Genetic AD (Alzheimer's disease) accounts for only few AD cases and is almost exclusively associated with increased amyloid production in the brain. Instead, most patients are affected with the sporadic form of AD and typically have altered clearance mechanisms. The identification of factors that influence the onset and progression of sporadic AD is a key step towards understanding its mechanism(s) and developing successful therapies. An increasing number of epidemiological studies describe a strong association between AD and cardiovascular risk factors, particularly hypertension, that exerts detrimental effects on the cerebral circulation, favouring chronic brain hypoperfusion. However, a clear demonstration of a pathophysiological link between cardiovascular risk factors and AD aetiology is still missing. To increase our knowledge of the mechanisms involved in the brain's response to hypertension and their possible role in promoting amyloid deposition in the brain, we have performed and investigated in depth different murine models of hypertension, induced either pharmacologically or mechanically, leading in the long term to plaque formation in the brain parenchyma and around blood vessels. In the present paper, we review the major findings in this particular experimental setting that allow us to study the pathogenetic mechanisms of sporadic AD triggered by vascular risk factors.
Sporadic AD (Alzheimer's disease) compared with familial AD
AD is the leading cause of dementia in the elderly and, although some treatments help the symptoms and slow the progression of the disease, no definitive cure exists. Early-onset/genetic AD accounts for only a small proportion of total cases and is associated with an increased production of the Aβ (amyloid β-peptide). In contrast, the vast majority of patients are affected by late-onset/sporadic AD, with no evident increase of Aβ production, but probably with a failure in the mechanisms responsible for the clearance of Aβ from the brain. Both vascular efflux of Aβ across the BBB (blood–brain barrier) and influx of circulating Aβ in the CNS (central nervous system) are mechanisms that, when dysregulated, may contribute to Aβ accumulation, leading to its aggregation and deposition [1,2].
Actually, it seems that the only step forward for the development of novel successful rational therapies is the identification of factors that influence onset/progression of the sporadic AD for an understanding of its mechanism(s).
Despite the old belief that AD, in contrast with vascular dementia, has a non-vascular origin, a growing body of epidemiological studies strongly associate cardiovascular risk factors, such as hypertension, with increase of AD [1–4]. Indeed, it is now well known that hypertension exerts a strong affect on cerebral circulation, favouring hypoperfusion through endothelial function impairment, increased smooth muscle constrictor tone and structural remodelling that significantly modifies vascular compliance [5,6]. Moreover, recent studies on genetic AD animal models further suggest an interaction among anti-hypertensive therapies and a benefit on cerebral amyloid pathology and cognitive decline [7,8].
However, despite these clear links between hypertension and AD, few basic science studies have investigated this relationship so far.
AD, amyloid deposition and cerebral vasculature
Many molecular lesions have been detected in AD, but the dominant theme to emerge from the data is that an accumulation of misfolded proteins in the aging brain results in oxidative and inflammatory damage, which in turn leads to energy failure, synaptic dysfunction and eventually cognitive decline.
Over the last few decades, most attention has been focused on genetic animal models that could resemble the AD pathological features by overproducing the Aβ, resulting in its aggregation and deposition in brain parenchyma [9–12]. However, the question still unanswered is: do they really represent what happens in the natural evolution of the human pathology?
Nevertheless, the huge amount of studies conducted in these animal models allowed us to understand that Aβ spontaneously self-aggregates into multiple coexisting physical forms. One form consists of oligomers, which coalesce into intermediate assemblies. β-Amyloid can also grow into fibrils, which arrange themselves into β-pleated sheets to form the insoluble fibres of advanced amyloid plaques. It is now clear that soluble oligomers and intermediate amyloid are the most neurotoxic forms of Aβ and that the severity of the cognitive defect in AD correlates with levels of oligomers in the brain, not the total Aβ burden [13,14]. However, it is also true that misfolded proteins of a different nature could lead to the formation of aggregates depending on the different maturation stage, going from monomeric peptides to intermediate oligomers and eventually to complex fibrillar structures. It is clearly emerging that this mechanism could operate all over the body and intriguing new data recently demonstrated that, for example, it takes place in the failing heart, providing an unanticipated cause for at least a subset of patients affected by idiopathic dilated cardiomyopathy . Moreover, the fact that sporadic AD is not associated with increased production of Aβ through mutated APP (amyloid precursor protein) or presenilins opens the possibility that aggregates of other proteic nature besides β-amyloid could also be pathogenetic.
Indeed, the origin of the different amyloid proteins deposited in the cerebral vasculature is still poorly understood. As most cell types are able to express APP and could potentially release Aβ, several hypotheses for mechanisms leading to parenchymal plaque formation and CAA (cerebral amyloid angiopathy) have been suggested. Three major hypotheses have been proposed: systemic, vascular and drainage. These mechanisms are not necessarily mutually exclusive, and it is conceivable that they might even occur at the same time.
The systemic hypothesis proposes that Aβ is transferred from blood to the vasculature. In vivo studies have shown the existence of a receptor-mediated bidirectional transport of Aβ across the BBB: RAGE (receptor for advanced glycation end-products) mediates the influx from circulation into the brain and is present on the luminal side of the endothelium; LRP-1 (low-density lipoprotein receptor-related protein-1) mediates the inverse process, localized at the apical face [16–18]. Interestingly, ApoE (apolipoprotein E), one of the most important risk factors associated with AD and CAA, is an LRP-1 ligand and modulates the rate of Aβ transport [16–18]. The exchange of Aβ between CNS, CSF (cerebrospinal fluid) and blood is an important process determining the concentration of Aβ in the brain. Alternatively, blood-borne Aβ may enter the brain if the BBB integrity is compromised, which indeed has been shown to be the case in AD and in the brains of transgenic mice for the pathology. Leakage of the BBB can also be mediated by Aβ itself by impairing endothelial regulatory function and endothelial cell death [19,20].
The vascular hypothesis proposes local production of Aβ from cerebrovascular cells. Several observations support this view. APP has been detected in extracts from vessels of AD brains and in vessels walls, coexisting with amyloid fibrils. It has been suggested that Aβ in CAA is derived from SMCs (smooth muscle cells) in the media of cerebral arteries. They are closely associated with vascular amyloid, and have been shown to express APP and to produce Aβ [21,22]. In addition to myocytes and pericytes, endothelial, adventitial and perivascular cells have been shown to express APP .
The drainage hypothesis suggests that neuronally produced Aβ drains with the ISF (interstitial fluid) along perivascular spaces of parenchymal and leptomeningeal vessels to cervical lymph nodes. CAA occurs owing to deposition of Aβ along these drainage pathways. In favour of the drainage hypothesis is the presence of CAA in transgenic mice that express human APP in the brain, in most instances, under the control of neuron-specific promoters .
It is thus conceivable that, with the onset of pathological conditions that render the artery walls more rigid, such as hypertension, the amplitude of pulsations is reduced and the passage of Aβ along the vessel walls is slowed. Such slowing allows the soluble Aβ in the vessel walls to precipitate as insoluble or β-pleated sheet amyloid resulting in CAA. The amyloid then blocks the elimination of Aβ, leading to increased concentration of soluble Aβ in the brain, which may by itself be associated with dementia, leads to precipitation of Aβ in the form of plaques, the development of tau pathology and neuronal and synaptic loss .
All of these hypotheses are intriguingly intertwined with haemodynamic challenges triggered by vascular factors, such as hypertension.
Murine models of hypertension-induced AD
As mentioned previously, most studies use animal models that mimic genetic AD, constituting only a small fraction of total AD cases. In contrast, our group has pursued the strategy to develop the only animal model of hypertension-related ‘Alzheimer-like’ pathology so far. In fact, we have found that mice that have been subjected to high blood pressure, by different stimuli, show accumulation of amyloid aggregates, the main histological finding of AD, as well as other AD-related features such as initiation of immune response and dysfunction of the BBB.
This animal model appears to be a unique opportunity, which enables us to dissect the molecular mechanisms that underlie AD development, focusing on the so far unexplained epidemiological data that high blood pressure is able to facilitate the onset of the disease. Thus this exclusive experimental strategy may allow picking molecular determinants of AD, which could be used as therapeutic targets against the disease, which would otherwise escape the research on AD.
To this aim, we have studied cerebral effects of a particular animal model of hypertension, mechanically obtained by TAC (transverse aortic coarctation) (Figure 1).
TAC experimental model
We have first characterized the acute impact of coarctation of the aortic arch between the two carotid arteries on CBF (cerebral blood flow), finding that the hypertension induced is this particular murine model determined an increase of CBF in the right brain hemisphere, with a concomitant reduction of the CBF in the contralateral one, within the first week of surgery . The acute blood pressure increase evokes trigger factors of neurodegeneration such as oxidative stress and inflammation . The derived brain injury is mainly localized in selected brain areas that control cognitive functions, such as the cortex and hippocampus, and could be a consequence of a defect in the BBB permeability . It is noteworthy that, even if these latter events are not enough to produce ischaemic/haemorrhagic injury, they are able to induce a subtler damage, altering the mechanisms that are fundamental for maintaining normal brain function [27,28].
We then extended the observation period of this murine model and also used another animal model of chronic hypertension obtained by subcutaneously infusing angiotensin II. In particular, we found that longstanding conditions of high blood pressure determine Aβ deposition in cortex and hippocampus as early as 4 weeks after TAC, as is evident from staining with anti-Aβ antibody and Congo Red . Interestingly, both approaches evoked the same effects, clearly indicating that the strategy used to reach the hypertensive condition does not influence the phenotype, but the chronic increase in blood pressure itself causes the brain damage observed . Furthermore, in this study, we demonstrated that passive immunotherapy, obtained by administration of anti-Aβ IgG, is able to rescue hypertensive brains from Aβ deposition and plaque formation, supporting the previously discussed hypotheses for this phenomenon .
In a subsequent study, we reinforced our previous observations by demonstrating also with Thioflavin-S, the occurrence of Aβ deposition, both in cerebral parenchyma and in proximity to blood vessels . In addition, we monitored over time cerebral blood perfusion with a novel technique based on contrast ultrasound imaging, revealing a whole brain hypoperfusion, suggesting that Willis's Circle distributes the asymmetric hypertensive challenge of TAC on the whole cerebral circulation . Interestingly, we found that this cerebral haemodynamic response to hypertension is also accompanied by profound changes in brain metabolism, as indicated by a selective down-regulation of the specific endothelial transporter for glucose (GLUT1) expression, pointing to a specific impairment of endothelial mechanisms of energy supply to the brain. Interestingly, the expression of GLUT1 in capillaries as well as the total surface area at the BBB available for glucose transport, have been found to be substantially reduced in AD brains [31,32]. Thus hypertension-induced hypoperfusion and hypometabolism contribute to the continuous shortage of pivotal energy nutrients observed in sporadic AD.
Role of neuroinflammatory response in neurodegenerative disorders
Among the typical pathological hallmarks described to occur in AD, a prominent activation of inflammatory processes is a keynote feature both in the brains of people with AD and in genetic animal models of the pathology . In this regard, the neuroinflammatory process has been strongly suggested as a possible driver of the pathology, as well as a result of an ongoing disease process . Interestingly, it has also been suggested that inflammation may be the link between hypertension and target organ damage , making conceivable the hypothesis that hypertension-induced neuroinflammatory processes could influence onset/progression of late-life dementia/sporadic AD.
The inflammatory process that takes place in most neurodegenerative diseases, like AD, consists mainly of elements of the innate immune response [36,37]. The brain microenvironment is constantly protected from noxious agents or injurious processes primarily by microglia, the resident immune cells of the brain, endowed with numerous receptors capable of detecting pathophysiological stimuli and promptly become activated [38,39]. Once activated, microglia may proliferate and undergo a morphological transformation from a ramified to amoeboid appearance. On the basis of the nature of activation stimuli, microglia become ‘aware’ of the problem and act in an appropriate manner to perform a defined task [38,39].
In view of that, it is becoming increasingly clear that, depending on the environment in which microglia are activated, they can either take on a ‘classically activated’ phenotype or an ‘alternatively activated’ phenotype [40,41]. Usually, alternatively activated microglial cells have an activation profile less inflammatory than the classical one [40,41].
Moreover, inflammation is a biological process that involves and recruits different cell types and molecules that may have many other functions besides inflammation. This could be the reason this response appears complex and highly variable. Indeed, most cells and molecules have plasticity sufficient to make them promote or inhibit an inflammatory response, depending on the environmental triggering factors.
The availability of mouse models for AD is ideal to dissect the different aspects of the neuroinflammatory process and to test whether specific immune and inflammatory mediators may affect brain pathological hallmarks, neurodegeneration and functional outcomes.
Interestingly, epidemiological studies have shown conflicting data about long-term use of anti-inflammatory drugs, and particularly NSAIDs (non-steroidal anti-inflammatory drugs), and the consequent reduced risk of AD . From here, the idea arose to treat AD models with drugs or substances that are able to interfere with the neuroinflammatory response, to better investigate this relationship. Pharmacological treatments in experimental models have the advantage that time points of treatment and dosage can easily be varied, and, on the other hand, the disadvantage that drugs may not be specific, and off-target effects are difficult to rule out .
Lots of studies have been carried out to investigate the effect of treatments of APP mice with NSAIDs [43,44]. Another important tool to influence the inflammatory processes associated with AD is the possibility of treating mice either acutely or chronically with the bacterial endotoxin LPS (lipopolysaccharide) , thus producing a stronger activation of innate and acquired immune responses. In particular, LPS shifts the kind of microglia activation (from alternatively to classically activated and vice versa). These studies have helped in dissociating microglial activation patterns and phenotypic markers associated with Aβ pathology.
On this issue, we have recently used our model of hypertension-induced ‘Alzheimer-like’ pathology to unravel the role of neuroinflammation in a spontaneous condition leading to Aβ deposition in brain. First, we found that hypertension itself triggers neuroinflammation before Aβ deposition in our ‘Alzheimer-like’ murine model . To dissect the role of neuroinflammation during the pathology, we used the two strategies to influence the immune system. Interestingly, the finding that only immune system activation, obtained with LPS, but not its inhibition, through NSAIDs, strongly reduced amyloid burden, suggested that stimulating inflammation in the appropriate time window may represent a promising strategy to limit vascular-triggered AD-pathology .
The brain haemodynamic challenge triggered by hypertension is an emerging pathological factor of chronic brain damage. Great steps forward have been made in unravelling the biological and molecular bases of hypertension and its powerful effects on the brain circulation. Nonetheless, numerous unsolved questions do still exist and this seems to be an intriguing chance to dissect at the same time the mechanism of sporadic AD. Animal models of hypertension (Figure 2) are the best means by which some of these intriguing questions can be addressed, allowing investigators to combine the use of genetic manipulation to achieve a spontaneous evolution towards ‘Alzheimer-like’ pathology.
Model of hypertension-induced ‘Alzheimer-like’ pathology
Models of Dementia: the Good, the Bad and the Future: A Biochemical Society Focused Meeting held at Robinson College, Cambridge, U.K., 15–17 December 2010. Organized and Edited by Stuart Allan (Manchester, U.K.), Christian Hölscher (University of Ulster, Coleraine, U.K.), Karen Horsburgh (Edinburgh, U.K.), Simon Lovestone (King's College London, U.K.) and Calum Sutherland (Dundee, U.K.).
amyloid precursor protein
cerebral amyloid angiopathy
cerebral blood flow
low-density lipoprotein receptor-related protein-1
non-steroidal anti-inflammatory drug
transverse aortic coarctation
This work was supported by the Italian Ministry of Health ‘Ricerca Corrente’ and ‘Cinque per mille’ to G.L.