Age‐related small vessel disease: a potential contributor to neurodegeneration in multiple sclerosis

Abstract

Multiple sclerosis (MS) is a chronic inflammatory demyelinating disorder of the central nervous system wherein, after an initial phase of transient neurological defects, slow neurological deterioration due to progressive neuronal loss ensues. Age is a major determinant of MS progression onset and disability. Over the past years, several mechanisms have been proposed to explain the key drivers of neurodegeneration and disability accumulation in MS. However, the effect of commonly encountered age‐related cerebral vessel disease, namely small vessel disease (SVD), has been largely neglected and constitutes the aim of this review. SVD shares some features with MS, that is, white matter demyelination and brain atrophy, and has been shown to contribute to the neuronal damage seen in vascular cognitive impairment. Several lines of evidence suggest that an interaction between MS and SVD may influence MS‐related neurodegeneration. SVD may contribute to hypoperfusion, reduced vascular reactivity and tissue hypoxia, features seen in MS. Venule and endothelium abnormalities have been documented in MS but the role of arterioles and of other neurovascular unit structures, such as the pericyte, has not been explored. Vascular risk factors (VRF) have recently been associated with faster progression in MS, though the mechanisms are unclear since very few studies have addressed the impact of VRF and SVD on MS imaging and pathology outcomes. Therapeutic agents targeting the microvasculature and the neurovascular unit may impact both SVD and MS and may benefit patients with dual pathology.

INTRODUCTION

Multiple sclerosis (MS) is a chronic inflammatory demyelinating disease of the central nervous system with neurodegeneration contributing to long‐term disability 112. After a phase of active inflammatory demyelination, translating into transient neurological worsening 37, patients enter into a progressive phase wherein accumulation of neurologic disability is driven by neuronal/axonal loss disproportionate to inflammatory activity 123, 211.

Age is one of the stronger predictors of entry into the progressive phase and accumulation of severe disability 38, 178, 207. This implies that factors known to be associated with aging, such as hypoxia163, mitochondrial dysfunction 43 and iron accumulation 215 may be important contributors to neuronal damage, and by extension, disability in MS 53, 124, 125, 191.

Cerebral small vessel disease (SVD) is another age‐related phenomenon and affects cerebral small cerebral arterioles, capillaries and venules 60, 152. SVD associates with microinfarcts, microbleeds and with periventricular white matter (WM) pathology. Age and vascular risk factors (VRF), such as hypertension, are the most important predictors of SVD 174. SVD associates with neurodegeneration in the elderly 229 and in young adults with VRFs 32, 64 and may contribute to age‐related neurodegenerative pathology seen in MS (Figure 1).

Figure 1.

Aging with multiple sclerosis (MS): review outline. BBB, blood–brain barrier; RR, relapsing–remitting MS; P, progressive MS.

There are several reasons why the interaction between MS and age related—SVD warrants further study, including: (i) MS patients have a longer life expectancy 128 reaching 60 years of age and beyond 187 and thus susceptible to the accumulation of vascular comorbidities; (ii) vascular comorbidities contribute to MS progression 129, reduced life expectancy 27, 73, 131, increase load of WM lesions and brain atrophy 97; (iii) in MS focal demyelination occurs in watershed areas 78, where hypoperfusion and tissue hypoxia are features 47, 53, 124, suggesting an important relationship between MS pathology and cerebral arterial perfusion; (iv) the chronic inflammatory milieu in MS may predispose to SVD and (v) drugs targeting the microvasculature may be beneficial not only in SVD but also in MS 33, 135, 148.

Herein, we present a comprehensive review on MS and SVD pathology. Through a critical analysis of clinical, radiographic and pathologic data, we explore the impact of these diseases on the cerebral microvasculature, their overlapping features and the possible additive effect of SVD on MS clinical outcomes and neurodegeneration. In so doing, we hope to shed light onto the striking age‐related accumulation of neurologic disability that characterises the later stages of MS.

CEREBRAL MICROCIRCULATION

The cerebral circulation can be classified according to vessel size into macro‐ and microcirculation 94, 98, 106 as illustrated in Figure 2. Microvessels do not provide sufficient collateral flow to perfuse tissue when a penetrating arteriole or venule is blocked 20. This contributes to deep WM vulnerability to ischaemia because its major blood supply is via long medullary arterioles 24, 217. Communication between the brain parenchyma and microvasculature is mediated through the neurovascular unit (NVU) that includes different structures including the pericytes that act as capillary sphincters and control blood flow 80. The normal microvasculature structural features and functions have been reviewed elsewhere 11, 89, 106, 113, 156, 182, 226 and are summarized in Table 1.

Figure 2.

A. Cerebral circulation is divided in macrocirculation (a) that includes the internal carotid and vertebral arterial systems and in microcirculation. The microcirculation is composed of pial and perforating arterioles, capillaries and venules. Pial arterioles give rise to superficial perforating arterioles that course centripetally entering the brain at right angles to its surface, branching and terminating as end arteries. Perforating arterioles also arise from the main arteries branches and irrigate the basal ganglia, deep WM and brainstem. Perforating arterioles lie within the perivascular spaces (PVS), where vessel wall components are in close contact with the astrocyte endfeet, and terminate as end arteries (ie, without shunts) in capillary beds. The capillary plexus drains into venules (not shown). B. Potential interaction between Small Vessel Disease (SVD) and Multiple Sclerosis (MS). Longstanding MS Brain MRI (FLAIR coronal section), showing focal and diffuse periventricular white matter hyperintensities adjacent to the lateral ventricles. A perforating artery affected by SVD is illustrated where enlargement of the perivascular space, thickening of the vessel wall with significant reduction of the vessel lumen is present. These vessel changes are associated with subsequent vessel occlusion leading to lacunes, vessel rupture causing microbleeds, hypoperfusion, blood–brain barrier damage and subsequent myelin break down. This could cause additional white matter damage and neuronal loss in MS.

Table 1.

Microcirculation: normal structure and changes with aging, small vessel disease (SVD) and in multiple sclerosis (MS).

Microcirculation	Structure	Function	Aging	SVD	MS
Pial arteries Perforating arteries	Surrounded by CSF and pia‐arachnoid and glia limitans. Three vessel wall layers: 1. tunica adventitia (mostly collagen and fibroblasts) 2. a tunica media (smooth muscle) 3. a tunica intima (single layer of endothelial cells and internal elastic lamina)	1. Nutrition 2. Cerebral blood flow regulation 3. Vasophylic neuroblast migration	1. Tortuousity 2. Thickened walls 3. Loss of elastin and smooth muscle 4. Reduced coverage by pericytes and perivascular nerve plexus	1. Severe thickening vessel wall 2. Arteriosclerosis/fibrinoid necrosis 3. Enlarged PVS 4. Vascular ectasia 5. Amyloid wall infiltration 6. PVS Hemosiderin deposition	1. Enlarged PVS 2. Scarce information
Capillary	Single layer of endothelial cells	1. Nutrition 2. Cerebral blood flow regulation 3. Immune response	1. Increased diameter 2. Basal membrane thickening, reduplication and vacuolization 3. Increased vessel wall hyalinosis and fibrosis 4. Reduced coverage by pericytes	1. Perivascular rarefaction 2. Increased endothelial permeability	1. Increase endothelial permeability, with protein leakage, red blood cell extravasation 2. Perivascular inflammation
Venules	Thin‐walled vessels: 1. tunica adventitia (mostly collagen and fibroblasts) 2. a tunica media (thin layer of smooth muscle) 3. a tunica intima (single layer of endothelial cells) 4. No valves	1. Metabolite clearance 2. Cerebral blood flow regulation	1. PVS collagenosis	1. Vessel wall and PVS collagenosis	1. Thickening of vessel wall, vessel tortuosity 2. Intramural fibrinoid deposition, wall reduplication 3. Iron deposition 4. Enlarged PVS 5. PVS collagenosis 6. Perivascular inflamation 7. Thrombosis
Neurovascular Unit	1. Endothelial cells 2. Pericytes 3. Vascular smooth muscle cells 4. Astrocytes 5. Basal membrane 6. Neurons 7. Perivascular macrophages	1. Immune response regulation (adaptive and innate immune responses regulation of leukocyte transfer) 2. Transport of substances 3. Blood flow 4. Angiogenesis regulation 5. Coagulation and fibrinolysis	1. Endothelial cell increased permeability 2. Reduced pericytes	1. Endothelial cell increased permeability	1. Changes in endothelial cell tight and adherens junction expression (occludine, claudins, caderins, zonula‐ocludens), with increase permeability 2. Basal membrane/extracellular matrix abnormalities

DIFFERENTIATING MS FROM SVD

Classical MS features detected by conventional MRI and its pathologic correlates

MS lesions occur in different areas of the CNS and can be detected in vivo using brain magnetic resonance imaging (MRI) and on postmortem examination. Typical conventional MRI and pathology features are summarized in Table 2.

Table 2.

Summary of core brain MRI and pathology features in MS and SVD.

	MS		SVD
	Conventional Brain MRI	Histopathology	Conventional Brain MRI	Histopathology
T2/FLAIR WMH	Focal lesions Sharply defined ! U‐ or S‐shaped juxtacortical ! Perpendicular to the lateral ventricles (Dawson fingers) ! Corpus callosum ! Optic nerve ! Spinal cord ! Pons periphery	Myelin loss (all constituents) Different degrees of axonal loss At high venule density and arterial watershed areas Active—rich in macrophages and lymphocytes Inactive—minimal macrophage infiltration	Focal lesions Watershed regions ! Central pons ! Sparing of the spinal cord ! Less frequent in the corpus callosum ! Sparing U fibres	Myelin loss Axonal loss
	Diffuse WM lesions Periventricular	Myelin loss with selective reduction of phospholipids	Diffuse WM lesions Symmetrical Mild periventricular WM	Selective loss of phospholipids and MAG with PLP preservation Loosening of the fibre network around “tortuous venules.” Minor arteriosclerotic vessel changes
			Irregular periventricular WM	Severe myelin loss and reactive gliosis Incomplete infarcts Fibrohyalinotic and arteriosclerotic vessels
			Punctate deep WM	Mild tissue changes surrounding dilated PVS Myelin loss and atrophic neuropil around fibrohyalinotic arterioles
			Deep partial confluent/confluent WM	Axonal loss and astrogliosis. Myelin, axon and oligodendrocyte loss, focal transitions to complete infarcts
T1	Transient hypointensities Widespread WM Persistent hypointensities Periventricular > juxtacortical	Increase extracellular space due to oedema Tissue loss	Areas >2 and <15 mm in the perforating arteries territory more in BG, pons (central), internal capsule and corona radiate	Irregular cavitations with scattered fat‐laden macrophages, reactive gliosis and myelin loss
SWI	Diffuse hyposignal BG Cortex	Perivascular iron deposition R* signal changes do not always correspond to iron deposition Iron in activated macrophages/microglia at the edges of WM lesions Iron precipitation in aggregates typical of microbleeds	Small rounded hypointensities visualized on brain MRI, 2–10 mm not well visualized on T2 Artheriosclerosis SVD—deep WM, BG and brainstem in arteriosclerosis‐related SVD Cerebral amyloid SVD—cortex, convexity/sulcus hypointensities	Microscopic bleeds, small lacunes or to hemosiderin containing macrophages in PVS Microscopic and macroscopic bleeds; β‐amyloid vessel wall deposition
PVS	Along convexity perforating medullary arteries	Perivascular collagenosis and inflammatory cuffs within these enlarged spaces	Arteriosclerosis SVD—lenticulostriate arteries entering the BG through the anterior perforated substance Cerebral amyloid SVD—pial and superficial perforating arteries	Enlarged PVS Perivascular collagenosis Enlarged PVS
CEL	Widespread in WM Different patterns: nodular, punctate, ring or open‐ring	Macrophage/microglia,lymphocytic infiltrates BBB disruption	Only in the context of acute stroke	Acute ischemic changes—red hypoxic neurons, inflammatory infiltrate

!Differentiating features between multiple sclerosis and small vessel disease, WMH, white matter hyperintensities; FLAIR, fluid‐attenuated inversion recovery; PVS, perivascular spaces; SWI, susceptibility weighted imaging; CEL, contrast enhancing lesions; MAG, myelin‐associated glycoprotein; PLP, proteolipid protein.

WM hyperintensities (bright) lesions (WMH) on T2/fluid‐attenuated inversion recovery (FLAIR) imaging sequences are key features of MS, corresponding to different degrees and types of pathology 51, 69, 102, 111, 114, 142. Myelin constituents are differentially affected in well‐defined WM lesions (plaques), diffusely abnormal WM (DAWM) or areas of normal WM (NAWM) on conventional MRI. Global loss of all myelin constituents is a key feature of plaques, whereas in areas of DAWM, myelin phospholipids and certain proteins such as myelin associated glycoprotein (MAG) are reduced with relative preservation of other constituents such as proteolipid protein (PLP) 114. However, in practice, differentiating histological changes in NAWM from DAWM may be challenging. Lesions also change over time, older patients with longer disease duration having more inactive than active lesions 69.

While no area is typically spared, there are regions of predilection for WM pathology 32, 37, 111, 134. The underlying substrate for this topographic predilection is not fully understood, but some observations suggest that the relationship between MS lesions and the cerebral vessels is important. Persistent T2 lesions are more common in central areas of the brain relative to the peripheral regions, whereas acute, transient contrast enhancing lesions are more evenly distributed throughout the cerebral WM 116. This suggests that additional factors, such as periventricular WM susceptibility to hypoxia, may contribute to persistent tissue damage 66 in some areas. Focal WM lesions occur at sites of high venous density and also in watershed areas of low arterial blood flow 78 further supporting a dynamic interaction between MS pathology and the cerebral circulation.

While cortical and deep gray matter pathology are important features of progressive MS that associate with clinical disability, 51 sensitivity to detect lesions in these areas on conventional MRI is relatively poor. This is particularly true in the cerebral cortex. As gray matter signal change is conspicuously absent in MRI diagnostic criteria, which concentrate on brain WM lesions (periventricular, juxta‐cortical) 15, we have focussed the current review on WM changes classically described radiographically in the disease. That being said, it should be acknowledged that non‐conventional MRI techniques and recent pathology studies demonstrate damage in the cerebral cortex and deep gray matter including demyelination, inflammation and neuronal loss 51. The relationship between these lesions and cerebral vasculature is yet to be explored and therefore is beyond the scope of the current review.

Age‐related small vessel disease

The aging brain

Brain volume and myelin loss occur during “normal aging” 195. At a cellular level, there is astrocyte and microglial hyperactivity, cellular senescence, stem cell exhaustion, altered intercellular communication, genomic instability, mitochondrial dysfunction and free radical generation, loss of proteostasis and dysregulated nutrient sensing 22, 34, 92, 120, 153 to name a few. These changes also affect the cerebral vessels 223 particularly arterioles and also veins 23, 62, 95. Cerebral vasoreactivity is impaired in aged brains and changes in the nitric‐oxide pathways 121, reduction of endothelin‐A and beta‐adrenergic receptors 181 have been reported. This age‐related degeneration of brain vessels may impair local perfusion 89. With aging, there is an increased BBB permeability and BBB‐pericyte injury 63, 141, 149. Other neuronal signalling changes, serotonin, acetylcholine and other vasoactive neurotransmitters may also affect cerebral blood flow control 55. All these factors may predispose the aging brain to increased vulnerability to ischemic/hypoxic injury. Structural age‐related microvascular changes have been summarized in Table 1.

Typical clinical, conventional MRI and pathology features of age‐related SVD

VRF such as hypertension, diabetes, smoking, hypercholesterolemia and obesity, associate with increased risk of cerebrovascular disease in general, even in younger populations 210. Clinical manifestations of arteriolosclerosis‐related SVD are varied, including stroke, cognitive decline, urinary and walking dysfunction 9, 28, 34, 118, 153, 155, 186, 189. SVD lesions can also be asymptomatic 93, 217. Amyloid deposition is also an important cause of age‐related SVD 31, 60, 77, 122. Although this principally affects cortical arterioles, the underlying subcortical WM may be affected and show signs of damage because these arterioles supply blood to this area. It is generally accepted that the spinal cord is relatively spared in SVD, though this is based on relatively scant literature 67, 205. Morphological vascular changes in age‐related SVD are summarized in Table 1 16, 52, 88, 186 and illustrated in Figure 2. Of the microvasculature, arterioles are particularly affected with vessel wall thickening and tortuosity. That being said, thickening of vein vessel walls and perivenous collagenosis have also been described 23, 95. Key imaging‐pathology features SVD are leukoaraiosis 64, 66, 74, defined as WMH on FLAIR/T2 brain MRI images without prominent hypointensity on the T1 images, lacunes 28, 56, 217, microbleeds 41, 74, 90 and enlarged perivascular spaces 119, 217 as summarized in Table 2. Leukoaraiosis occurs in 5‐10% of patients aged 20–40 years 32 and in up to one‐third of people aged 65–84 years 22. Lesion distribution patterns differ according to the different VRF 174. A selective loss of MAG and relative preservation of PLP is a feature of SVD WM abnormalities 12. The role of energy failure in SVD‐related disease is highlighted by the observation that there is a decrease in the number of mitochondria in leukoaraiosis 193.

VRF and SVD also associate with generalized and focal brain atrophy 18, 82, 115, 143, 147, 206 and reduced perfusion in NAWM 200, 213.

Imaging discriminators between MS from SVD

Several WM lesion locations and characteristics segregate more clearly with MS 14, 21, while others more with SVD 4, 13, 15, 16, 21, 44, 71, 107, 166, 199 (Table 2). The distinction between these two disease processes on imaging can be difficult in longstanding and late onset MS 184 where both conditions are more likely to coexist. Comparative quantitative measures have shown more heterogeneous lesions in MS 138, 199. Medial lemniscus T2 hyperintensity in the dorsal pons has been reported more frequently in SVD than in MS patients 79. SWI 100, 109 improves differentiation between MS and vascular lesions based on lesion location, perivascular orientation, and the presence of hypointense (rims around) lesions as well as detection of central veins within WM lesions 136. The exact specificity of central vein sign needs to be assessed in larger studies, particularly taking into account vascular comorbidities. Differentiating MS lesions without a central vein from ischemic lesions still remains difficult, particularly in diffuse WM lesions. Annual rate of SVD‐related lesion volume increase was similar to the rate of MS‐related lesion burden increase in secondary progressive MS observed in natural history studies or the placebo arms of treatment trials 140, 179, 180 and thus not very useful in distinguishing the two disorders. Only small studies using nonconventional imaging techniques have compared MS with SVD 39, 164. In magnetization transfer studies, normal appearing WM seems to be spared in SVD 164 in contrast to MS. Diffusion coefficient measurements 150, as well as magnetic resonance spectroscopy 96, may be useful in the differential diagnosis.

MS AND SVD: EVIDENCE FOR A POSSIBLE INTERACTION

Most of the current literature has concentrated on differentiating MS white matter change from that associated with SVD (Figure S1). In the previous section, we compared the two disorders and identified distinctive clinical and imaging features. Nevertheless, DAWM changes and brain atrophy are common in both disorders, particularly in older patients. SVD may coexist in longstanding MS and be responsible for additional brain damage and clinical disability. In this section, we will summarise the direct and indirect evidence of a possible interaction between MS and SVD. MS predominantly affects the veins and venules while SVD predominantly affects the arterioles. Despite this, factors such as aging, VRFs and chronic inflammation could predispose to microvasculature damage, including in arterioles, that leads to hypoperfusion and tissue hypoxia that contributes to the extent and distribution of MS‐related pathology.

“MS vascular theory” and microcirculation morphological changes in MS

Several vascular changes in the MS brain have been described over the years and a comprehensive review of the history of these observations has recently been published 160. Early pathology studies traced MS lesions to draining CNS veins and capillaries, surrounded by perivascular inflammatory cell infiltrates 2, 48, 170. Structural microvasculature changes reported in MS are summarized in Table 1 and include thickening of vessel walls and sometimes vessel thrombosis 1, 26, 70, 158, 212. The latter findings led to disappointing trials using anticoagulants and hyperbaric oxygen as a treatment for the disease 159 that were followed by reduced interest in the role of vessels in MS pathology 160. During the last decade, the hypothesis of venous insufficiency in MS patients was raised given the claim that venous blood flow alterations were more prevalent in MS patients 224, an observation not replicated in larger, more controlled studies 204. Recently using new MRI techniques, many MS lesions have been demonstrated in vivo to centre around veins of RR and PPMS 101, 105, 196 further supporting postmortem observations that the inflammatory process dominantly involves the veins and venules in MS. However, it is somehow surprising that the arterioles remain intact in longstanding MS and the current published literature does not rule out that structural or functional arteriole changes play a role in MS pathology.

In MS, enlargement of the PVS, particularly at the brain convexity, associates with brain atrophy 99 and perivascular protein changes 70, 139, 208 further pointing to a possible relationship between vascular pathology and MS disease severity. Structural abnormalities of PVS adjacent to venules have been mentioned above, but it is not clear if PVS adjacent to pial and penetrating arterioles are spared. Meningeal inflammation spreads into convexity PVS 85, but it is not always clear if this PVS inflammation occurs exclusively around veins or also surrounds penetrating arterioles, since this information is not always specified. The temporal profile of these structural vascular and perivascular changes in MS, in the context of age‐related changes, requires further study. Even if not directly related to MS, arteriole changes are predictable as part of aging and increased VRF, though they have not been systematically quantified so far.

NVU disintegration is increasingly recognised as a contributor to neurodegeneration 229 and has been described in MS not only in active demyelinating lesions but also in NAWM 3, 101, 154 (Table 1). NVU dysfunction can lead to clotting and fibrinolytic pathway abnormalities 11, may impair repair mechanisms such as angiogenesis in chronic MS 72, 83 and potentially affect CBF regulation.

Indirect evidence of vascular dysfunction in MS

Hemodynamic changes in MS

Several studies have found that patients with MS have reduced cerebral perfusion 45, 76, 162. Cerebral blood flow is reduced in nonenhancing WM lesions 188, cortical and subcortical gray matter 108 associating with higher disability scores and T2 lesion load 151. Chronic MS plaques are more prevalent in WM regions with lower relative perfusion 84 and hypoperfusion associates with T1 hypointensities 146. The observed hypoperfusion in MS appears to be a primary phenomenon and not merely a consequence of neuronal death 176, taking into account the fact that reduced cortical and deep gray matter cerebral blood flow is present in all disease course subtypes 59 even in the absence of corresponding volume loss 50, 151. However, large longitudinal studies are still needed to confirm that hypoperfusion precedes neurodegeneration. Reduced cerebrovascular reactivity has also been reported in MS and impaired dilator capacity of cerebral arterioles to vasomotor stimulation has been proposed as a possible contributor to MS hemodynamic changes 132.

Hypoxia in MS

MS inflammation and associated NVU dysfunction 3 may lead to tissue hypoxia 45, 110. In the animal model, experimental autoimmune encephalitis (EAE), acute tissue hypoxia develops rapidly in response to inflammation, triggers enlargement of vessel lumen and increased vessel number and is related to neurological deficits 47. Reducing tissue hypoxia may be an underestimated therapeutic target since it may reduce demyelination in animal models 53. There has been an increased interest in the relationship between angiogenesis and MS 72, 117, 177, but few studies have investigated the relationship between tissue hypoxia and structural microvasculature abnormalities in MS 47. MS WM lesions and classic ischaemic WM disease share some histopathological changes 49, 110. Using vascular distance maps, larger MS lesions tend to be further from vessels 104. Recently, it has been shown that MS lesions tend to accumulate not only in areas of high venule density but also in watershed areas 78 where hypoxia due to low arterial perfusion may contribute to and/or amplify mitochondrial dysfunction, previously documented in MS 125. It is not clear if only arteriolar and venule damage could explain these findings or if they relate to downstream changes at the capillary level, including pericyte and astrocyte dysfunction 49, 80. Indeed, pericytes when exposed to hypoxia and ischemia constrict and die 80. However, abnormalities at the capillary level, particularly pericyte injury, have been relatively neglected in MS and warrant investigation. Additionally, the link between vessel fibrin deposition and tissue hypoxia/ischemia needs to be clarified. Differences between acute focal demyelinating lesions and diffuse WM changes may associate to different vascular changes, the former to perivenular inflammation and the later with arteriolar changes.

MS and VRF

Epidemiological studies

Studies have shown that MS patients have a greater prevalence of VRF, such as smoking and increased body mass index, compared to the general population while other VRF, such as hypertension, diabetes and hyperlipidaemia do not differ significantly between groups 130. However, the presence of vascular comorbidity (diabetes, hypertension, hypercholesterolemia, heart disease and peripheral vascular disease) 197 associates with an increased risk of more severe MS‐related disability 46, 129, 131. Tobacco smoking has also been associated with increased disability and faster progression of clinical disability in MS in some 81 but not in all studies 144. Studies on other nonclassical VRF have been less explored 192. Higher homocysteine, also associated with SVD 91, has been found in MS when compared to healthy controls in most studies 198 and this could be due to vitamin B12 deficiency 167, 168. The presence of VRF also affects disability (eg, gait impairment) and increases mortality in people without MS. It is not clear if the effect on MS outcome is due to the additive effect of non‐MS pathology or due to worse MS pathology.

MRI studies on the effect of VRF in MS

Few MS imaging studies have taken into account the effect of VRF on lesion distribution and size and brain atrophy (Table S1) 81, 219, 228. North American MS patients with one or more vascular risk factor(s) have an increased lesion burden and more brain atrophy 97 compared to MS patients without VRF. However, how SVD contributes to these more severe imaging features is not clear. It is possible that increased T2 lesion load in MS patients with VRF is due to additive periventricular WM vascular lesions, lacunes and microbleeds, as a mere association of two disorders (Figure 2). Similarly, vascular risk factor associated atrophy could simply be additive to that in MS patients 115.

MS and cerebrovascular disease

Epidemiological studies

MS patients have reduced life expectancy that is estimated to be between 7 and 14 years compared to individuals without the disease 178. Though increased mortality can be directly related to MS, cardiovascular diseases are also important contributors 73. MS is associated with an increased rate of ischemic stroke, myocardial infarction and heart failure in the first years after diagnosis 35, 36. The reported cardiovascular disease excessive risk early after MS diagnosis might be due to surveillance bias and later in the disease course may relate to venous disorders in progressive MS, suggesting that immobilization may be a predisposing factor 173. The heterogeneity among studies on the incidence and prevalence of cardiovascular disorders in MS makes it difficult to fully understand the epidemiology of vascular comorbidity in MS 127. There is scarce information of global atherosclerosis burden in MS patients assessed clinically or with vascular imaging, though an increase in subclinical markers of atherosclerosis has been reported in a small group of MS patients particularly in those with reduced physical activity 161.

SVD imaging and pathology correlates in MS

As previously mentioned, MS is associated with haemodynamic changes. These could be related to MS but the contribution of additional factors, such as concomitant SVD, has not been evaluated. Though this may not be a significant concern in young MS patients with no VRF, it is somehow surprising that the potential presence of SVD has not been considered in older patients since SVD dramatically increases with age. Most of the imaging studies have excluded patients with previous symptomatic cerebrovascular disease (eg, stroke and ischemic heart disease) and recording of VRF has not been documented systematically.

After an extensive search (supplementary figure 1) for publications assessing SVD imaging characteristics in MS only two MRI studies with conflicting results assessed the prevalence of microbleeds in MS patients 58, 227. An imaging pathological study aiming to track iron in two MS brains disclosed iron precipitation in aggregates typical of microbleeds. Indeed, both cases were older than 60 years, and one had significant VRF and evidence of severe atherosclerosis at autopsy 10. An older pathology study also reported perivenular hemosiderin deposition related to MS plaques in 21 out of 70 MS cases, of which four had coexistent cardiovascular disease 1. There are no publications quantifying concomitant arteriolar SVD in MS. Differentiating venous collagenosis associated with leukoaraiosis from MS‐associated venous abnormalities may be challenging due to the lack of specific markers. In longstanding MS, basal ganglia T1 hypointensities along with diffuse WM changes have been reported, but no information regarding vascular comorbidities was provided 165. Since no imaging‐pathological correlation had been performed, the findings could be related to MS and/or SVD pathology.

The chronic inflammatory milieu in MS could contribute to cerebral small vessel damage and vascular damage may impact brain inflammatory response

Of interest is whether there is an interaction between MS and SVD pathology, and if there is in what direction this lies, since the chronic inflammatory milieu of MS could exacerbate SVD 207 and VRF could exacerbate MS pathology. Rheumatoid arthritis, a systemic inflammatory disorder, is associated with an increased risk of cardiovascular disease 8 and accelerated atherosclerosis 225. Inflammation can per se be deleterious to the vessel wall and is thought to be an important risk for systemic atherosclerosis 75, 216. The relationship between inflammation and SVD has not been sufficiently explored, but the following data suggest that it deserves more attention: (i) systemic inflammation measured by interleukin‐6 is associated with SVD 137; (ii) genes associated with inflammatory pathways are upregulated in SVD brains 171, 209 and (iii) prominent inflammatory infiltrates are found in some amyloid angiopathy subtypes 7. In MS, perivascular inflammation may increase cerebral vessel vulnerability to vascular risk factor‐related damage, thus contributing to increase of SVD.

MS and SVD may share common genetic and/or environmental factors

It is possible that common genetic factors simultaneously affect MS and vascular disease phenotypes 214. Apolipoprotein E, an important atherosclerosis risk factor 218 does not seem to affect MS clinical course in humans 25 though it may impact the inflammatory response in MS animal models 185. Mitochondrial genetic variants have been implicated in both MS and leukoaraiosis 194. Underlying variants of fibrinolytic systems can also produce an effect on MS inflammation 57 and also on cerebral ischemia 221.

Environmental factors may also potentially trigger and/or contribute to both MS and atherosclerosis pathology. Chlamydia pneumoniae infection has been associated with both disorders 40, but its role in their pathogenesis has not been demonstrated. Sodium chloride intake strongly correlates with hypertension 190 and has been associated with increased clinical and radiological MS activity 61 possibly secondary to the induction of Th17 lymphocytes as demonstrated in EAE 103. Smoking, another risk factor shared by both disorders 19, 133, associates with endothelial and BBB disruption 133, 145 and may cause brain damage through multiple pathways 30. The deleterious effect of smoking in cerebral vessels could explain increased lesion load in MS smokers when compared to non‐smokers (Table S1), but this not been investigated.

Potential interaction between MS and arteriosclerosis therapies

Current MS treatments, such as fingolimod 148, may have an antiatherogenic effect. Also alpha‐beta1 integrin (VLA‐4) blockade has been shown to be effective in reducing CNS inflammation in MS 169 but may also reduce neointimal growth following vascular damage 17, since this integrin is involved in vascular remodelling and atherosclerosis 87. Fumarates may have a cardioprotective effect 6 and may improve CNS response to hypoxia 220. On the other hand, statins, known to delay atheroma plaque progression and prevent ischemic cardiovascular events, may delay MS progression 33. The potential benefit and risks of aspirin in MS have been recently reviewed 203. Interestingly, a 1961 publication a trial comparing prednisolone, placebo and calcium aspirin, reported no deterioration in patients on aspirin, whereas there was clear clinical worsening in the other two groups 135. Moreover, antihypertensive drugs with a protective effect on cerebrovascular disease, such as amiloride 183, may exert a neuroprotective effect in progressive MS 5. Finally, biotin, an important cofactor for many mitochondrial enzymes that protects against hypoxia associated energy failure may reverse disability in progressive MS 201. These effects may relate to a pluripotent mechanism of these drugs that interfere with different pathophysiological cascades but could reduce the effect of one pathology, on the other hand. Either way to develop individualized treatments, comorbidities, such as SVD, should be taken into account and the influence of therapeutic interventions in these comorbidities should not be overlooked.

SVD AS A POSSIBLE CONTRIBUTOR TO NEURODEGENERATION IN MS

Older MS patients: less inflammation, more age‐related disorders

Although the pathological hallmark of MS is the presence of multifocal areas of demyelination with relative axonal preservation, imaging and pathological studies have shown that neuronal/axonal injury can occur early and associate with active inflammation and demyelination 65, 202 However, age and disease duration affect the inflammatory response in MS lesional and nonlesional WM and GM, all of which seem to decrease in older patients with a longer disease duration 67, 68. In these patients, neurodegeneration is related, in part, to not only on‐going low‐grade inflammation but also several mechanisms including increased energy deficiency, oxidative injury, hypoxia and exhaustion of functional reserve capacity 124. Age‐related pathology, such as Alzheimer's and vascular disease, may amplify all of these mechanisms and thus contribute to increased neuronal damage 67.

“Second hit” hypothesis: could SVD contribute to hypoperfusion and brain atrophy in progressive MS?

As previously mentioned, VRF 18, 29 and SVD 206 associate with brain atrophy and cognitive impairment. The exact relationship between SVD and brain damage is incompletely understood 152, 172, 217, but factors such as hypoperfusion/ischemia due to reduced vessel lumen size, impairment of perivascular lymphatic drainage, BBB dysfunction 215 and subclinical inflammation may lead to oligodendrocyte damage and loss of myelin causing WM lesions and neuronal loss 86, 152, 172, 217. NVU dysfunction leading to neuronal‐vascular uncoupling has been implicated in perpetuating tissue damage in ischemia 80. The presence of SVD in MS patients may represent the extra hit that hinders compensatory mechanisms. In this case, vascular dysfunction and hypoperfusion with consequent chronic hypoxia could contribute to neuronal death leading to slow neurological deterioration independent of relapses. Cerebrovascular disease has been shown to contribute to neuronal damage in neurodegenerative diseases such as Alzheimer disease 54, 229 and vascular cognitive impairment 157 promoting cycles of chronic hypoperfusion, pericyte and astrocyte dysfunction with BBB permeability changes, oxidative stress, inflammation and mitochondrial impairment 43, 124, 193. This “vasculo‐neuronal‐inflammatory” model of neurodegenerative diseases, centred in NVU dysfunction, could be applied to MS 229. In MS energy deficiency and tissue hypoxia due to mitochondrial dysfunction leads to ionic imbalance and axonal degeneration and this could be potentiated by concomitant SVD, in particular in watershed areas, where MS lesions tend to accumulate 124. Not only detailed mapping of MS lesions related to arterial and venous blood supply needs to be investigated 124, but also characterization and quantification of SVD, including scoring of arterial vessel wall changes, microbleeds, microinfarcts and its relationship to energy failure in MS is warranted.

Oligodendrocyte regeneration mechanisms can also be impaired due to dual pathology affecting ventricular–subventricular zone‐derived progenitor cells 126, since this area is frequently affected by SVD and MS. These mechanisms may particularly cause or potentiate hypoperfusion and brain damage in MS 49 not only affecting focal lesion topography 78 but also diffuse abnormal WM lesions, where there are imaging and pathologic similarities to SVD. Additionally, chronic inflammation in MS may predispose to microvasculature and NVU damage leading to abnormalities in fibrinolytic pathways 11, 57, impairing angiogenesis 117 and repair after ischemic injury and thus perpetuating neuronal injury.

Could SVD potentiate tissue damage related to acute inflammation in MS?

The effect of the interaction between SVD and MS on acute inflammation and subsequent neuronal damage has not been sufficiently explored but may contribute to the age‐related decline of recovery after a relapse 42. Vessel integrity is essential for many steps of the immune response, including leukocyte priming, activation and migration 222. The effect of cerebral age‐related SVD on each of these steps is not well characterized. Taking into account the previously described structural vascular changes, it is expected that the immune responses will be compromised to some extent. Cerebral SVD‐related BBB dysfunction has been shown to associate with endothelial cell and monocyte/macrophage activation 175, which could contribute to inflammation‐related neuronal damage in MS. As previously mentioned, there is tissue hypoxia associated with inflammation in MS 45, 53, 110, 117 and it is plausible that, if present in MS patients, SVD impairs compensatory mechanisms to acute inflammation‐related hypoxia, potentiating tissue damage.

CONCLUSIONS AND DIRECTIONS FOR FUTURE RESEARCH

Understanding the contribution of SVD to brain damage in MS patients and in particular its role in the neurodegenerative process is a high priority. Some common imaging and pathological features may be markers of SVD and MS interaction: diffuse periventricular WM lesions and enlargement of perivascular spaces. The challenge lies in identifying specific biomarkers that differentiate these two pathologies and ensuring that identified MS‐specific features have arisen from cohorts where VRF have been excluded. VRF and comorbidities are associated with faster MS progression and increased lesion load through unclear mechanisms. Future imaging research on brain volumetrics and WM lesions should take into account VRF and comorbidities in MS patients.

In MS, there is hypoperfusion and reduced vascular reactivity. Venule and endothelium abnormalities have been described but the contribution of arterioles and the NVU to these hemodynamic changes is still to be explored. Human imaging‐pathological studies would allow to better dissecting of the interplay between MS and age related vascular changes/SVD, MS animal models should be set up to look at the direct effect of vascular comorbidities on MS pathology and at potential common MS and SVD pathogenic pathways. Understanding the impact of SVD in MS is important in planning treatment trials, particularly in older progressive patients and may lead to better neuroprotective therapies in the future.

CONFLICTS OF INTEREST

None.

Supporting information

Additional Supporting Information may be found in the online version of this article at the publisher's web‐site:

Figure S1. SVD and MS interaction review. Search strategy: arterioscleros*, basal, basal ganglia hemorrhage, bleed*, brain*, cereb*, cerebral small vessel diseases, crani*disease*,disseminated, ganglia, haemorrhag*, hemorrhag*, intra‐crani*, intracrani*, intracranial, intracranial arteriosclerosis, intracranial hemorrhage, hypertensive, lacun*, leukoaraios*, leukoaraiosis, micro‐bleed*, microangiopath*, microbleed, ms, multiple, multiple sclerosis, multiple sclerosis, chronic progressive, multiple sclerosis, relapsing‐remitting, scleros*, small, stroke*, stroke, lacunar, vessel.

Table S1. Brain MRI studies on the impact of vascular risk factors on MS brain lesions.