This editorial introduces a special themed issue of Clinical Science on vascular causes of dementia, particularly small vessel disease (SVD), the commonest cause of vascular dementia and a central common pathway in the pathogenesis of stroke and dementia.
The themed issue, initiated by the Vascular Experimental Theme of the Dementia Platform UK (DPUK), will include critical reviews on several key topics from experts in the field–these are in progress–but we are also keen to receive proffered reviews and original work addressing molecular and translational topics that aim to unravel the complex and multifactorial processes that contribute to development and clinical expression of human cerebral SVD. It complements a workshop, to be held in January 2017, bringing together researchers in dementia and stroke to discuss the multifactorial mechanisms contributing to neurodegeneration acting around the complex vascular–brain interface, including models, their relevance and ways to ensure reliability.
These are interesting times for research in stroke and dementia, and for SVD(s) as central underpinning pathologies. Stroke and dementia both generate astoundingly large health and societal problems. Nearly 17 million people have a first stroke each year worldwide , of whom about one-third have died from the stroke by 6 months, one-third are dependent and only one-third can continue living independently. Stroke has been the second commonest cause of death in the world and the commonest cause of dependency in older adults consistently for many decades [2,3], and is a common cause of cognitive decline and dementia . SVD causes one-fifth of all strokes (lacunar ischaemic stroke; many haemorrhagic strokes in older people)  and worsens outcomes after stroke . There are estimated to be approximately 47.5 million people in the world living with dementia , in up to 45% of whom SVD is a major contributing or the sole pathology . This may be an underestimate since dementia may be under-diagnosed, particularly in low- and middle-income countries where healthcare provision is thinly spread and family support within a familiar environment may allow failing cognition to go undetected. Many governments have now made dementia a priority for research to accelerate development of prevention and treatment, with very large research spends.
What are the treatment options? The prevention and treatment of stroke have changed almost beyond recognition in the last 25 years, from nihilism, to becoming one of the most, if not the most, preventable and treatable neurological diseases. Consequently, stroke incidence is falling in most developed countries , rapid implementation of medical  and surgical treatments after transient ischaemic attack or minor stroke can reduce early recurrent stroke to negligible levels , and many more patients remain independent after acute ischaemic stroke thanks to stroke units, thrombolysis and thrombectomy [12,13]. There is now evidence that improved stroke prevention and treatment are leading to parallel reductions in diagnoses of dementia .
In contrast, although several different causes of dementia are now recognized (Alzheimer's disease, fronto-temporal, Lewy body, vascular), truly effective treatments for dementia have yet to emerge, and, apart from possibly adopting a healthy lifestyle  and managing modifiable vascular risk factors such as hypertension (both of which are generally good ideas anyway) , little can yet be done to prevent dementia, including once early signs of cognitive impairment are evident [17,18].
Dementia research is dominated by Alzheimer's disease which has largely focused on amyloid plaques and tau protein. However, despite decades of research, millions spent, hundreds of interventions tested (e.g. over 300 in the Tg2576 mouse model alone ), the lack of effective treatment or prevention has led to calls for wider thinking about Alzheimer's disease pathogenesis .
It has long been evident that most dementias in the elderly are mixed vascular and Alzheimer's disease . Human studies have shown repeatedly that the expression of cognitive impairment or dementia in life, for a given age and burden of Alzheimer's pathology, is most strongly influenced by the amount of vascular disease rather than by the Braak stage or burden of Alzheimer's disease pathology [21–23]. The recent finding of reductions in incidence of dementia paralleling improvements in stroke prevention and treatment [14,24,25], as well as data modelling exercises , appear to provide further evidence for important vascular contributions to dementia.
A clear mechanistic understanding is fundamental to the development of effective interventions. In stroke, this was accelerated by the advent of neuroimaging from the late 1980s, which allowed more accurate epidemiology to differentiate haemorrhagic from ischaemic stroke, tracking of brain responses to arterial occlusion and re-opening  and encouraged trials of thrombolysis and thrombectomy. Imaging with MRI and positron emission tomography (PET) also provided pathophysiological insights into stages of tissue injury in relation to blood flow in human ischaemic stroke . A similar approach, taking advantage of the range of neuroimaging tools which are now substantially more advanced, has been set out recently for Alzheimer's disease . In stroke, information from neuroimaging in patients has also helped to balance suggestions emerging from experimental models that were simply incorrect, for example that reperfusion worsened ischaemic tissue damage after acute ischaemic stroke . As we know now, reperfusion accelerates recovery  and is the basis of acute stroke treatment.
In stroke, there is a long-running healthy debate on the balance of information from experimental models compared with human clinical research . This has helped to shape awareness of the strengths and limitations of both approaches and led to improved standards for, and reliability of, experimental studies (e.g. the ARRIVE reporting guidelines, http://www.equator-network.org). Models of cerebral SVD, subcortical ischaemic stroke and monogenic SVDs have been reviewed extensively [33,34], including of vascular dementia and the limitations [35,36]. These have also provided valuable confirmatory mechanistic insights that are hard to obtain in humans.
Have dementia models received the same scrutiny? Most research into Alzheimer's disease pathophysiology is performed in rodents, but the rodent brain has proportionately much more grey and much less white matter than the human brain. As the interaction between the two is a fundamental component of human cognition, this must influence model relevance and interpretation in relation to human dementia pathologies. There has been less critical appraisal of Alzheimer's disease models but recent work is revealing . A critical appraisal of transgenic mouse models of Alzheimer's disease identified 427 papers describing 55 transgenic models, n=11 118 animals used in 838 experiments testing 357 interventions: blinding, randomization and sample size estimation were mostly lacking, but of even more concern, the actual outcome of the intervention was described for only 16 of 357 interventions tested (4%) . Studies did not give reasons for not reporting results of the other 341 (96%) interventions, but the information would be important to the field.
The wider thinking called upon to advance understanding of Alzheimer's disease pathophysiology and treatment  can equally be applied to vascular dementia  and small vessel stroke. It has long been assumed that microvascular brain damage is ‘ischaemic’ and due to vessel narrowing from atheroma or spasm . This is partly because the white matter lesions commonly seen on brain magnetic resonance scanning appear ‘white’, similar to the appearance of an acute infarct due to a large blocked artery. Although narrowing of the perforating arterioles does occur, there are also areas of dilation, and occluded arterioles have rarely been found in lacunar ischaemic stroke . Cerebral blood flow (CBF) falls with increasing white matter hyperintensity (WMH) burden in humans, but there are no/little data to indicate a longitudinal association between low CBF and worsening WMH ; instead, worsening WMH leads to lower CBF . An alternative hypothesis for SVDs suggests that the cerebral vascular endothelium–the blood–brain barrier (BBB)–becomes more leaky and allows plasma components (and cells) into and through the arteriolar and capillary wall, causing secondary damage in the vessel wall and surrounding tissues . The process is accompanied by inflammatory cell infiltrates in the vessel wall and perivascular tissues, secondary myelin and axonal damage. This theory is now gathering momentum [42–45], is supported by studies in patients ([46–48] and summarized in ) and experimental models, both of sporadic  and monogenic SVDs  and translational genetic studies .
There is still much to unravel. These topics will be addressed in this themed issue, discussed at the workshop, and we welcome additional proffered papers describing original research and reviews.
- Why does the BBB become leaky? Is this primary, or just secondary to factors such as hypertension, or a bit of both? What is the role of the pericyte in modifying BBB permeability [51,53]? Why do some people seem to be more vulnerable to developing SVD brain damage than others–is it just genes, or do environmental and socioeconomic factors contribute? What other factors can influence development of SVD brain damage? Why do some people present with cognitive impairments whereas others present with stroke?
- How does the BBB leakage influence perivascular tissues such as basement membranes, myelin and axons ? Is the myelin and axonal damage all secondary to the vessel changes, or is there an underlying vulnerability? For example, could the endothelium influence the integrity of myelin formation directly? Might this be developmental or acquired?
- Where does the inflammation come from? Is this primary or can systemic inflammation lead to secondary cerebral SVD? Might a similar process apply in Alzheimer's disease , where inflammation is also recognized to be an important component of the pathology?
- What is the role of hypoperfusion and/or ischaemia–ischaemia undoubtedly occurs and must add to brain injury, but does this reflect endothelial dysfunction and vascular dysregulation  or is it a simpler arteriolar narrowing ‘mechanical’ process? Is the vascular dysregulation of the aging neurovascular unit key to accumulation of Alzheimer's disease pathology in the brain and failure to clear toxic waste ? How does the perivascular space regulate brain fluid and waste clearance  and what goes wrong in Alzheimer's disease and SVD?
- Indeed, is Alzheimer's disease a vascular disease? Vascular risk and lifestyle factors increase the risk of Alzheimer's disease [59,60], as well as of vascular dementia, but the correct approach to managing vascular risk in older people with failing cognition is very unclear [61,62]. Alzheimer's disease is associated with altered vasoreactivity of the cerebral microvasculature , similar to that seen in patients with overt cerebrovascular disease . If so, efforts to improve function of microvessels might delay or even prevent clinical expression of Alzheimer's dementia.
- More practically, is there a relevant animal model of mixed dementia ? Do current models reflect mechanisms that are known to be related to vascular disease or dementia in humans? Where the model does not use a mechanism that reflects common pathogenesis in humans, but nonetheless appears to cause at least some aspects of the human pathology , is that telling us something useful about multiple disease pathways that is worthy of exploitation?
- Are humanized models such as induced pluripotent stem cells (iPSCs) useful in probing molecular mechanisms relevant to SVD? Can they realistically recapitulate the neuron–glial–vascular interface and interactions? If so, will it be feasible to define the important molecules that underlie disruption of cellular interactions in SVD as drug targets and use human iPSCs in drug screening?
Considering that a broad consensus may be emerging in the dementia field, and recognizing that multifactorial or mixed dementias are probably more common than pure dementia of any type, it may be more valuable to describe an individual's mixed dementia in terms of the proportion due to Alzheimer's disease or vascular pathology , because in turn this may assist in determining patient treatment and management. It may be a more realistic reflection of the messy world of multiple concurrent pathologies that accumulate in later life. In which case, the science of clinical research and laboratory methods, particularly the experimental models, need to reflect this reality.
Meantime, in the field of stroke, we can at least do a lot to manage vascular risk factors and, while evidence that individual vascular risk factor interventions prevent or delay progression of Alzheimer's disease dementia has so far been disappointing [8,68,69], the falling dementia rates in population-based studies  and improved cognitive function in at risk older individuals using a multifactorial ‘vascular health’ approach in a randomized trial , give considerable hope that joint scientific endeavours between vascular and dementia researchers at the highly integrated vascular–brain interface are likely to be effective in reducing the burden of dementia and stroke. This requires researchers in stroke, dementia and vascular biology to work together, benefitting from each other's knowledge, rather than pursuing parallel duplicative work, or, worse, repeating earlier erroneous work, which is the risk of remaining on home ground, however appealing that may be.
↵1 Dementia Platform UK (Vascular Experimental Medicine Theme).
- © 2016 The Author(s). published by Portland Press Limited on behalf of the Biochemical Society