Introduction

A key missing piece in dementia research is the elucidation of the neurovascular basis of cognitive impairment [230, 258, 308]. The term vascular cognitive impairment (VCI) is used to describe a wide spectrum of conditions characterized by cerebrovascular disease ranging from subjective cognitive decline to vascular dementia (VaD) (Fig. 1). While VaD remains the second most common type of dementia worldwide after Alzheimer’s disease (AD), its prevalence may be underestimated—especially in populations with significant concomitant small vessel disease burden, such as those in Asia [42, 48, 49]. Vascular factors may also exacerbate the pathology of AD [221], giving rise to the argument that vascular pathology could even be the most common contributor to dementia in elderly populations [222]. Moreover, VaD is associated with a high mortality rate and rapid stepwise disease progression (Fig. 1) [4, 5, 116]. Therefore, the identification of interventions to potentially benefit VCI patients and reduce its socioeconomic burden is of critical importance.

Fig. 1
figure 1

Stepwise progression to VaD. The road from risk factors to disease manifestation in VCI is a complicated one, as the multiple demographics, lifestyle and comorbid disease risk and mitigating factors interact through the progression from asymptomatic vascular lesions, cognitive impairment, and finally to VaD. Furthermore, these complex interactions give rise to several distinct cerebrovascular diseases underlying different forms of vascular brain injuries, leading to the clinical heterogeneity of VaD. However, regardless of the specific nature of vascular injury (occlusive, thrombotic, etc.), a state of chronic cerebral hypoperfusion can be considered to be the common etiological link. *See Table 1 for details and summary of supporting research

Full size image

The pathophysiology and clinical characteristics of VCI have been extensively reviewed [99, 281, 283, 304]. Briefly, VCI is characterized by brain lesions that occur due to vascular pathology which leads to diverse cognitive impairments. These lesions result in ischemic, hemorrhagic, and hypoperfusive states that can manifest as various clinical symptoms. Such vascular pathophysiological states in turn lead to a range of downstream effects and structural changes on the brain, including infarcts, lacunes, microbleeds, white matter injury and parenchymal lesions [108, 139, 140, 147, 228, 247]. VCI thus exists as a heterogeneous group of diseases which can be divided into various subtypes, such as multi-infarct dementia, post-stroke dementia and subcortical ischemic vascular dementia, each having unique features that may manifest clinically as dementia over time. While these disease subtypes provide an avenue to categorize the differing disease etiologies, they have a common set of risk factors, including demographic factors, lifestyle factors or presence of co-morbid conditions (summarized in Table 1). Each of these risk factors is known to independently contribute to the progression of cerebrovascular disease and is therefore associated with cognitive dysfunction and impact the progression to dementia. While there are several treatment options to manage VCI and their underlying risk factors as highlighted in Table 2, the root causes of the problem remain unclear. Hence, there remains several knowledge gaps in the understanding of VCI pathophysiology. Although several mechanisms have been reported to have roles in VCI progression, including neuroinflammation, oxidative stress-induced brain damage, neurodegeneration and brain atrophy [254], we still lack a thorough understanding of this complex disease, and even the aforementioned mechanisms have not been fully elucidated. Furthermore, the clinical signs and symptoms of VCI vary between patients given the heterogeneity of the severity and site of injury. Hence, a consensus on the underlying causes of VCI needs to be reached.

Table 1 Risk factors for vascular dementia (VaD)
Full size table
Table 2 Current pharmacotherapeutic options for vascular dementia (VaD)
Full size table

Chronic cerebral hypoperfusion (CCH) refers to chronically inadequate brain perfusion [56]. Given that CCH is intimately associated with various risk factors, pathophysiological processes and pathological lesions known to be involved in VCI (Fig. 1), we propose that CCH is the central underlying cause for the progression of VCI. We emphasize CCH as the underlying cause as it ties together some of the known mechanisms of VCI such as chronic inflammation, oxidative stress, neurodegeneration and brain atrophy. Age-related vascular changes lead to a state of global CCH and induce pathophysiological changes such as blood–brain barrier dysfunction, resulting in increased vulnerability to disease even in the absence of risk factors [278]. Recently, CCH was identified as the common feature observed in multiple subtypes of VCI [74, 283]. Furthermore, it was reported that global cerebral blood flow was significantly lower in VaD than in age-matched controls [227, 231] but significantly higher than in AD [227]. The reductions in cerebral blood flow is one of the earliest features observed from early VCI to VaD [130, 145, 224], and is consistently observed in different brain regions, such as a reported 31% decrease in cerebral blood flow in the frontal cortex and a 39% decrease in the parietal cortex [234]. Compromised cerebral blood flow in the deep white matter of the brain is also associated with hemodynamic ischemic injury, and therefore leads to a higher volume of white matter lesions (WMLs) [18, 286]. Moreover, cerebral hypoperfusion was shown to be a good predictor for WMLs in VCI patients [176, 198]. Given the above evidence, we postulate that CCH is a common driver of VCI pathologies such as WMLs, lacunes, infarcts, and subsequent cognitive impairment.

CCH is strongly associated with stepwise cognitive decline in VCI, where much of what is known about its clinical manifestation along the spectrum from normal to end-stage VCI comes from multiple longitudinal studies involving recruited subjects [128, 135, 200, 224, 268, 275, 286]. Neuropathological evidence, neuropsychological assessments and imaging are important adjuncts in many of these studies to ensure accurate study recruitment. In this review, we explore the concept that CCH is the main mediator of VCI pathology and cognitive impairment. We highlight the links between mechanisms and the development of structural neuropathological changes during VCI and provide a landscape of how each of these changes lead to the development of cognitive impairment in patients. Understanding how CCH drives VCI progression from pre-clinical to severe dementia is essential for both researchers and clinicians in diagnosing and developing novel therapeutics for early intervention.

Pathophysiology of chronic cerebral hypoperfusion (CCH)

In VCI, isolated instances of vascular injury may accumulate into widespread damage that overcome intrinsic repair mechanisms in areas critical for cognitive functions. There are several possible mechanisms that govern the transition from a physiological to a pathological state in the brain during cerebral vascular disease. In order to understand how this transition occurs, animal studies employing CCH are commonly used to model the underlying pathology of VCI. Mouse CCH models are generally generated by manipulating the common carotid artery, for instance, via bilateral common carotid artery stenosis [242] or asymmetric common carotid artery surgery [114], both of which could lead to reduced cerebral blood flow, white matter rarefaction, glial activation, as well as subsequent cognitive impairment. A cascade of molecular and cellular events has been shown to be involved in the pathogenesis of CCH including energy imbalance, oxidative stress, endoplasmic reticulum stress, mitochondrial dysfunction and inflammation (Fig. 2).

Fig. 2
figure 2

Pathological drivers of CCH-associated VCI. Several CCH-induced pathological drivers have been long associated with the pathogenesis of VaD including energy imbalance, inflammation, endoplasmic reticulum (ER) stress, oxidative stress and mitochondrial dysfunction. Decreased ATP production impairs ATPase pumps, results in neuronal depolarization, and leads to a deregulation in the glutamate homeostasis at the synaptic cleft and excitotoxicity in the brain. Low cerebral blood flow triggers the brain to utilize anaerobic respiration to produce ATP and this results in the accumulation of lactate within the neurons, leading to acidosis. Neuronal death following CCH is primarily attributed to the increase in the pro-inflammatory cytokine release during the chronic inflammatory response. Danger associated molecular patterns (DAMPs) released by the brain cells can also trigger glial activation and leukocyte infiltration, both of which can also produce pro-inflammatory cytokines. At the cellular level, increase in the reactive oxidative species from various sources, including the mitochondria, induces the oxidative stress state. While the increase in reactive oxygen species can contribute to the redox dynamics and hemodynamics imbalance, it can also induce chronic ER stress. Persistent ER stress leads to an accumulation of misfolded proteins and can have fatal effects on neuronal survival and integrity via the terminal unfolded protein response (UPR) pathway, as well as contributing to Ca2+ homeostatic imbalance. Mitochondrial deterioration causes a further decrease in the ATP production, leading to proteosomal dysfunction, as well as contributing to the frequency of mutagenesis events at the mitochondrial DNA. In a chronic state of CCH, these drivers are pathological and ultimately pave the way for downstream disease mechanisms

Full size image

Pathogenic mechanisms

Energy imbalance

During VCI, blood vessels in the brain including arterioles, veins and capillaries are partially occluded or hypoperfused [34, 145, 234]. Disruption to glucose and oxygen supply compromises the production of adenosine triphosphate (ATP) [118, 119]. The resultant state of energy imbalance impairs the function of ATP-dependent sodium–potassium pumps [82] which are critical for maintaining the resting membrane potential of neurons. As such, neurons spontaneously depolarize and release the excitatory neurotransmitter glutamate into the synaptic cleft. The excess accumulation of glutamate in the synaptic cleft is exacerbated by other defective ion pumps that fail to recycle the glutamate, leading to persistent depolarization and overstimulation of neighboring neurons [14, 38]. This excessive activation of glutamate receptors (i.e. NMDA and AMPA receptors) due to energy imbalance that results in neuronal dysfunction and death is called excitotoxicity, which has been reported to occur in chronic diseases such as VCI [265]. In order to compensate for the lack of glucose, the brain will begin to undergo anaerobic glycolysis, which produces lactate. Accumulation of lactate in the brain in turn leads to acidosis and acidotoxicity [143, 301].

Oxidative stress

Oxidative stress is defined as an environment where pro-oxidant species dominates over anti-oxidant species [95]. It is one of the central drivers of pathology in many diseases and it is implicated in the cognitive decline in VCI [26, 53, 166, 172]. Correspondingly, a reduction of circulatory antioxidant enzyme levels (e.g., superoxide dismutase, catalase) and antioxidant capacity (e.g., glutathione, ergothioneine) have been observed in VCI patients [80, 241, 297, 298]. In the brain, CCH causes a disruption in calcium (Ca2+) homeostasis which leads to acute and chronic production of reactive oxygen species [82, 166] from various sources, including electron transport chain, nicotinamide adenine dinucleotide phosphate oxidases (Nox) and nitric oxide synthase. In animal models of CCH, reduction of endothelial nitric oxide synthase expression [189] disrupts the vascular tone and exacerbates cerebral blood flow hypoperfusion [83]. CCH also increases Nox-1 expression in neurons, inducing apoptosis and contributing to cognitive impairment [53]. Oxidative stress also increases levels of circulating nitric oxide synthase inhibitor, reducing nitric oxide bioavailability, leading to vasodilation impairment as evident in cognitive impairment [67]. As the stiffness and pulsatility of the vessels increase, higher sheer stress is generated which disrupts normal continuous blood flow [19]. These vascular changes have been reported to be associated with reduced blood supply to white matter regions, thus precipitating the formation of white matter lesions and lacunes [266, 293].

Endoplasmic reticulum stress

Endoplasmic reticulum stress is emerging as a pathological mechanism in the etiology of VCI [195]. The endoplasmic reticulum is involved in the synthesis and post-translational modifications of molecules that are important in maintaining Ca2+ homeostasis [300]. Being the site of translation, protein folding and transport, disruptions to endoplasmic reticulum’s physiological function in the form of endoplasmic reticulum-calcium depletion, hypoxic conditions and oxidative stress, are known to result in misfolding and accumulation of unfolded integral proteins. Such stressors to endoplasmic reticulum activate an adaptive stress response pathway known as the unfolded protein response (UPR) [229]. This pathway involves three independent endoplasmic reticulum membrane-associated sensors which are protein kinase R-like endoplasmic reticulum kinase (PERK), inositol-requiring protein 1 (IRE1) and activating transcription factor 6 (ATF6) [300].

Under prolonged endoplasmic reticulum stress, cellular proteostasis becomes unsustainable, resulting in accumulation of misfolded proteins and activation of terminal UPR [120]. Attenuation of endoplasmic reticulum stress-induced apoptosis has been found to confer protection against ischemia and reperfusion injury [296]. More specifically, studies using neuronal models of vascular dementia have shown the contribution of zinc-induced neurotoxicity to its pathogenesis, upregulating endoplasmic reticulum stress-related genes like CCAAT-enhancer-binding protein homologous protein (CHOP) and growth-arrest- and DNA-damage-inducible gene 34 (GADD34) [263]. The same group also found that the endoplasmic reticulum stress pathway is involved in zinc-induced neurotoxicity thus implying its possible roles as both a cause and consequence in driving VaD [144].

Mitochondrial dysfunction

Given the high energy demands of the brain, mitochondria, as the ‘powerhouse of the cell’, play a central role in producing energy in the form of ATP. Mitochondria are also vital in regulating brain cell survival and death by controlling the movement of calcium ions between the cells and the extracellular surroundings. Reactive oxygen species produced by the mitochondrial energy-redox axis can signal for apoptosis when the cells are damaged [104, 159, 196, 218].

Mitochondrial dysfunction results in decreased energy production, thus altering cellular redox dynamics in the brain. Under these conditions, mitochondria begin producing an excess of \({\text{O}}_{2}^{ \cdot - }\) and H2O2 molecules in response to increased oxidation of proteins, phospholipids and DNA that pushes the redox equilibrium towards a pro-oxidative state [192]. There is also a global reduction in mitochondrial protein complexes over time [104, 159, 161, 310]. Therefore, it is not surprising that mitochondrial dysfunction has been observed in VaD [162]. In particular, mitochondrial damage such as increased mitochondrial bioenergetic deficits in the hippocampus plays important roles in the spatial learning and memory decline in both human patients and in CCH rodent models [16, 72, 162, 172]. Defects in mitochondrial metabolism lead to altered patterns in the mitochondrial respiratory rate, altered membrane potential, decreased pyruvate hydrogenase levels, increased oxidative stress as manifested by increased hydrogen peroxidase levels. However, mitochondrial dysfunction may not occur independently of other pathological processes but is commonly observed to overlap with other mechanisms such as oxidative stress and proteasome dysfunction [70].

Mitochondrial DNA contains genes that encode the cell’s mitochondrial energy production machinery, and defects or mutations in the mitochondrial DNA have been associated with age-related dementia and neuropathology [29, 31, 68, 102, 287]. The m.3316G > A mutation has been identified in early-onset VaD patients who did not manifest typical vascular symptoms [155]. Rather, the mutation causes a reduction in the activity of the respiratory chain complex I, and hence is associated with the well-established link that cerebrovascular damage increases when mitochondrial energy chain complexes are compromised [125].

Neuroinflammation

Inflammation involves a complex range of responses that are known to play a role in disease conditions. While inflammation is important in tissue repair and recovery, under disease conditions, chronic activation of inflammatory responses results in a destructive phenotype that is observed during disease development and progression [62]. Both acute and chronic inflammation have been implicated in cellular injury associated with a hypoxic state of VCI [136, 174, 217]. Under CCH, reduced blood supply disturbs cellular integrity, activates glial cells and recruits peripheral immune cells to the brain [23, 141, 309], causing death of neighboring cells and secondary tissue damage. Various molecular mechanisms such as activation of inflammatory pathways and inflammasome activation have been shown to play a role in inflammation during CCH.

Systemic inflammation serves as the initial signal of a stressed cell involving the release of damage associated molecular patterns (DAMPs), which are recognized by the pattern recognition receptors, namely Toll-like receptors (TLR) and NOD-like receptors (NLR) on neighboring cells, initiating an inflammatory response [82, 260]. Further studies have established the association of NLR family pyrin domain containing 3 (NLRP3) and Absent in melanoma-2 (AIM2), both of which are involved in inflammasome activation, with VaD and CCH [81, 177, 206, 207]. Regulatory pathways such as nuclear factor kappa B (NF-κB) and mitogen-activated protein kinase (MAPK) are subsequently activated to upregulate a wide range of inflammatory proteins [105, 146] including interleukin (IL)-1β, IL-6 and tumor necrosis factor (TNF) [24, 317]. These inflammatory cytokines can cause cell death, oligodendrocyte damage and demyelination [264, 305]. Attenuation of IL-1β production was shown to ameliorate hypoperfusion-induced brain injury in mice [206]. Microglia and astrocytes also release adhesion molecules and chemokines, which activate and facilitate leukocyte infiltration [15, 20, 115]. In a mouse model of cerebral ischemia, genetic deletion of the chemokine CCL2 has been shown to reduce brain injury via modulation of inflammation. Other proinflammatory proteins such as c-reactive protein are also upregulated to facilitate cerebral inflammation in VCI patients [77, 215, 232].

The complement system has also been implicated in stroke [168]. Complement proteins promote inflammation via glial activation and induce neuronal injury through the C5 activating membrane attack complex (MAC). Formation and deposition of C5b-9/MAC complexes damages the myelin sheath [175, 233], and abrogation of C5 protein reduces glial activation and white matter ischemia under CCH [167]. The central effector protein in the system is the C3 convertase enzyme complex [28, 197, 240]. It has been demonstrated that under CCH, microglial cells aggravate white matter injury via the C3-C3aR pathway in rat brains [311]. Together, the entire inflammatory process facilitates astrogliosis and scar formation, oligodendrocyte and endothelial cell dysfunction and blood–brain barrier disruption [248, 282, 313], leading to neurodegeneration, neurovascular dissociation and eventually structural damage to the brain. Therefore, inflammation serves as a critical mechanism that drives subsequent pathological changes in CCH. In summary, the pathological mechanisms of CCH covered in the above section lay a foundation to comprehend the complex mechanistic underpinnings of the disease. The pathological mechanisms of CCH include the involvement of multiple molecules and signaling pathways, especially those related to inflammation and oxidative stress.

Neuropathological features of CCH

In this section, we will examine how the pathogenic mechanisms described above contribute to the neuropathological features that have been described in VCI.

Glial activation

The term neurovascular unit describes the structural and functional interactions between neurons, glial cells, pericytes, extracellular matrix components and endothelial cells in the brain. The neurovascular unit maintains homeostasis within the brain microenvironment ensuring optimal conditions for function of neurons and other cells. During CCH, the entire neurovascular unit is affected by the combined effects of the pathological mechanisms described above that can cause reduced integrity of the neurovascular unit [238]. This results in a homeostatic imbalance in the brain.

Glial cells, especially microglia, drive inflammatory responses by releasing proinflammatory molecules. Increased number of glial cells are commonly seen in VCI patients especially at the white matter regions [245, 269]. Mechanistic studies using animal models have shown that upon CCH, activated microglial cells participated in both systemic and complement-activated inflammation; whereas attenuation of microglial activity reduces proinflammatory cytokine levels, increases myelin density and eventually improves cognitive performance [138, 311].

Astrocytes are also involved in the process of inflammation during CCH. Astrogliosis has direct influence on blood–brain barrier integrity and induces damage when constitutively activated astrocytes form glial scars or swelling at the end feet processes [88, 211]. With CCH, a study reported decreased astrocyte polarity and structural support to the endothelial cells eventually contributing to blood–brain barrier damage [127].

In the white matter, oligodendrocytes are the predominant glial cell type, and produce the myelin sheath around myelinated axons. As CCH damages oligodendrocytes and white matter, repair mechanisms are often impaired due to inflammation and loss of growth factors released by neurons, microglia and astrocytes. The myelin-independent axonal support from oligodendrocyte is also affected, causing significant axonal loss [93, 181, 313]. Upon ischemia, oligodendrocytes also release inhibitory proteins Nogo-A and MMP-9, preventing neuronal remodeling, and initiating a deleterious cascade within white matter to cause blood–brain barrier damage [93, 181].

Together, dysfunction in each of the components within the neurovascular unit can result in the disruption of brain homeostasis, which can eventually lead to neuronal loss and white matter infarctions at the grey matter and the deep white matter territory [35, 252].

Activation of cell death

Programmed cell death is a critical role in animal development and tissue homeostasis. Abnormal regulation of programmed cell death is associated with various human diseases including neurodegeneration. Different forms of cell death such as apoptosis, pyroptosis and autophagy have been observed in cerebral ischemia and reperfusion injury [76, 84, 141]. Of these, there has been ample evidence in the literature implicating apoptosis in VCI. In postmortem studies of VCI patients, apoptotic vascular cells were identified in the basal ganglia and subcortical white matter regions [103]. Apoptotic neuronal cells were also observed at cortical layers 3 and 5; and extensive ischemic lesions and axonal damage were observed in severe dementia [103]. Furthermore, protein expression and proteomics studies have revealed a decreasing anti-apoptotic proteins expression pattern in the cortex of VCI patients compared to controls [64]. Within regions of leukoaraiosis, significant increases in apoptotic oligodendrocytes were observed compared to adjacent white matter [33]. Mirroring the evidence seen in human patients, animal models of CCH present similar results, with increased markers of apoptosis observed. Specific changes observed in these animal studies include increased visualization of apoptotic bodies, increased expression of apoptotic proteins such as caspase 3, and reduced expression of anti-apoptotic proteins such as Bcl-2 [194, 243, 257, 272, 290].

More recently, other forms of cell death such as pyroptosis have also been investigated in human patients [24] as well as rodent models of CCH [206, 312]. Autophagy has gained interest as well, having been shown to be upregulated specifically in VaD [41] and in CCH rodent models [47, 51, 126, 306].

These findings, in relation to cell death mechanisms being implicated in the pathophysiology of CCH, reinforce the concept of degeneration over the course of VCI progression, and may suggest that therapeutic interventions in cell death pathways may prove effective in curbing the pathological progression of VCI.

Blood–brain barrier dysfunction

The blood–brain barrier is a selectively permeable barrier that separates the circulating blood from the parenchymal tissue. The endothelial cells of the BBB are characterized by expression of tight junction proteins between adjacent cells, reduced rate of transcytosis and other transcellular movement across the barrier into or out of the brain. This property of blood–brain barrier establishes a finely tuned microenvironment for the brain by maintaining homeostasis and defending against pathogenic infections. In CCH, increased blood–brain barrier permeability has been observed [40, 191, 214, 277, 292], and is associated with neuronal loss and white matter degeneration during disease progression [271]. Blood–brain barrier damage can be induced through increased excitotoxicity, inflammation and oxidative stress, which can contribute to further brain injury via mechanisms such as increased leukocyte infiltration.

Excitotoxicity causes a persistent activation of endothelial cells causing cell death and uncontrolled movement of substance across the blood–brain barrier [6, 59]. Separately, high Ca2+ levels in the cytosol of the endothelial cells can also lead to activation of cell death mechanisms and the increased propagation of Ca2+ levels through the intracellular sources of Ca2+ such as mitochondria, and endoplasmic reticulum [92] can relocate the endothelial tight junction proteins [32]. Presence of proinflammatory cytokines can directly damage the blood–brain barrier, reducing its integrity by inducing endocytosis of the tight junction proteins thereby weakening the tight junction assembly [96, 307]. The internalised tight junction proteins are directed to lysosomal degradation, leading to long-term blood–brain barrier dysfunction [250, 279]. Reactive oxygen species have also been implicated in the progression of VaD and could possibly contribute to the breakdown of the blood–brain barrier [211]. Increased reactive oxygen species in endothelial cells downregulates epithelial cadherin levels [2] and bioavailability of nitric oxide, leading to endothelial and blood–brain barrier dysfunction [55, 89].

Blood–brain barrier dysfunction, oxidative stress [112] and inflammation induce matrix metalloproteinases (MMPs)-mediated proteolytic degradation of the extracellular matrix [63, 223, 250]. In both VaD and experimental CCH, increased levels of gelatinases (MMP-2 and MMP-9) have been reported [46, 223], which are associated with the degradation of basement membrane and tight junction proteins of the blood–brain barrier [160, 276]. Blood–brain barrier damage also increased the size of the perivascular spaces leading to cellular damage of pericytes within, a common observation in VaD. Association of pericyte damage with CCH, white matter damage, neuronal loss and cognitive impairment is evident [185] although a direct link between pericytes to VaD has yet to be demonstrated.

The blood–brain barrier regulates immune cell infiltration by maintaining low levels of leukocyte adhesion molecules on endothelial cells with inhibitory effects derived from pericytes [289]. In VaD, inflammation increases expression of adhesion molecules and chemokines such as intercellular adhesion molecule-1 (ICAM-1) and vascular adhesion molecule (VCAM) in endothelial cells [259], facilitating leukocyte infiltration [165, 171]. Upon crossing the damaged blood–brain barrier, activated leukocytes cause irreversible damage to the blood–brain barrier and contribute to further release of pro-inflammatory cytokines and reactive oxygen species, which forms a vicious feedback loop of activating endothelia [50]. Although increased ICAM levels have been reported in post-mortem studies of VaD patients [180], evidence showing specific temporal dynamics of leukocyte movement into the brain is still lacking. Overall, the establishment of endothelial cell activation upon CCH not just damages the blood–brain barrier, it also causes a reduction in the resting cerebral blood flow, and thus further contributes to a hypoperfused state within the brain. The biggest challenge here is the myriad aspects of blood–brain barrier damage and its downstream mechanisms in the context of VCI and other neuropathologies that remain unknown.

White matter lesions

One of the major pathological hallmarks of VCI is the formation of white matter lesions (WMLs) [18, 286]. The white matter functions to connect and preserve neural circuit signaling, thus implying the clinical importance of WMLs as markers of brain dysfunction due to cerebral vessel disease. Pathologically WMLs represent processes ranging from demyelination, astrogliosis, axonal loss and venular damage. These are in turn a consequence of the combined effects of increased oxidative stress and inflammation in the brain induced by CCH and blood–brain barrier breakdown [150, 163]. Our group has reported that disruption to the structural integrity of white matter can cause cognitive dysfunction [107, 123]. Anatomically, the white matter region comprises of numerous nerve fiber tracts that are surrounded by myelin. During disease progression, demyelination may occur due to various reasons. Excitotoxicity, oxidative stress and inflammation lead to oligodendrocyte damage through the loss of cellular function, mitochondrial dysfunction and production of pro-apoptotic signaling proteins, eventually contributing to their death and white matter injury [179, 255], and causing primary or secondary myelin destruction in white matter regions [178].

There is limited evidence regarding remyelination at the site of white matter injury following CCH. Remyelination is uncommon during CCH as the chronic hypoxic and pro-oxidative states block the ability of oligodendrocyte progenitor cells from being able to differentiate into newly matured oligodendrocytes [86, 91], a process further impeded by surrounding damaged endothelial cells and scar-formation during astrogliosis [10]. The age-dependent Wnt signaling pathway, which plays a role in the oligodendrocyte progenitor cells differentiation, is also compromised in VCI disease states leading to further remyelination dysfunction [130]. Nevertheless, despite being vulnerable to vascular injury, evidence showed restoration of oligodendrocyte progenitor cells and oligodendrocytes after prolonged CCH (i.e. 1 month of bilateral common carotid artery stenosis), suggesting their potential regeneration ability [181]. Studies suggested that oligodendrogenesis and regeneration are facilitated by reactive astrocytes, which secrete trophic factors such as brain-derived neurotrophic factor in response to white matter injury [169, 183, 184].

Epigenetic and genetic mechanisms

While VCI progression is mostly sporadic in nature, some forms of VCI are known to be influenced by the interplay of genetics and epigenetics. With technological advancements, researchers have improved access to diagnostic tools which carry out high throughput genomic-based investigations. The following section highlights recent research in the genetic factors of VCI as well as emerging interest in the epigenetics of VCI progression.

Epigenetics

Epigenetics refers to the alteration of gene expression without altering corresponding DNA sequences [170]. Epigenetic mechanisms are driven primarily by environmental stimuli including stress, diet and other behavioural factors. Given that most of the risk factors of VaD are associated with lifestyle-associated conditions such as hypertension and diabetes mellitus, the role of epigenetics seems to be critical in explaining the pathophysiology of the disease [203]. In fact, there are several lines of evidence for epigenetic contribution in the pathophysiology of dementia in general. These include studies where DNA methylation and hydroxymethylation were observed to be significantly reduced in the hippocampus, entorhinal cortex, cerebellum, and prefrontal cortex of AD patients compared to healthy controls [170]. Epigenetic modifications in AD neuropathology have been increasingly studied with the findings also implicated to other neurodegenerative diseases [79]. While there is limited evidence for the role of epigenetics specifically in VCI or CCH [203, 237, 299], this is likely due to the nascent nature of this topic, pointing to the need for further studies. Nevertheless, with epigenetic changes seen as drivers of pathological conditions, they may be regarded as biomarkers for early disease detection [199]. As such, further study of epigenetics would provide insights into VCI and perhaps aid in the stratification within VCI.

Genetic mutations

Certain forms of VCI onset and progression are known to have a familial component though the majority are sporadic cases [173]. Monogenic influences of tissue responses to VCI include NOTCH3 mutations causing cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL), a rare form of cerebrovascular disease. The Notch pathway is important in the regulation of cell fate [121]. In particular, the Notch 3 receptor-mediated pathway is involved in the vascular smooth muscle survival [173]. In CADASIL, the mutation of the NOTCH3 gene occurs within the epidermal growth factor—like repeat domains in the N terminal of the receptor. Brains of CADASIL patients manifest an aberrant oligomerization of mutant Notch 3 proteins, leading to altered protein–protein interactions [165]. The pathophysiology of CADASIL still remains unknown, but the NF-κB pathway has been reported to play an essential role in the inflammatory responses in the CADASIL-associated angiopathy. NF-κB promotes the expression of genes coding for cytokines that leads to an amplified vascular inflammation level and hence vascular dysfunction [149]. Notch 3 misfolding phenomenon can cause an increase in free radical production in the brain, although the levels produced may not be directly pathogenic [39].

There are also overlapping genes with AD which are known to be involved in the VCI pathogenesis, namely the presenilins, the amyloid precursor protein (APP), and the apolipoprotein E (APOE) [158, 220]. It has been reported that the presence of even a single allele of the APOE4 variant could be a potential risk factor for progression of VCI [220], thus providing evidence for a commensal interaction between AD and other CVD conditions.

Unravelling the potential of early detection and intervention strategies

Currently, the treatment options for VCI remain sparse. Understanding the underlying pathophysiology of VCI through CCH provides critical insights to the discovery of biomarkers and targets for disease-modifying treatments.

The identification of specific biomarkers for VCI will be critical for more specific and sensitive diagnosis. These biomarkers may allow for early detection of VCI in at-risk patients. While the diagnostic criteria for VCI is based primarily on neuroimaging, blood-based biomarkers are nevertheless useful as surrogate disease indicators. Many blood and cerebrospinal fluid biomarkers have been identified over the years. Several proinflammatory molecule, such as C-reactive protein, IL-1α and IL-6, have been proposed as potential biomarkers [57, 133, 288]. Given that the increase in plasma level of inflammatory proteins precedes cognitive impairment in VCI, the identification of proinflammatory proteins in early stages of the disease not only offers prognostic advantage but also possible therapeutic intervention [77]. Other than inflammation, the classic marker for blood–brain barrier dysfunction, matrix metalloproteinases, has also been reported multiple times in VCI patients and found to be an early biomarker for cognitive dysfunction [73, 78, 191]. While evidence for oxidative stress in VCI patients are limited, it has been shown that oxidative stress is increased in mild cognitive impairment and AD [44, 210]. Our team has contributed to the field in establishing several possible blood biomarkers for white matter hyperintensities and microinfarcts in clinical cohorts of VCI patients such as serum hepatocyte growth factors, IL-8 and growth differentiation factor-15 [45, 106, 122, 314, 315].

Beyond specific aspects of pathophysiology, several multi-purpose therapeutic interventions have been proposed. The basis of these therapeutic interventions is built upon the evidence that cerebrovascular injury is not always progressive but may instead be reversible. For instance, white matter hyperintensities which are indicative of white matter lesions may regress and be amendable to treatments [87]. Therefore, to the extent that cerebrovascular disease such as white matter hyperintensities is related to CCH and can contribute to the risk of VCI development, markers which allow for the early identification of these lesions may enable early mitigation of cerebrovascular disease and in turn, VCI development. In a similar vein, a deep understanding of the molecular underpinnings of CCH which are relevant to cerebrovascular disease allows for the identification of potential treatment markers as well as drug targets. In the latter case, this then facilitates a potential for development of disease-modifying treatments.

Given that chronic diseases such as VCI are linked to diet and lifestyle factors, interventions at this stage are important for managing the disease. Recent studies have found association of VCI with dietary habits, shedding light on using intermittent fasting as a possible treatment for VCI [201, 280]. Intermittent fasting has been shown to improve cognitive ability, neurotropic factor production, synaptic plasticity, mitochondrial biogenesis, and has also been shown to ameliorate vascular pathology and cognitive impairment in rodent VCI models [11, 85, 110, 213, 237, 273]. Additional systemic beneficial effects of intermittent fasting include attenuation of inflammation, oxidative stress, mitochondrial dysfunction and DNA damage [65]. Separately, in clinical studies on VaD, increasing physical activity has been suggested to reduce the risk of dementia manifestation, albeit not with entirely consistent results due to lack of standardized methods [1, 71, 249, 285].

These approaches are not only potentially disease-modifying at the pathophysiological level, but they also serve as a preventive strategy to mitigate an at-risk patient’s risk of disease. More studies are required for a more robust conclusion in order to delineate the role of intermittent fasting and exercise in VCI clearly [124, 156, 157, 235]. Notably, our team have been investigating the effect of multiple lifestyle interventions on the prevention of cognitive decline under the Singapore Geriatric Intervention Study to Reduce Cognitive Decline and Physical Frailty (SINGER) [52], which is an adaptation of the pioneering Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER) [151]. Lifestyle interventions provide a promising potential in managing VCI and reframing the public health perspective of the disease.

Summary and future directions

VCI is now a widely accepted term introduced to embody the entire spectrum of vascular-related cognitive alterations or cerebrovascular disease-related burdens that can manifest into cognitive impairments [100]. Cognitive deficits associated with VCI include slower mental processing and impaired executive functioning such as poor planning, poor judgement and poor decision-making. Non-cognitive behavioural manifestations include apathy, anxiety and even depression are also common. VCI has been gaining interest in the field as it is potentially preventable, prior to reaching the end-stage dementia [132, 294]. Given such emphasis on early detection and diagnosis in the field, there is a need to better understand the pathophysiology of VCI. As reviewed above, current experimental evidence indicates that a chronic state of hypoperfusion in the brain drives the various pathophysiological mechanisms and structural changes in the brain. CCH therefore holds promise in shedding some light on the molecular and mechanistic underpinnings of VCI.

As reviewed above, there are several common mechanisms that occur during the progression of CCH-induced injury such as energy imbalance, oxidative stress, endoplasmic reticulum stress, mitochondrial dysfunction and inflammation (Fig. 2). These mechanisms drive the downstream structural neuropathological changes in the brain including glial activation, cell death activation, blood–brain barrier breakdown and white matter lesion formation. The pathological features begin from a hypoperfused state and can coexist and interact to adversely influence cognitive function as reported in animal models [25, 27, 43, 166]. This suggests that the effects of CCH on cognition are mediated by mechanistic drivers and structural changes in the brain. Indeed, CCH may be the earliest, insidious indicator of VCI, while brain atrophy and white matter lesions may occur downstream from CCH as more dynamic and detectable changes.

It is exciting to witness the field of VCI rapidly expanding and moving towards sharper definitions and deeper insights into underlying mechanisms. The heterogeneity of the disease is widely recognized to be due to the complex interactions between vascular injuries and risk factors that are involved prior to, and during disease manifestation. Across these subtypes, variations also exist at the clinical, neuroimaging, and pathological levels. Yet, a strong argument may be made that all subtypes of VCI include a CCH state, which we believe to be the main driver for subsequent pathological progression. Prolonged cerebral hypoperfusion may therefore serve as the transition from the at-risk state to the VCI state. The observational and experimental evidence from CCH models presented in this review help reinforce the importance of CCH as a critical feature in our efforts to unravel the underlying molecular mechanisms of VCI. Further identification of specific biomarkers of CCH may provide the rationale for the evaluation of these markers in the clinic which can bring us closer to detecting VCI at an early stage as well as introduce treatment options which may delay disease onset or slow disease progression.