stroke.ahajournals.org.

Statistical Analysis

We analyzed groups defined by MRI (“lesional” or “nonlesional”), anatomical location (PVLs or DSCLs), and dementia status (demented or nondemented). Four categories of CAA (none, mild, moderate, and severe) were collapsed to a dichotomy between “none-mild” and “moderate-severe.”

Data were analyzed (STATA version 8.0 STATA Corporation) including ANOVA and Pearson χ2 tests for heterogeneity. ANOVA was adjusted for multiple readings within individuals. All analyses were checked using nonparametric statistics (eg, Kruskal–Wallis). The most conservative P values have been used. In the analysis of 3 categories (control, PVLs, and DSCLs) P values have been adjusted for multiple comparisons.

Results

Vascular Morphology

Hematoxylin and eosin sections of MRI-scanned coronal slices from a subgroup of the tissue resource were examined for qualitative assessment of vascular changes in the WM.13 Thickening of WM arterioles and small arteries was frequent. Although we did not use a specific stain appearances considered to represent periventricular venous collagenosis were rare.20 Estimations of arteriolar morphology were made in 390 arteriolar profiles from 19 brains (Table 1). Tangential or incomplete vascular profiles were avoided. The regions in which measurements were made were selected to include: nondemented individuals, WMLs (109 vessels), nonlesional WM (80 vessels); and demented individuals, WMLs (107 vessels), nonlesional WM (94 vessels).

Internal diameter was not reduced in either WMLs or in the presence of dementia (P>0.2, Figure 1a). Wall thickness within WMLs (nondemented [median] 116 μm; interquartile range [IQR] 47 to 194 μm; demented 95 μm [IQR 45 to 160 μm]) was significantly increased compared with nonlesional WM (nondemented 36 μm [IQR 21 to 96]; demented 24 μm [IQR 15 to 51; P<0.001]) and was not related to dementia (P>0.2). The perivascular space was enlarged within WMLs (P<0.001) and was slightly greater in those with dementia (P=0.04). There was no difference between cases with high neocortical neurofibrillary tangle counts (Braak stages 5 and 6; 66 vessels) compared with moderate or low counts (Braak stages 1 through 4; 324 vessels).

Figure 1. a, Bar chart (mean and SD) of vascular dimensions in DSCL (solid bars) compared with normal WM from the same anatomical area (dotted bars) showing significant vessel wall thickening and perivascular enlargement in lesional tissue. b, Bar chart (mean and SD) of ependymal grade compared with severity of MRI PVL showing increasing ependymal loss in more severe lesions; c, Bar chart (mean and SD) of collagen IV staining as a percentage of total field area showing no change in capillary density in lesional tissue. d, The ratio of ICAM staining to Col IV is increased in DSCL compared with normal WM in both demented (white) and nondemented (gray). MR lesions defined as “diffuse” change also show enhanced ICAM staining. e, Comparison of the ratio of ICAM to Collagen IV by gender. PVS indicates perivascular space.

Ependymal Morphology

In PVLs but not DSCLs, there was increased denudation of the ventricular lining compared with the ventricular lining in the absence of WMLs (P=0.004; Figure 1b). This effect is somewhat clustered within individuals and reduces in effect when multiple readings within individuals are adjusted for (P=0.04).

ICAM/Col IV Ratio

The percentage of area stained by Col IV was similar across all of the groups studied, indicating no significant variation in the density of the capillary network (Figure 1c). There was no difference in the ratio of ICAM/Col IV in PVLs compared with PVWM (P=0.17; Figure 1d). There was a highly significant increase in the ICAM/Col IV ratio for DSCLs compared with DWM (P<0.001). We also examined 32 areas of WM that were classified radiologically as “diffuse” change,13 implying subtle radiological changes and intermediate or mild myelin attenuation by histology, which showed elevation of the ICAM/Col IV ratio compared with nonlesional DSC WM (P=0.05) but no significant difference from DSCLs (P>0.2). Analysis for age and dementia showed no significant relationship with ICAM expression (Figure 1d and 1e).

Microglial Activation and HIF Expression

The expression of CD68, HIF1α, and HIF2α was estimated in DSC and PVWM using 43 samples from 20 brains with the shortest fixation times. They consisted of 13 DSCLs and 11 DWM from 14 brains and 10 PVLs and 9 PVWM from 14 brains. The cellular morphology of HIF immunoreactivity is consistent with either microglia or astrocytes so that double-labeling techniques are required to address the cell-type specificity of expression. Quantitative differences are shown in Figure 2. Significant differences between DSCLs and DWM samples were demonstrated for CD68 (Figure 2a; P=0.002), HIF1α (P=0.003), and HIF2α (P=0.005; Figure 2a). In PVLs compared with PVWM, there was no significant difference in expression of CD68 (P>0.2). HIF1α and HIF2α expression was higher but did not reach conventional significance (both P=0.1). There were no differences between demented and nondemented respondents (all P>0.2). All these relationships are affected by multiple readings within individuals; however, the sample is too small to properly adjust for this effect.

Figure 2. The frequency immunoreactive profiles for CD68, HIF1α, and HIF2α (a,b,c), neuroglobin (d), MMP7 (e), neuromedin B receptor (f), and VEGF receptor2 (g) in WMLs (outliers shown as circles). HIF1α and HIF2α expression compared with severity of leptomeningeal (b) and parenchymal (a) CAA. Both are significantly associated with increased HIF1α profiles in DSCL when CAA is present at moderate/severe grades (white boxes) compared with mild/none (gray boxes).

Effects of CAA on HIF Expression

The DWM of the centrum semiovale includes a watershed between perforating arteries derived from the cerebral cortex and long penetrating branches of arteries in the base of the brain. This arrangement potentially jeopardizes WM perfusion in the face of any diffuse arterial disease affecting the cerebral cortex. In the CFAS study, CAA is a frequent pathological finding among both demented and nondemented respondents.18 We compared the severity of HIF1α expression in WMLs and nonlesional WM with increasing severity of CAA affecting the cortex overlying the WM regions imaged. CAA was scored (Consortium to Establish a Registry of Alzheimer Disease Protocol) into none, mild, moderate, and severe. Because dementia status did not influence HIF1α expression, DSCLs from demented and nondemented respondents were pooled and compared with CAA scores, stratified into none/mild versus moderate/severe, and analyzed separately for parenchymal and meningeal involvement.

Higher CAA load, both in parenchymal and leptomeningeal vessels, was associated with elevated HIF1α expression in DSCLs compared with DWM (Figure 2b and 2c; P=0.04). This relationship was not observed to reach statistical significance for PVLs. HIF2α expression was not related to either leptomeningeal or parenchymal amyloid load (all P>0.2). The lack of association between HIF2α and CAA may reflect the higher constitutive levels of this protein, which is less clearly linked to the hypoxic response cascade. Double-labeling ICC for the HIFs showed HIF1α expression in both astrocytes (Figure 3) and microglia. In contrast, HIF2α expression was restricted to microglia.

Figure 3. Double labeling immunocytochemistry demonstrates cellular markers of microglia (CD68: a, g, m, s, a1 and g1) or astrocytes (glial fibrillary acidic protein; d, j, p, w, and d1). Neuroglobin (b and e) is colocalized in microglia (composite image c) but not astrocytes (f). MMP7 is colocalized to both microglia and astrocytes (i and l). Neuromedin B receptor is uniformly expressed in astrocytes (r) but only in a subpopulation of microglia (o; arrow). VEGF receptor 2 is colocalized to microglia (u) but not astrocytes (y). HIF1α is expressed in both microglia (c1) and astrocytes (f1). HIF2α is expressed only in microglia (i1) but not astrocytes (data not shown).

Expression of Hypoxia-Regulated Proteins

The immunocytochemical demonstration of HIF1α is capricious and is highly dependent on agonal factors and postmortem delay in the human brain. To strengthen the evidence for a hypoxic response within WMLs, we quantified the frequency of cellular profiles immunoreactive for 4 other candidate markers that have been shown to be upregulated by hypoxia.14,15 There was evidence that the frequency of Ngb-expressing cell profiles were different in the 3 groups (P=0.02), being increased in DSCLs compared with control (P=0.14) and a small nonsignificant decrease in PVLs compared with control (P=0.4; Figure 2d). MMP7 immunoreactive profiles varied considerably between groups (P<0.001) with increases in both DSCLs (P=0.009) and in PVLs (P<0.001; Figure 2e). Neither NMBR-expressing (P=0.35) nor VEGFR2-expressing (P=0.25) cell profiles showed variations by lesion type, although small differences may exist (Figure 2f and 2g). Double-labeling ICC (Figure 3) shows the expression of Ngb, MMP7, NMBR, and VEGFR2 in activated microglial cells. MMP7 and NMBR are also colocalized within astrocytes. Many of the cases show a population of clasmatodendritic astrocytes within both PVLs and DSCLs.

Discussion

WMLs and Markers of Hypoxia

The study demonstrated that markers associated with vascular degenerative morphology, endothelial and microglial activation, and molecular markers of a hypoxic response are altered in WMLs in the aging brain. These changes would be expected to impair, or reflect impairment of, vascular perfusion. The effect of CAA in the arterial vessels that supply DWM, and the sclerotic changes present within lesional WM, would both likely reduce perfusion in the centrum semiovale. Increased capillary endothelial activation and microglial activation are relatively nonspecific findings associated with a variety of vascular, hypoxic, and inflammatory insults, among other causes. However, the persistence of HIF-immunoreactive cell profiles and the increased expression of molecules that are demonstrated to be upregulated by HIF are strong evidence that WM attenuation is associated with ongoing hypoxia.14 HIF1α is normally controlled via ubiquitin-mediated proteolysis through rapid degradation under normoxic conditions.21,22 HIF1α is not readily demonstrable by ICC in normally oxygenated tissues.23 Persistence of HIF1α to the extent that it becomes demonstrable by ICC is regarded as strong evidence that the tissue is in a hypoxic state. The situation regarding HIF2α is somewhat less clear because of greater constitutive expression in normoxic conditions within macrophage lineage cells.23 HIF1α is especially capricious in immunocytochemical studies, and we were only able to demonstrate staining in cases with the shortest postmortem interval between death and tissue donation and processing. We therefore also looked for the expression of various molecules known to be upregulated either directly by HIF or as a response to hypoxia and show that both MMP7 and Ngb are significantly upregulated.

PVLs Versus DSCLs

The results suggest a difference between PVM and DWM. There are anatomical differences in the vascular supply of PVWM versus DWM. The lack of an association between the severity of CAA and HIF expression in PVLs may relate to the minimal perfusion contributed by cortically derived vessels in this part of the brain. It is possible that WM attenuation may differ between PVLs and DSCLs in terms of the extent of axonal loss, demyelination, and increased interstitial fluid. The PVWM may be prone to increased fluid accumulation related to the proximity of the ventricles associated with reduced integrity of the ventricular ependyma. However, some of our data for DSCLs can be interpreted in terms of increased retention of tissue fluid related to impaired perivascular fluid drainage. In the cerebral cortex, there is a significant retrograde fluid drainage pathway associated with the perivascular spaces of arteries and arterioles.24 The perivascular space associated with WM and basal ganglia arteries and arterioles is contiguous with this drainage system and is likely to function in a similar way. Enlargement of perivascular spaces in WMLs may represent enlargement caused by the accumulation of tissue fluid through impaired drainage.

The observed association between the thickened arterioles with enlarged perivascular spaces and the areas of lesional WM remains unexplained. Its absence from normal WM is difficult to explain in terms of a systemic disorder such as hypertension or arteriosclerosis, which ought to cause generalized arteriolar disease. The presence of arteriolar sclerosis in WMLs has been reported previously, and our findings confirm this previous work.25 In a cross-sectional observational study such as this, the possibility that WM changes lead to the vascular changes or vice versa cannot be ascertained.

Of the 5 hypoxia-associated markers included in this study HIF1α, HIF2α, VEGFR2, and Ngb were significantly upregulated only in DSCLs. MMP7 was significantly upregulated in DSCLs and in PVLs. We were unable to demonstrate significant upregulation of NMBR and VEGFR2, although both markers show a nonsignificant trend toward increased expression more marked in DSCLs (Figure 2f and 2g). For the first time, we demonstrated that both astrocytes (HIF1α, MMP7, and NMBR) and microglia (HIF1α, HIF2α, Ngb, MMP7, NMBR, and VEGFR2) express molecules that are involved in the adaptive and cytoprotective response to hypoxia (Figure 3). Coexpression of HIF1α in astrocytes is a novel observation. The astrocytic population in this cohort showed clasmatodendritic changes (dendritic pruning and cytoplasmic enlargement) that have been described previously in WMLs in AD and in Binswanger disease.26–28 Clasmatodendritic astrocytes in WMLs contain plasma proteins, suggesting impairment of blood–brain barrier dysfunction and phagocytosis of extracellular proteins. It is possible that observed immunolocalization of HIF1α, MMP7, and NMBR in astrocytes is also attributable to phagocytosis. Ongoing studies, including in situ hybridization, will determine which cell types synthesize these hypoxia-response proteins and whether there is any dysregulation of the downstream adaptive pathways that account for the failure of WM to adapt to hypoxia in brain aging.

This study provides direct evidence that hypoxia plays an important role in the pathogenesis of incidental WMLs in brain aging. The pathology of these incidental WMLs includes changes demonstrated previously in the WMLs associated with AD and with Binswanger disease, suggesting that they are part of a pathological continuum common to all 3 clinical scenarios.

The Cognitive Function and Ageing Neuropathology Study is funded by MRC United Kingdom. We are grateful to the respondents, their families, and their carers for agreement to participate in the brain donation program.

Footnotes

Correspondence to Professor Paul G. Ince, ‘E’ Floor, Royal Hallamshire Hospital, Glossop Road, Sheffield S10 2JF UK. E-mail

References

  • 1 Pantoni L, Garcia J. The significance of white-matter abnormalities 100 years after Binswanger’s report. Stroke. 1995; 26: 1293–1301.CrossrefMedlineGoogle Scholar
  • 2 de Groot J, de Leeuw F, Oudkerk M, Hofman A, Jolles J, Breteler M. Cerebral white matter lesions and subjective cognitive dysfunction: the Rotterdam Scan Study. Neurology. 2001; 56: 1539–1545.CrossrefMedlineGoogle Scholar
  • 3 de Leeuw F-E, de Groot JC, Achten E, Oudkerk M, Ramos LMP, Heijboer R, Hofman A, Jolles J, van Gijn J, Breteler MM. Prevalence of cerebral white matter lesions in elderly people: a population based magnetic resonance imaging study. The Rotterdam Scan Study. J Neurol Neurosurg Psychiatry. 2001; 70: 9–14.CrossrefMedlineGoogle Scholar
  • 4 Mirra SS, Heyman A, McKeel D, Sumi S, Crain B, Brownlee L, Vogel FS, Hughes JP, van Belle G, Berg L. The consortium to establish a registry of Alzheimer’s disease (CERAD) Part II. Standardization of the neuropathologic assessment of Alzheimer’s disease. Neurology. 1991; 41: 479–486.CrossrefMedlineGoogle Scholar
  • 5 Brun A, Englund E. A white matter disorder in dementia of the Alzheimer type: a pathoanatomical study. Ann Neurol. 1986; 19: 253–262.CrossrefMedlineGoogle Scholar
  • 6 Awad I, Spetzler R, Hodak J, Awad C, Carey R. Incidental subcortical lesions identified on magnetic resonance imaging in the elderly, I: correlation with age and cardiovascular risk factors. Stroke. 1986; 17: 1084–1089.CrossrefMedlineGoogle Scholar
  • 7 Fazekas F, Kleinert R, Offenbacher H, Schmidt R, Kleinert G, Payer F, Radner H, Lechner H. Pathologic correlates of incidental MRI white matter signal hyperintensities. Neurology. 1993: 1683–1689.MedlineGoogle Scholar
  • 8 Scheltens P, Barkhof F, Leys D, Wolters E, Ravid R, Kamphorst W. Histopathologic correlates of white matter changes on MRI in Alzheimer’s disease and normal aging. Neurology. 1995; 45: 883–888.CrossrefMedlineGoogle Scholar
  • 9 Haglund M, Englund E. Cerebral amyloid angiopathy, white matter lesions and Alzheimer encephalopathy—a histopathological assessment. Dement Geriatr Cogn Disord. 2002; 14: 161–166.CrossrefMedlineGoogle Scholar
  • 10 Tian J, Shi J, Bailey K, Mann M. Relationships between arteriosclerosis, cerebral amyloid angiopathy and myelin loss from the cerebral white matter in Alzheimer’s disease. Neuropathol Appl Neurobiol. 2004; 30: 46–56.CrossrefMedlineGoogle Scholar
  • 11 Pantoni L, Garcia J. Pathogenesis of leukoaraiosis. Stroke. 1997; 28: 652–659.CrossrefMedlineGoogle Scholar
  • 12 Tomimoto H, Akiguchi I, Wakita H, Suenaga T, Nakamura S, Kimura J. Regressive changes of astroglia in white matter lesions in cerebrovascular disease and Alzheimer’s disease patients. Acta Neuropathol. 1997; 94: 146–152.CrossrefMedlineGoogle Scholar
  • 13 Fernando MS, O’Brien JT, Perry RH, English P, Forster G, McMeekin W, Slade JY, Golkhar A, Matthews FE, Barber R, Kalaria RN, Ince PG; Neuropathology Group of MRC CFAS. Comparison of the pathology of cerebral white matter with post-mortem magnetic resonance imaging (MRI) in the elderly brain. Neuropathol Appl Neurobiol. 2004; 30: 385–395.CrossrefMedlineGoogle Scholar
  • 14 Burke B, Giannoudis A, Corke KP, Gill D, Wells M, Ziegler-Heitbrock L, Lewis CE. Hypoxia-induced gene expression in human macrophages: implications for ischemic tissues and hypoxia-regulated gene therapy. Am J Pathol. 2003; 163: 1233–1243.CrossrefMedlineGoogle Scholar
  • 15 Sun Y, Jin K, Mao XO, Zhu Y, Greenberg DA. Neuroglobin is up-regulated by and protects neurons from hypoxic-ischemic injury. Proc Natl Acad Sci U S A. 2001; 98: 15306–15311.CrossrefMedlineGoogle Scholar
  • 16 Folstein M, Folstein S, McHugh P. Mini-mental state: a practical method for grading the cognitive state of patients for the clinician. J Psychiatric Res. 1975; 313: 1419–1420.Google Scholar
  • 17 Copeland J, Dewey M, Griffiths-Jones H. Computerized psychiatric diagnostic system and case nomenclature for elderly subjects. Psychol Med. 1986; 16: 89–99.CrossrefMedlineGoogle Scholar
  • 18 Neuropathology Group. Medical Research Council Cognitive Function and Aging Study. Pathological correlates of late-onset dementia in a multicentre, community-based population in England and Wales. Neuropathology Group of the Medical Research. Lancet. 2001; 357: 169–75.CrossrefMedlineGoogle Scholar
  • 19 Scheltens P, Barkhof F, Leys D, Pruvo J, Nauta J, Vermersch P, Steinling M, Valk J. Semiquantative rating scale for the assessment of signal hyperintensities on magnetic resonance imaging. J Neurol Sci. 1993; 114: 7–12.CrossrefMedlineGoogle Scholar
  • 20 Moody D, Brown W, Challa V, Anderson R. Periventricular venous collagenosis: association with leukoariosis. Radiology. 1995; 194: 469–476.CrossrefMedlineGoogle Scholar
  • 21 Maxwell PH, Wiesener MS, Chang GW, Clifford SC, Vaux EC, Cockman ME, Wykoff CC, Pugh CW, Maher ER, Ratcliffe PJ. The tumor suppressor gene VHL targets hypoxia-inducible factors for oxygen-dependent proteolysis. Nature. 1999; 399: 271–275.CrossrefMedlineGoogle Scholar
  • 22 Wang GL, Jiang B-H, Rue EA, Semenza GL. Hypoxia-inducible factor 1 is a basic-helix-loop-helix-PAS heterodimer regulated by cellular O2 tension. Proc Natl Acad Sci U S A. 1995; 92: 5510–5514.CrossrefMedlineGoogle Scholar
  • 23 Talks KL, Turley H, Gatter KC, Maxwell P, Pugh CW, Ratcliffe PJ, Harris AL. The expression and distribution of the hypoxia-inducible factors HIF-1alpha and HIF-2 alpha in normal human tissues, cancers, and tumor associated macrophages. Am J Pathol. 2000; 157: 411–421.CrossrefMedlineGoogle Scholar
  • 24 Preston S, Steart P, Wilkinson A, Nicoll J, Weller R. Capillary and arterial cerebral amyloid angiopathy in Alzheimer’s disease: defining the perivascular route for elimination of amyloid beta from the human brain. Neuropathol Appl Neurobiol. 2003; 29: 106–117.CrossrefMedlineGoogle Scholar
  • 25 van Swieten J, van den Hout J, van Ketel B, Hijdra A, Wokke J, van Gijn J. Periventricular lesions in the white matter on magnetic resonance imaging in the elderly. A morphometric correlation with arteriolosclerosis and dilated perivascular spaces. Brain. 1991; 114: 761–774.CrossrefMedlineGoogle Scholar
  • 26 Akiguchi I, Tomimoto H, Suenaga T, Wakita H, Budka H. Blood-brain barrier dysfunction in Binswanger’s disease; an immunohistochemical study. Acta Neuropathol (Berl). 1998; 95: 78–84.CrossrefMedlineGoogle Scholar
  • 27 Sahlas D, Bilbao J, Swartz R, Black S. Clasmatodendrosis correlating with periventricular hyperintensity in mixed dementia. Ann Neurol. 2002; 52: 378–381.CrossrefMedlineGoogle Scholar
  • 28 Tomimoto H, Akiguchi I, Suenaga T, Nishimura M, Wakita H, Nakamura S, Kimura J. Alterations of the blood-brain barrier and glial cells in white-matter lesions in cerebrovascular and Alzheimer’s disease patients. Stroke. 1996; 27: 2069–2074.CrossrefMedlineGoogle Scholar
strokeahaStrokeStrokeStroke0039-24991524-4628Lippincott Williams & Wilkins
Annex – Supplemental Online-Only Content01062006

Members of the Cognitive Function and Ageing Neuropathology Study Group. Cambridge: Carol Brayne, Mark Chatfield, Fiona Matthews, John H. Xuereb; Liverpool: John Broome, Dan Crooks, Cherie McCracken, James MacKenzie; Newcastle: Raj N. Kalaria, Ian G. McKeith, John T. O’Brien, Tuomo Polvikoski; Nottingham: James S. Lowe; Oxford: Margaret M. Esiri, James H. Morris; Sheffield: Gill Forster, Paul G. Ince, Julie Simpson, Stephen B. Wharton.

Supplemental Information: Methods

Braak Staging Protocol

Standardized brain areas in formalin-fixed, paraffin-embedded blocks were used for Braak staging. The blocks used were from the hippocampus, including the temporal neocortex, sampled approximately at the level of the lateral geniculate body and from the occipital or parietal neocortex, depending on the study center. We were unable to retrieve the hippocampal block/slide from 5 brains. Neurofibrillary degeneration (including both neurofibrillary tangles and neuropil threads were assessed using either a silver stains (Bielschowsky’s; Gallyas) or a mouse monoclonal antibody to tau protein (PHF-1) with diaminobenzidine as the chromogen; all sections were counterstained with hematoxylin. The brains were staged using a modification1 of the Braak and Braak2 staging system on thin sections,3 which has been shown to have good inter-rater reliability.4 When the occipital striatal block was unavailable, the temporal isocortex, or parietal cortex,5 was rated.

TABLE I. Demographic Features of the Individuals by Cohort Center

CharacteristicCambridge n=37Liverpool n=70Newcastle n=34Nottingham n=39Oxford n=26Total n=206
Sex
    Male12 (32%)36 (51%)14 (41%)18 (46%)9 (35%)89 (43%)
    Female25 (68%)34 (49%)20 (59%)21 (54%)17 (65%)117 (57%)
Age at death
    <80 y12 (32%)7 (10%)13 (38%)9 (23%)6 (23%)47 (23%)
    80–90 y16 (44%)32 (46%)10 (29%)21 (54%)15 (58%)94 (45%)
    >90 y9 (24%)31 (44%)11 (32%)9 (23%)5 (19%)65 (31%)
Dementia status
    Nondemented21 (57%)31 (44%)17 (50%)24 (62%)14 (54%)107 (52%)
    Demented16 (43%)40 (56%)17 (50%0)15 (38%)12 (46%)99 (48%)

TABLE II. Sex Specific Percentage Frequency Distribution of DSCL Severity/Region/Age Categories

LocationGrade<80 y80–90 y>90 yOverall
Male (n=28)Female (n=20)Male (n=42)Female (n=60)Male (n=17)Female (n=37)Male (n=88)Female (n=117)
*P=0.05 (Kendall’s tau-b).
Frontal lesionsNone50.045.052.456.766.745.954.551.3
Mild17.920.016.76.716.718.917.012.9
Moderate10.720.014.310.016.713.513.612.8
Severe21.415.016.626.70.021.6*14.823.0
Parietal lesionsNone46.450.059.561.076.567.658.661.2
Mild14.325.016.715.35.910.813.815.5
Moderate10.720.011.98.517.610.812.611.2
Severe28.55.011.915.30.010.814.912.0
Occipital lesionsNone64.360.073.258.666.755.669.057.9
Mild7.110.09.813.816.716.710.314.1
Moderate7.110.04.96.95.611.15.78.8
Severe20.820.012.220.611.216.614.919.2

eLetters(0)

eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.