AimsWhile vascular pathology is a common feature of a range of neurodegenerative diseases, we hypothesized that vascular changes occur in association with normal ageing. Therefore, we aimed to characterize age‐associated changes in the blood–brain barrier (BBB) in human and mouse cohorts.MethodsImmunohistochemistry and Evans blue assays were used to characterize BBB dysfunction (tight junction protein expression and serum plasma protein accumulation), vascular pathology (pericyte loss and vascular density) and glial pathology (astrocyte and microglial density) in ageing neurological control human prefrontal cortex (a total of 23 cases from 5 age groups representing the spectrum of young adult to old age: 20–30 years, 31–45 years, 46–60 years, 61–75 years and 75+) and C57BL/6 mice (3 months, 12 months, 18 months and 24 months, n = 5/6 per group).ResultsQuantification of the tight junction protein ZO‐1 within the cortex and cerebellum of the mouse cohort showed a significant trend to both increased number (cortex P < 0.001, cerebellum P < 0.001) and length (cortex P < 0.001, cerebellum P < 0.001) of junctional breaks associated with increasing age. GFAP expression significantly correlated with ageing in the mice (P = 0.037). In the human cohort, assessment of human protein accumulation (albumin, fibrinogen and human IgG) demonstrated cells morphologically resembling clasmatodendritic astrocytes, indicative of BBB dysfunction. Semiquantitative assessment of astrogliosis in the cortex expression revealed an association with age (P = 0.003), while no age‐associated changes in microglial pathology, microvascular density or pericyte coverage were detected.ConclusionsThis study demonstrates BBB dysfunction in normal brain ageing, both in human and mouse cohorts.

No abstract

This article presents analyses of attainment variations for ® ve cohorts of school leavers between 1988 and 1995. Social class, ethnicity and gender variations in educational attainment at 16 are examined using data from over 80,000 young people in England and Wales. This is the only representative data set that allows analyses of educational variations in England and Wales across social class, ethnic and gender groups simultaneously. The preliminary analyses showed an underrepresentation of some ethnic and lower social class groups in the independent education sector. The attainment analyses thereafter are for state school pupils only. Between 1988 and 1995, attainment differences relating to social class, ethnicity and gender are all seen to increase. In terms of the percentages leaving compulsory schooling with ® ve or more high grade passes (grade A*± C at General Certi® cate of Secondary Education), the social class gap of 50 percentage points widened to 56 percentage points, the ethnicity gap from 14 to 20 percentage points and the gender gap from 5 to 10 percentage points. The uneven improvement in educational attainment across ethnic, social class and gender groups is a cause for concern. For certain subgroups there appears to be little or no improvement over the 8 years. These ® ndings relate to a period of considerable political intervention within the British compulsory education system. They suggest that policies focused on raising educational standards may be at variance with the aim of reducing educational inequality.

IntroductionDeposition of abnormally phosphorylated tau (phospho-tau) occurs in Alzheimer’sdisease but also with brain ageing. The Braak staging scheme focused on neurofibrillary tangles, butabundant p-tau is also present in neuropil threads, and a recent scheme has been proposed by theBrainNet Europe consortium for staging tau pathology based on neuropil threads. We determined therelationship of threads to tangles, and the value of staging for threads in an unselected population-representative ageing brain cohort. We also determined the prevalence of astroglial tau pathologies, and their relationship to neuronal tau. Phospho-tau pathology was determined by immunohistochemistry (AT8 antibody) in the MRC-CFAS neuropathology cohort. Neuropil threads were staged using the BrainNet Europe protocol for tau pathology, and compared with Braak tangle stages. Astroglial tau pathology was assessed in neo-cortical, mesial temporal and subcortical areas.ResultsCases conformed well to the hierarchical neuropil threads staging of the BrainNet Europe protocol and correlated strongly with Braak staging (r=0.84, p < 0.001). Based on the areas under the receiver operator curves (AUC), incorporating either threads or tangle staging significantly improved dementia case identification to a similar degree over age alone (Braak stage X2(1)=10.1, p=0.002; BNE stage X2(1)=9.7, p=0.002). Thorn-shaped astrocytes, present in 40 % of cases, were most common in mesial temporal lobe, then brainstem, and were associated with subpial tau-positive neurites (mesial temporal: X2(1)=61.3, p < 0.001; brainstem: X2(1)=47.9, p < 0.001). Adding thorn astrocytes did not improve dementia prediction (AUC: X2(1)=0.77, p=0.381), but there was a weak relationship between numbers of areas involved and staging for threads or tangles (r=0.17, p=0.023). Neuropil threads develop hierarchically in parallel with neurofibrillary tangles.ConclusionsThe BrainNet Europe staging protocol is straightforward to apply, and offers similar predictive value for dementia to Braak tangle staging. Astroglial tauopathy, particularly thorn shaped astrocyte formation, does not relate to dementia status, but the association with phospho-tau neurites may suggest a pathogenic relationship to neuronal tau pathology.

IntroductionMagnetic resonance imaging (MRI) cerebral microbleeds (CMB) arise from ferromagnetic haemosiderin iron assumed to derive from extravasation of erythrocytes. Light microscopy of ageing brain frequently reveals foci of haemosiderin from single crystalloids to larger, predominantly perivascular, aggregates. The pathological and radiological relationship between these findings is not resolved.MethodsHaemosiderin deposition and vascular pathology in the putamen were quantified in 200 brains donated to the population-representative Medical Research Council Cognitive Function and Ageing Study. Molecular markers of gliosis and tissue integrity were assessed by immunohistochemistry in brains with highest (n = 20) and lowest (n = 20) levels of putamen haemosiderin. The association between haemosiderin counts and degenerative and vascular brain pathology, clinical data, and the haemochromatosis (HFE) gene H63D genotype were analysed. The frequency of MRI CMB in 10 cases with highest and lowest burden of putamen haemosiderin, was compared using post mortem 3T MRI.ResultsGreater putamen haemosiderin was significantly associated with putaminal indices of small vessel ischaemia (microinfarcts, P < 0.05; arteriolosclerosis, P < 0.05; perivascular attenuation, P < 0.001) and with lacunes in any brain region (P < 0.023) but not large vessel disease, or whole brain measures of neurodegenerative pathology. Higher levels of putamen haemosiderin correlated with more CMB (P < 0.003).ConclusionsThe MRI-CMB concept should take account of brain iron homeostasis, and small vessel ischaemic change in later life, rather than only as a marker for minor episodes of cerebrovascular extravasation. These data are of clinical relevance, suggesting that basal ganglia MRI microbleeds may be a surrogate for ischaemic small vessel disease rather than exclusively a haemorrhagic diathesis.