Perivascular drainage of solutes is impaired in the ageing mouse brain and in the presence of cerebral amyloid angiopathy
The deposition of amyloid-β (Aβ) peptides in the walls of leptomeningeal and cortical blood vessels as cerebral amyloid angiopathy (CAA) is present in normal ageing and the majority of Alzheimer’s disease (AD) brains. The failure of clearance mechanisms to eliminate Aβ from the brain contributes to the development of sporadic CAA and AD. Here, we investigated the effects of CAA and ageing on the pattern of perivascular drainage of solutes in the brains of naïve mice and in the Tg2576 mouse model of AD. We report that drainage of small molecular weight dextran along cerebrovascular basement membranes is impaired in the hippocampal capillaries and arteries of 22-month-old wild-type mice compared to 3- and 7-month-old animals, which was associated with age-dependent changes in capillary density. Age-related alterations in the levels of laminin, fibronectin and perlecan in vascular basement membranes were also noted in wild-type mice. Furthermore, dextran was observed in the walls of veins of Tg2576 mice in the presence of CAA, suggesting that deposition of Aβ in vessel walls disrupts the normal route of elimination of solutes from the brain parenchyma. These data support the hypothesis that perivascular solute drainage from the brain is altered both in the ageing brain and as a consequence of CAA. These findings have implications for the success of therapeutic strategies for the treatment of AD that rely upon the health of the ageing cerebral vasculature.
KeywordsAlzheimer’s disease Amyloid-β Cerebral vasculature Perivascular drainage Basement membranes Cerebral amyloid angiopathy
Cerebral amyloid angiopathy (CAA) results from the deposition of amyloid-β (Aβ) peptides in the walls of leptomeningeal and cortical arteries and capillaries and is present in the majority of patients with Alzheimer’s disease (AD) . The vascular Aβ deposits, which are also found in an estimated 30–40% of non-demented elderly individuals, are associated with capillary thinning and vessel tortuosity, vasoconstriction, intraluminal thickening, inhibition of angiogenesis and the death of pericytes, endothelial and smooth muscle cells [9, 22, 39, 49, 61]. CAA is also associated with cerebral hypoperfusion, microhaemorrhages and cognitive impairment [15, 44, 50, 51, 56].
While the aetiology of sporadic CAA is unclear, it is likely caused by the failure of Aβ elimination from the brain parenchyma. Aβ is degraded by enzymes and removed from the brain by multiple mechanisms, including receptor-mediated transport across the endothelium, uptake by microglia and macrophages and, from its distribution in CAA, by bulk flow along the basement membranes of capillary and artery walls [4, 69]. With ageing and in AD, enzymatic degradation, absorption of Aβ into the blood and microglial uptake have been reported to be reduced or dysfunctional [19, 35, 41]. However, to date, the effects of ageing on the perivascular drainage of solutes, and the implications for the development and progression of CAA, have not been examined.
Basement membranes are thin sheets of highly specialized extracellular matrix that, outside the brain, are typically found at the epithelial-mesenchymal interface and regulate cell growth, differentiation and migration [3, 18, 62]. Cerebrovascular basement membranes are composed of laminins, nidogens and collagen IV, as well as heparan sulphate proteoglycans such as perlecan, agrin and fibronectin . Multiple lines of evidence suggest that basement membranes may play a role in the aetiology of CAA. Kinetic and biophysical studies have shown that laminin, collagen IV and nidogen interact directly with Aβ to prevent its aggregation and promote disaggregation of pre-formed fibrils, whereas agrin and perlecan induce and stabilize Aβ fibrillization [7, 8, 13, 17, 31, 32, 43]. Transgenic mice overexpressing transforming growth factor-β1 (TGF-β1) have thickened basement membranes preceding the development of murine CAA [45, 72]. Furthermore, thickening, reduplication and vacuolization of basement membranes, along with upregulation of heparan sulphate proteoglycans, have been reported in the AD brain [49, 55, 67]. Collectively, these findings indicate that alterations in the composition, level of expression or morphology of cerebrovascular basement membranes may disturb perivascular drainage of Aβ from the parenchyma and initiate deposition of Aβ within the walls of the cerebral vasculature. Moreover, clearance of Aβ and other toxic metabolites from the brain may be further impeded by the accumulation of vascular Aβ deposits as CAA.
In the present study, we tested the hypothesis that ageing of cerebral blood vessels and the development of CAA alter the pattern of perivascular drainage of solutes from the brain in naïve mice and in the Tg2576 mouse model of AD. We report that clearance of small molecular weight dextran is impaired in the hippocampal microvasculature of 22-month-old wild-type mice compared to 3- and 7-month-old animals. This was associated with age-dependent changes in both capillary density and basement membrane protein composition. The presence of CAA affected the drainage of dextran mainly in the walls of veins of 22-month-old Tg2576 mice, compared to younger mice without vascular Aβ deposits. Dextran also bound directly to parenchymal and vascular Aβ in aged Tg2576 mice. These data support the hypothesis that perivascular drainage is altered both in the ageing brain and as a consequence of the development of CAA. Our findings have implications for the success of therapies that rely on the diffusion and clearance of parenchymal Aβ via bulk flow movement along perivascular drainage pathways in the elderly brain. There are also wider consequences related to the kinetics of drug distribution within the elderly and AD brain.
Materials and methods
Male and female Tg2576 mice expressing the human Swedish (K670N, M671L) APP mutations were kindly provided by Dr. Karen Ashe (University of Minnesota, USA) and maintained on an outbred C57BL6 background. For perivascular drainage experiments, 3-, 7- and 22-month-old Tg2576 mice or age-matched wild-type littermates were used (n = 6/group). A separate group of 3-, 8- and 16-month-old BalbC mice (n = 5/group) was used for immunoblotting experiments. All experiments were performed in accordance with animal care guidelines stipulated by the Animal Care and Use Committees at the Universities of Leipzig and Southampton and conformed to UK Home Office regulations and to the European Communities Council Directive (86/609/EEC).
Tg2576 mice and wild-type littermates were anaesthetized with an intraperitoneal injection of etomidate (Hypnomidate; 33 mg/kg body weight; Janssen-Cilag, Neuss, Germany). Additionally, local anaesthesia of the skull was achieved with a subcutaneous injection of lidocaine hydrochloride (Licain; 1%; 17.5 mg/kg body weight; DeltaSelect, Pfullingen, Germany). All animals received a stereotaxic injection of 0.7 μL FITC-conjugated 10 kDa dextran amine (Sigma-Aldrich, Dorset, UK; 1 mg/ml in sterile PBS) into the left dorsal hippocampus over a 4-min period (coordinates from Bregma: AP = −2 mm; ML = 1.8 mm and DV = 1.6 mm). Injection capillaries were left in situ for 2 min to allow for diffusion of dextran in the brain parenchyma and mice were killed 15 min post-injection.
Mice were deeply anaesthetised with an overdose of sodium pentobarbital, intracardially perfused with 0.1 M phosphate buffered saline (PBS; pH 7.4) followed by 4% paraformaldehyde, brains were post-fixed overnight, cryoprotected in 30% sucrose and sectioned at 10 μm. For thioflavin S staining, sections were treated with 1% thioflavin S for 5 min, differentiated twice in 70% ethanol, washed in PBS and coverslipped with anti-fading solution (Vector Labs, Peterborough, UK). For Aβ immunolabelling, tissue sections were incubated for 15 min with 3% hydrogen peroxide, treated with 70% formic acid (5 min), blocked with 15% normal goat serum and incubated overnight at 4°C with anti-Aβ17–24 (clone 4G8; 1:500; Millipore, Watford, UK). For glucose transporter-1 (glut-1) immunostaining, sections were blocked in 3% hydrogen peroxide, followed by 15% normal goat serum and incubated overnight with anti-glut-1 (1:750; Merck, Nottingham, UK). Sections were washed with PBS, incubated with anti-mouse or anti-rabbit horseradish peroxidase conjugates (1:400; Vector Labs, Peterborough, UK) and developed with nickel-enhanced diaminobenzidine as chromogen. For double immunofluorescence labelling, sections were blocked in 15% goat serum and incubated overnight at 4°C with rabbit anti-laminin (1:500; Sigma-Aldrich, Dorset, UK) and anti-α smooth muscle actin (1:300; Sigma-Aldrich). The next day, sections were rinsed in PBS, developed with the appropriate fluorescently conjugated secondary antibodies (1:200; Invitrogen, Paisley, UK) and coverslipped with anti-fading solution (Vector Labs). For double labelling of Aβ and laminin, brain tissue sections were incubated overnight with anti-laminin (1:500), developed with AlexaFluor 546-conjugated anti-rabbit (1:200), treated for 2 min with 70% formic acid and washed in PBS before being incubated overnight with anti-Aβ17–24 IgG (1:100). Sections were developed the following day with AlexaFluor 633-conjugated anti-mouse (1:200). Photomicrographs were captured using a Lexica SP2 confocal laser scanning microscope and exported to Photoshop CS or Image J imaging software (NIH, Maryland, USA).
Isolation of cortical blood vessels
Mice were deeply anaesthetised with an overdose of sodium pentobarbital, intracardially perfused with 0.1 M PBS (pH 7.4), and the brains rapidly removed and the cerebral cortex removed. To isolate blood vessels, cortical tissues were mechanically disrupted in 500 μL of 0.1 M NH4CO3 + 7% sodium dodecyl sulphate (SDS) (plus protease inhibitor cocktail set III, Merck) and agitated for approximately 4 h. Capillaries were separated from large-diameter vessels by passing the vascular tufts through a 10 μm nylon mesh filter and collected in the filtrate. Large-diameter vessels were suspended in RIPA lysis buffer [20 mM Tris–HCl (pH 8.0), 150 mM NaCl 1 mM EDTA, 1% Igepal, 0.1% SDS, 50 mM NaF, 1 mM NaVO3] sonicated, aliquoted and stored at −80°C.
For laminin and fibronectin blots, samples were diluted in NuPAGE® LDS sample buffer (Invitrogen) containing 2.5% β-mercaptoethanol and heated at 75°C for 5 min. Collagen samples were prepared without β-mercaptoethanol or heat. For the detection of perlecan, samples (30 μg) were digested with 5 mU heparinase III (Sigma-Aldrich) in the presence of protease inhibitor cocktail set for 2 h at 37°C before loading. Proteins were separated by gel electrophoresis on 3–8% Tris–acetate gels (15 μg/lane; Invitrogen) and transferred to nitrocellulose membranes. Membranes were blocked for 1 h at room temperature with 8% non-fat milk powder and incubated overnight at 4°C with anti-collagen IV (1:2,500; Abcam, Cambridge, UK), anti-laminin (1:500), anti-fibronectin (1:3,000; AbD Serotec, Kidlington, UK) or anti-perlecan (1:750; Millipore) antibodies. Membranes were developed using an enhanced chemiluminescence detection kit, stripped and reproved with anti-glyceraldehyde-3-PDH (GAPDH) antibody (1:50,000; Sigma-Aldrich) to ensure equal protein loading.
Quantification and statistical analysis
The number of blood vessels containing dextran was quantified from five images per mouse throughout the hippocampus using Zeiss Image Analysis KS-400 software. The number and diameter of dextran-positive blood vessels in a 574 μm × 689 μm area were documented manually. Vessels were categorized as: (1) capillaries if their diameters were less than 10 μm, (2) arteries if their diameters were greater than 10 μm and positive for α smooth muscle actin or (3) veins if the diameters were greater than 10 μm and the vessels lacked α-smooth muscle actin immunoreactivity . Histograms were compiled using mean ± SEM values and analysed using one-way analysis of variance (ANOVA) with Newman–Keuls post hoc test, significance set at p < 0.05. For quantification of glut-1, micrographs were converted to binary images (4 sections/mouse), evaluated by densitometry using Image J software and evaluated using one-way ANOVA with Newman–Keuls post hoc test (significance p < 0.05). Immunoblots were quantified by densitometry as an optical density ratio of protein levels normalized to GAPDH levels and analysed using repeated measures two-way ANOVA with Bonferroni post hoc test (significance set at p < 0.05).
CAA development in Tg2576 mice
Effect of age and CAA on perivascular drainage of solutes
Among the Tg2576 mice, no differences were observed in the number of capillaries or arteries containing dextran in their walls between young, adult and aged mice (Fig. 3g, h, p > 0.05, one-way ANOVA). However, 22-month-old Tg2576 mice had significantly fewer dextran-positive capillaries compared to wild-type littermates (Fig. 3g, p < 0.001, one-way ANOVA). Similarly, a significantly lower number of labelled arteries were counted in 7-month-old Tg2576 mice compared to age-matched controls (Fig. 3h, p < 0.05, one-way ANOVA). Dextran labelling of veins in the Tg2576 mice was increased proportional to CAA severity, such that 22-month-old mice demonstrated significantly more labelling of veins compared with both 3- and 7-month-old Tg2576 mice (Fig. 3i, p < 0.001 and p < 0.05, respectively, one-way ANOVA), as well as age-matched controls (Fig. 3i, p < 0.01, one-way ANOVA). Taken together, these data suggest that perivascular drainage of solutes from the brain is disrupted by age and CAA, but with differential effects on the micro- and macrovasculature, respectively.
Changes in vessel density in wild-type and Tg2576 mice
Changes in basement membrane protein levels with age
There is now a strong body of evidence suggesting that failure of clearance mechanisms for the elimination of Aβ from the brain contributes to the development of CAA and AD [38, 70]. We report here the first experimental data demonstrating that drainage of intracerebrally injected dextran along the basement membranes of capillaries and arteries is altered in the ageing mouse brain. Furthermore, the presence Aβ deposits within the walls of blood vessels disrupts the normal route of elimination of solutes from the brain parenchyma. The disturbance in the pattern of perivascular drainage of tracer is associated with changes in blood vessel density related to increasing age and the development of CAA. The expression levels of specific cerebrovascular basement membrane proteins are also altered in the ageing brain.
Unlike other organs, the brain has no conventional lymphatics for the removal of interstitial fluid (ISF). Instead, ISF and solutes are eliminated by bulk flow along the basement membranes of capillary and artery walls [6, 27, 58]. Several lines of evidence support the hypothesis that dysfunction in the perivascular drainage of neuronally produced Aβ leads to the development of CAA: (1) transgenic mice expressing mutant human Aβ only in neurons develop vascular Aβ deposits [11, 65]; (2) tracers injected into the interstitial fluid of the adult mouse brain distribute along basement membranes in exactly the same pattern as Aβ deposition in human CAA ; (3) basement membrane abnormalities are observed in the ageing and AD brain and precede CAA development in the TGF-β1 transgenic mice [45, 49, 72], (4) Aβ deposits in the AD brain have been identified in perivascular compartments [52, 60] and (5) anti-Aβ immunotherapy experiments in mice and humans have demonstrated increased severity of CAA associated with decreased cerebral plaque load [5, 48, 51, 71]. The current findings provide strong experimental support for the hypothesis that perivascular drainage of solutes from the brain is altered with age and presence of CAA.
Abnormalities in the ageing cerebrovasculature, particularly within capillaries, have previously been described in both animal and human studies. These include morphologic (e.g., vessel looping, tortuosity and twisting), physiological (e.g., loss of smooth muscle and basement membrane thickening) and functional changes (e.g., decreased cerebral blood flow) [2, 10, 24, 26, 29, 36, 42, 53, 63]. In correspondence with these reports, we found that perivascular drainage was affected primarily in the small vessels of the aged mouse brain, with significant increases in the number of dextran-labelled capillaries between 3-, 7- and 22-months of age. Hippocampal capillary density was also significantly higher in 22-month-old mice compared to 3-month-old, but not 7-month-old wild-type mice, indicating that the number of capillaries increases in the maturing mouse brain but remains constant after 7 months of age. The stability of capillary density between 7- and 22-months of age, suggests that the increase in dextran labelling of capillaries observed in 22-month-old animals is due to a decline in the rate of perivascular drainage of solutes from the extracellular spaces of the parenchyma. This may be accompanied by a reduction in the rate of drainage of dextran from capillaries to arteries, as a decrease in the number of arteries labelled with dextran was also noted in the 22-month-old mice.
Basement membrane proteins play a critical role in the drainage of ISF from the brain. We found that intracerebrally injected dextran co-localized with laminin in hippocampal blood vessels, suggesting that basement membranes may interact directly with solutes to mediate their clearance along perivascular pathways. Therefore, alterations to basement membrane morphology and/or up- or downregulation of specific basement membrane proteins could influence Aβ aggregation, alter the rate of drainage and influence the development of CAA. Our immunoblotting results showed that the levels of laminin within capillary walls are significantly lower in the aged brain, while expression of fibronectin and perlecan is increased compared to young and adult mice, respectively. Considering that laminin and perlecan are known to have opposing effects on the prevention and promotion of Aβ aggregation, respectively, a corresponding shift in their expression levels in the aged human brain could promote vascular Aβ deposition [1, 8, 13, 43, 66]. Perlecan has also been shown to regulate the growth of smooth muscle and endothelial cells, and is thought to provide stability to the basement membrane scaffold [16, 33]. Although further experiments are needed, we suggest that the observed decrease in perlecan expression between 3 and 8 months may serve to prevent excess blood vessel formation and to stabilize the structure of established vessels in the adult brain. In the ageing brain, perlecan levels may be upregulated to compensate for the destabilization of the basement membrane induced by the decreased expression of laminin and other basement membrane proteins. Although little is known about the interaction between fibronectin and Aβ, a recent report has shown an association between fibronectin and the Aβ-degrading angiotensin converting enzyme-1 in the artery walls of human brains with CAA, suggesting a putative role for fibronectin in vascular Aβ processing . Collectively, these results suggest that ageing may be accompanied by a change in the biochemical composition of the cerebrovascular basement membranes, towards the promotion of Aβ deposition. Interestingly, although increased immunohistochemical staining for collagen IV has been reported in the ageing human brain [61, 64], we did not observe any changes in collagen IV expression between age groups in the mouse by Western blotting. Although further investigation is needed to account for this discrepancy, it is possible that mouse collagen IV is less susceptible to age-related changes or that other collagen isoforms are predominantly affected by ageing.
We have previously established that tracers injected into the extracellular spaces of the naïve adult mouse brain drain predominantly along the basement membranes of capillaries and arteries, with little or no labelling of veins . In the current study, we found that the number of capillaries labelled with dextran was significantly lower in 22-month-old mice Tg2576 mice compared to wild-type littermates. This finding likely correlates with the decrease in hippocampal capillary density in these mice, as reported here and previously [34, 46]. However, the most striking effect of CAA on perivascular drainage was the significant increase in the number of veins that were positive for dextran, which was not associated with alterations in total vessel number. This suggests that the deposition of vascular amyloid in capillary and artery walls alters the normal route of elimination of solutes from the brain. In addition, the reported decrease in vasodilatory response and cerebral blood flow induced by Aβ in artery walls of both young and aged Tg2576 mice may further disrupt the drainage of interstitial fluid (ISF) [23, 37, 47, 56]. Indeed, if the motile force for elimination of solutes along perivascular drainage pathways is generated from arterial pulsations , it possible that reductions in pulse amplitude related to alterations in cerebral blood pressure or flow and altered compliance of the aged vessel walls (i.e. stiffening), would slow the rate of drainage and encourage precipitation of Aβ in vessel walls .
Finally, our data indicate that dextran within the ISF co-localizes with the insoluble deposits of Aβ in the parenchyma and vasculature. This suggests that solutes, including soluble Aβ, brain metabolites or drugs that are contained in parenchymal ISF, may bind to Aβ plaques and thus affect their distribution within the brain, particularly in areas of high plaque and CAA load. Furthermore, the design of AD therapies and therapeutic agents that target parenchymal plaque removal should take into consideration the decreased efficiency of Aβ clearance mechanisms in the elderly and AD brain.
This work is funded by the Alzheimer’s Research Trust UK (C.A.H, J.N, R.O.C) and the German Alzheimer Forschung Initiative (R.S.). The authors would like to thank the Biomedical Imaging Unit (Southampton General Hospital), as well as Ute Bauer and Dr. Anke Hoffmann for excellent technical assistance. We would also like to express our gratitude to Dr. Karen Hsiao Ashe, Department of Neurology, University of Minnesota, USA, for kindly providing Tg2576 founder mice.
Conflict of interest
The authors declare they have no conflict of interest.
- 9.Burger S, Noack M, Kirazov LP, Kirazov EP, Naydenov CL, Kouznetsova E et al (2009) Vascular endothelial growth factor (VEGF) affects processing of amyloid precursor protein and beta-amyloidogenesis in brain slice cultures derived from transgenic Tg2576 mouse brain. Int J Dev Neurosci 27:517–523PubMedCrossRefGoogle Scholar
- 12.Carare RO, Bernardes-Silva M, Newman TA, Page AM, Nicoll JA, Perry VH et al (2008) Solutes, but not cells, drain from the brain parenchyma along basement membranes of capillaries and arteries: significance for cerebral amyloid angiopathy and neuroimmunology. Neuropathol Appl Neurobiol 34:131–144PubMedCrossRefGoogle Scholar
- 23.Han BH, Zhou ML, Abousaleh F, Brendza RP, Dietrich HH, Koenigsknecht-Talboo J et al (2008) Cerebrovascular dysfunction in amyloid precursor protein transgenic mice: contribution of soluble and insoluble amyloid-beta peptide, partial restoration via gamma-secretase inhibition. J Neurosci 28:13542–13550PubMedCrossRefGoogle Scholar
- 35.Krause DL, Muller N (2010) Neuroinflammation, microglia and implications for anti-inflammatory treatment in Alzheimer’s disease. Int J Alzheimers DisGoogle Scholar
- 39.Miao J, Xu F, Davis J, Otte-Holler I, Verbeek MM, Van Nostrand WE (2005) Cerebral microvascular amyloid beta protein deposition induces vascular degeneration and neuroinflammation in transgenic mice expressing human vasculotropic mutant amyloid beta precursor protein. Am J Pathol 167:505–515PubMedCrossRefGoogle Scholar
- 44.Natte R, Maat-Schieman ML, Haan J, Bornebroek M, Roos RA, van Duinen SG (2001) Dementia in hereditary cerebral hemorrhage with amyloidosis-Dutch type is associated with cerebral amyloid angiopathy but is independent of plaques and neurofibrillary tangles. Ann Neurol 50:765–772PubMedCrossRefGoogle Scholar
- 57.Steffensen B, Chen Z, Pal S, Mikhailova M, Su J, Wang Y et al (2010) Fragmentation of fibronectin by inherent autolytic and matrix metalloproteinase activities. Matrix BiolGoogle Scholar
- 61.Tian J, Shi J, Smallman R, Iwatsubo T, Mann DM (2006) Relationships in Alzheimer’s disease between the extent of Aβ deposition in cerebral blood vessel walls, as cerebral amyloid angiopathy, and the amount of cerebrovascular smooth muscle cells and collagen. Neuropathol Appl Neurobiol 32:332–340PubMedCrossRefGoogle Scholar
- 67.van Horssen J, Otte-Holler I, David G, Maat-Schieman ML, van den Heuvel LP, Wesseling P et al (2001) Heparan sulfate proteoglycan expression in cerebrovascular amyloid beta deposits in Alzheimer’s disease and hereditary cerebral hemorrhage with amyloidosis (Dutch) brains. Acta Neuropathol 102:604–614PubMedGoogle Scholar
- 71.Wilcock DM, Rojiani A, Rosenthal A, Subbarao S, Freeman MJ, Gordon MN et al (2004) Passive immunotherapy against Aβ in aged APP-transgenic mice reverses cognitive deficits and depletes parenchymal amyloid deposits in spite of increased vascular amyloid and microhemorrhage. J Neuroinflammation 1:24PubMedCrossRefGoogle Scholar