The Ties That Bind: Glial Transplantation in White Matter Ischemia and Vascular Dementia

Abstract

White matter injury is a progressive vascular disease that leads to neurological deficits and vascular dementia. It comprises up to 30% of all diagnosed strokes, though up to ten times as many events go undiagnosed in early stages. There are several pathologies that can lead to white matter injury. While some studies suggest that white matter injury starts as small infarcts in deep penetrating blood vessels in the brain, others point to the breakdown of endothelial function or the blood–brain barrier as the primary cause of the disease. Whether due to local endothelial or BBB dysfunction, or to local small infarcts (or a combination), white matter injury progresses, accumulates, and expands from preexisting lesions into adjacent white matter to produce motor and cognitive deficits that present as vascular dementia in the elderly. Vascular dementia is the second leading cause of dementia, and white matter injury–attributed vascular dementia represents 40% of all diagnosed dementias and aggravates Alzheimer’s pathology. Despite the advances in the last 15 years, there are few animal models of progressive subcortical white matter injury or vascular dementia. This review will discuss recent progress in animal modeling of white matter injury and the emerging principles to enhance glial function as a means of promoting repair and recovery.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13311-022-01322-8.

Introduction

Stroke affects 15 million people each year, worldwide. Of these, 5 million suffer from permanent disability, and approximately 5.5 million people live with varying levels of stroke-related disabilities [1]. The annual prevalence of stroke is estimated to be about 30.7 million worldwide, with 12.6 million people having moderate to severe disability following stroke. Subcortical white matter ischemia (WMI) constitutes up to 30% of all stroke subtypes [1, 2]. Significant white matter injury is present in most human stroke and occurs both in the stroke lesion site itself and as concomitant of the vascular risk factors of stroke, separately in the white matter [3]. Indeed, independent of its affects as direct stroke, WMI accumulates to cause vascular dementia. Clinical-pathological studies have highlighted the role of subcortical WMI, not only as a primary cause of cognitive and motor impairment, but also as an adjuvant to the expression of dementia caused by other factors, including Alzheimer’s disease (AD) and other neurodegenerative pathologies [4–6].

WMI and vascular dementia occur in a distinct cellular environment in the brain. Unlike large artery stroke, which spans brain regions of gray and white matters, WMI is localized to the cellular constituents of this brain area: endothelial cells, astrocytes, pericytes, oligodendrocytes, and oligodendrocyte precursor cells (OPCs) (Fig. 1). Axons course through the white matter, communicate with these non-neuronal cells in a form of “axo-glial signaling,” and shape white matter structure and function. These are the cellular constituents of the neurovascular niche of the white matter. WMI and the biology of white matter repair is understudied and remains an important topic in the field. A key concept in this emerging area is that glial cells, and particularly astrocytes, form central hubs of molecular signaling among the cells of the white matter neurovascular niche. Glia, derived of course from the Greek and meaning “glue,” tie elements together. This review will discuss an approach to enhance glial function in WMI as a means of promoting repair and recovery.

Fig. 1.

Overview of white matter injury progression in humans. White matter injury is initiated by vascular dysfunction, and it damages primarily the gliovascular unit, followed by axonal degeneration and demyelination. White matter lesions accumulate over time to produce motor and cognitive decline that presents as vascular dementia in elderly

Subcortical White Matter Ischemia and Vascular Dementia

Subcortical WMI is a distinct process from “large artery stroke,” which occurs with acute thrombotic or embolic occlusion in the anterior, middle, or posterior cerebral arteries or their branches [2, 7–9]. Subcortical WMI is a progressive vascular disease that starts as a primary abnormality associated with the vasculature of the white matter (Fig. 1). This disease has a pathology that appears heterogeneous—that is, several pathological processes participate [9] (Fig. 1). There is clear evidence by MRI that this disease starts in many circumstances as small infarcts in deep penetrating blood vessels in the white matter, and then this ischemia progresses locally and accumulates [10–12]. Some studies suggest that the disease is a primary breakdown in endothelial function or the blood–brain barrier, leading to leak of plasma proteins and local inflammation [9, 13, 14]. The arterioles that supply the subcortical white matter may also narrow and become fibrotic with age and with co-morbid conditions, such as hypertension, and produce local ischemia [7, 15]. These different pathologies lead to distinct types of animal modeling of the disease (below). Whether due to local endothelial or BBB dysfunction, or to local small infarcts (or a combination), the white matter lesions accumulate to produce hemiparesis with incomplete recovery, gait abnormalities, verbal processing deficits, and cognitive decline and difficulties in executive functioning that present as vascular dementia in elderly [16–18]. WMI is strongly age-associated. The incidence of subcortical WMI and VaD increases with age and is expected to increase in prevalence with the aging population [19]. In a concerning finding, over the age of 80, imaging suggests that all people will have MRI evidence of white matter ischemic lesions [20].

Clinical studies of human white matter disease indicate that this pathology progresses locally into adjacent white matter and does so over years as a chronic disease. New white matter strokes develop within pre-existing lesions and generate adjacent lesions in 71% of cases [21]. White matter ischemic lesions progress even under the controlled conditions of clinical trials [2, 22]. Repeat MRI imaging shows that new white matter lesions enlarge on average 14% per year [2]. The expansion of the white matter lesions correlates with the evolution of the cognitive impairment [23], new small strokes in this progression of white matter ischemia cause a steeper decline, especially in motor speed and executive functions [24]. This process of WMI-attributed vascular dementia is common, as it represents 40% of all diagnosed dementias and co-occurs with and accelerates Alzheimer’s pathology to represent the common disease entity of mixed dementia: dementia with mixed etiology—vascular and Alzheimer’s component [25]. These observational studies in human white matter disease and vascular dementia indicate two important biological principles: that this disease progresses locally and does so over a long time period. The local tissue adjacent to the white matter lesion is a target for a biological therapy to prevent disease progression, and the long time course of this disease means that strategies that induce repair of the injured white matter, or address this disease progression, have a long time window for delivery.

Animal Modeling of Subcortical White Matter Ischemia and Vascular Dementia

Despite an advanced pre-clinical literature of animal modeling in large artery, “gray matter” stroke, and animal models of Alzheimer’s disease, there are few animal models of progressive subcortical WMI or VaD. Experimental models of WMI and VaD have been difficult to develop because of the relatively smaller volume of cerebral white matter in rodents as compared to humans. There are three basic approaches to the rodent modeling of white matter stroke or vascular dementia (Table 1). Global ischemia models have been applied to WMI, such as progressive occlusion of the carotid arteries (Table 1). These do produce white matter lesions, but also ischemic damage in cortex, retina, and subcortical structures. While still a matter of debate, significant global brain hypoperfusion does not appear to be driver of vascular dementia [26–28]. Furthermore, by damaging the neurons that send axons to the white matter, and not just the white matter itself, this model of global brain hypoperfusion in rodents interferes with selective molecular, cellular, and behavioral studies of WMI. Signaling between axons in white matter and the cells of the white matter neurovascular niche (astrocytes, oligodendrocytes, OPCs, pericytes, and endothelial cells) is robust, is bilateral, and mediates key cellular events in the normal adult white matter environment [29, 30]. The activity of white matter axons, for example, stimulates OPC proliferation and differentiation [29, 31]. A study cannot infer, for example, that disordered oligodendrocyte precursor (OPC) responses to WMI or altered astrocyte or pericyte function is from the white matter disease if the axons that project to this white matter are also sick or dying because their neuronal cell bodies were hit with ischemia in that particular white matter model. A second model of WMI or VaD embraces the concept that this disease may not so much be due to the accumulation of small focal ischemic events in the white matter, but instead due to a more generalized dysfunction of endothelial cells and microvascular function in the subcortical white matter (Table 1). This approach utilizes the spontaneously hypertensive rat strain, or its derivatives, such as the spontaneously hypertensive-stroke prone rat line (SHR-SP). The SHR-SP spontaneously develops severe hypertension, renal failure, and brain abnormalities with end-organ damage throughout the body and early death [32, 33]. This rat model has the addition of the second most common contributing factor to subcortical white matter hyperintensities, hypertension (the first being age), and produces endothelial dysfunction as an early component [15]. However, the SHR-SP has widespread cortical and subcortical damage, neuronal loss, and reactive tissue scarring early in its disease course before MRI and behavioral indications of deficits [34]. The occurrence of widespread and particularly neuronal damage, outside of the white matter indicates that, just as with global brain hypoperfusion, there are sick and dying axons in the white matter from distant neurons, and this will directly affect the neurovascular unit of the white matter outside of direct WMI. A further area of concern for the SHR-SP is that it is a complex and incompletely characterized genetic mutant, with some of these mutations affecting gene systems that are involved in ischemia and tissue repair [35–37].

Table 1.

Summary of white matter Injury models. ~ Damage in other brain structures interfere with the behavioral readouts

WMI models		White matter volume	Exclusive WM damage	Vascular dementia driver	Behavioral readouts	References
Global ischemia		X	X	X	~	[26–31]
SHR-SP (hypertension)		X	X	✔	~	[32–37]
Vasoconstriction	Acute	X	✔	✔	X	[39–42, 44]
	Subacute/chronic	X	✔	✔	✔
WMI large mammals (piglets and primates)		✔	X	X	X	[51–58]

In addition, in a recent study, Quick et al. [38] demonstrated that while hypertension can lead to endothelial cell dysfunction and may be an important driver of WMI in humans, there are intrinsic factors that affect endothelial cell vulnerability and drive white matter damage in the absence of hypertension [38]. In a normotensive transgenic rat model, where Atp11b was deleted, small vessels in the brain and the retina showed endothelial dysfunction and oligodendroglia differentiation block that led to myelin disruption in an uneven pattern around some brain vessels, similar to what happens in the human brain [38]. Although with limitations, this suggests that the use of specific genetic mutations in vascular function genes has the potential to be a more useful way to model WMI in mice.

The third model of WMI or VaD creates small lesions produced by vasoconstriction in the small vessels of the corpus callosum (Table 1) [39–42]. These small lesions are induced by focal microinjection of the vasoconstrictor N5-(1-iminoethyl)-L-ornithine into the subcortical white matter ventral to the mouse forelimb motor cortex in rodents [39–42]. These micro-infarcts are sub-millimeter and only visible by light microscopy and detectable behaviorally only briefly or on MRI as very small lesions. These small white matter lesions mimic the initial stages of subcortical WMI or VaD. Human white matter ischemic lesions have a characteristic pattern close to the lateral ventricle and are found in progressively decreasing frequency with increasing distance from these locations [7, 9, 16, 17, 21, 43]. Recent studies have developed animal models that produce more diffuse and larger subcortical WM lesions below the motor cortex to resemble the size and the extent of those seen in the areas most commonly affected in WMI or VaD in humans [44]. Due to the initial silent nature of WMI, the vast majority of patients live in the chronic phases of this disease. These new models mimic closely the more moderate to advanced, chronic stages in human WMI that are commonly seen in clinic [2, 7, 43, 45–48], in the overall volume of subcortical white matter that is involved and in the progressive nature of the behavioral deficits. Clinically, most white matter lesions are believed to be secondary to vascular occlusion from changes in small vessels, as well as endothelial cell dysfunction [49, 50], although the underlying mechanisms for such vasculopathies may vary. These new models show that ischemia produced by vasoconstriction in the small vessels of the white matter not only produces damage to axons, myelin, and astrocytes throughout a broad region of subcortical WM [44] but also produces progressive, long-lasting motor control and cognitive deficits that resembled the cognitive, gait, and motor deficits seen in human WMI.

Although rodent WMI models present numerous advantages, differences between humans and rodents in brain structure are vast. In the search for animal WMI models providing translational advantages for biomedical research, pigs and non-human primates have emerged due to their similarities with the human brain. Domestic pigs and non-human primates have been considered excellent for modeling neurological diseases in the past because of their relatively large gyrencephalic brain, only seven times smaller than the human brain, white–gray matter ratio (white matter content: 10% in rodents vs 45% in humans [51]) and close resemblance between immune systems [52]. To date, the vast majority of WMI pre-clinical studies using domestic pigs or non-human primates have been focused on modeling intracerebral hemorrhage [53], hypobaric hypoxia [54], or endovascular middle cerebral artery occlusion (MCAO) [55–58]. While all these animal models cause different degrees of white matter injury, they also induce extensive damage to other brain structures [53–58]. In addition to the need for highly invasive and complex surgical procedures to induce WM damage, these large mammal WMI models produce subpar behavioral readouts that are deeply affected by the multi-structural damage. Moreover, while the use of domestic pigs for pre-clinical research is wildly accepted, the use of non-human primates is highly controversial and possess severe ethical concerns.

Biology of White Matter Repair in White Matter Ischemic Lesions

An important point that has come from several studies performed in the last 10 years is that while WMI seems to be triggered by vascular dysfunction [59, 60], it secondarily affects astrocytes, oligodendrocyte-lineage cells, and axons [9, 39–41, 44] (Fig. 1). Loss of oligodendrocytes and astrocytes due to WMI occurs progressively outside of the WM lesion, and WMI progresses or expands over time in part because axons in these adjacent areas lose their glial support (Fig. 1). It has been described that WMI initiates an oligodendrocyte progenitor cell (OPC) response [39, 40]. OPCs divide and migrate to the stroke, but the OPCs then do not differentiate into mature oligodendrocytes, leaving axons adjacent to stroke infarct chronically demyelinated and susceptible to progressive degeneration [9, 39, 40] (Fig. 2). This initial area of demyelination after WMI is larger than the axonal damage, leaving a region with hypo-myelinated axons [39–41]. This peri-infarct zone of mismatch—greater loss of myelin than axons after oligodendrocyte death—leaves axons exposed and vulnerable and the lesion expands over time, as do the cognitive and motor deficits indicating a progressive white matter ischemic condition as seen in VaD (Fig. 2).

Fig. 2.

Overview of the oligodendrocyte differentiation blockage due to white matter injury. White matter injury increases oligodendrocyte progenitor cell division and migration but inhibits oligodendrocyte differentiation

On the other hand, genetic fate mapping of OPCs after WMI indicates that 4% to 13% of these cells differentiate into astrocytes at any post-WMI time point examined [39] (Fig. 2). The astrocytic fate change of OPCs after WMI may account for the very limited differentiation of OPCs into mature oligodendrocytes. Alternatively, these OPC-derived astrocytes, which are located at the stroke margin, may be performing a barrier function around the infarct and induced by local factors at the infarct border. These pre-clinical data identify three important aspects of WMI: it generates a vulnerable peri-infarct region which can progress to further damage, disrupts neuronal connections to cause substantial disability, and does not trigger myelin repair as seen inflammatory diseases of the brain, such as multiple sclerosis.

These findings establish 2 major biological targets for white matter repair after subcortical WMI or during VaD: (1) removing the differentiation block to OPCs and promoting remyelination [42, 44] and (2) inducing new connections across brain regions, establishing links between sensorimotor cortical areas [44].

Glial Cell–Based Therapies to Treat White Matter Ischemia and Vascular Dementia

Currently, there is no therapy that enhances the brain’s own ability to recover from WMI and VaD or prevent the expansion of white matter lesions over time. However, there has been substantial progress in cellular transplantation in pre-clinical and recently in clinical stroke [53–56]. Those efforts have been focused on “large artery stroke” and have utilized mesenchymal stromal cells, fetal neural progenitors, and ES or iPS-derived neural progenitors [61–64]. WMI and VaD, unlike large artery or “gray matter” stroke, damages primarily astrocytes and axons and myelin that link local (areas within the same hemisphere) and distant brain areas (areas on opposite sides of the brain) (Fig. 1). Because of the different cellular constituents of white matter, it might be anticipated that a multipotent neural precursor or a neuronal precursor cell would not be optimum, while a therapeutic intervention targeting astrocytes is ideally suited to enhance white matter repair [65].

Recent studies have demonstrated that astrocytes not only guide axons but promote formation of new connections in the developing brain [66–72]. In CNS lesions like spinal cord injury, astrocytes help to re-connect damaged tissue. This role of astrocytes in axonal connection in development and axonal sprouting in injury runs counter to traditional ideas of “astrocyte scar” but is supported by substantial experimental work [70–72]. But astrocytes do not only have an impact in axonal formation. Immature astrocytes during CNS development and cell culture promote OPC proliferation [73–75]. In in vivo CNS lesions [73], astrocytes promote OPC differentiation into myelinating oligodendrocytes. This interaction may occur through astrocyte cytokine production, or modulation of the extracellular matrix [73, 75]. These data indicate that immature or non-reactive astrocytes promote OPC differentiation and may enhance re-myelination in CNS lesions. Transplantation of exogenous glial progenitors or immature astrocytes has promoted tissue repair and re-myelination in spinal cord injury models, genetic white matter diseases, multiple sclerosis, and radiation injury models [76, 77]. This approach has utilized progenitors directed toward an OPC fate, astrocytic fate, or at a bipotential stage of possible astrocyte or oligodendrocyte differentiation [76–79]. These data indicate that the disconnection of WMI, due to loss of axons and myelin in the white matter, may be treated by the production and local action of astrocytes.

Despite the importance of glial cells, there has been only one study that successfully demonstrated meaningful neural repair in WMI or VaD [44]. This is mainly because glial differentiation of hiPS cell protocols involves long (> 3 months) and labor-intensive processes that are inefficient and not suited for the clinical-level number of cells that would need to be generated [76, 80, 81]. In recent years, in pre-clinical studies, we and collaborators have developed a unique glial cell–based therapy for neural repair in WMI or VaD with great promise [44, 82–84]. By treating human-induced pluripotent stem cell-–derived neural progenitor cells (hiPSC-NPCs) with a short exposure of an HiF activator, deferoxamine (DFX), the cells become permanently biased to differentiate into hiPSC-glial enriched progenitor cells (hiPSC-GEPs) with a pro-repair astrocyte phenotype [44, 84]. This permanent “switch” in cell fate conferred by DFX should eliminate issues with lineage restriction that have previously hampered safe use in clinical trials [85] and will confer a decrease in off-target effects and an increase in repair capacity [86, 87]. This newly developed cell differentiation protocol allows for rapid (35 days) and efficient production of a potential therapeutic candidate for WMI/Vascular dementia (hiPSC-GEPs) with strong scale-up capabilities for a clinical application [44, 82–84].

The ability to differentiate patient-derived hiPSCs into GEPs provides a strong candidate for a cell-based therapy to target WMI, VaD, and other neurodegenerative diseases. Researchers have demonstrated that cellular transplantation of hiPSC-GEPs during the subacute stage of WMI replaces cells lost in the white matter and induces surviving cells to repair the damaged axons and lost myelin, allowing brain repair to occur at local and distant sites to the injection and leads to motor and cognitive recovery (Fig. 3). This capability is distinct from the common approaches in the stem cell transplant field in stroke, which target repair of neurons [84, 88]. Therefore, a cell-based therapy that can replace lost glia and induce structural repair to lead of recovery of neurological functions in WM is of great promise.

Fig. 3.

Overview of hiPSC-GEP transplant-induced white matter repair. hiPSC-GEP transplant after white matter injury enhances oligodendrocyte differentiation and promotes remyelination that leads to motor and cognitive improvement in a model of white matter injury in mice

What might this look like in a human therapy? Studies of stem/progenitor cell transplantation in developmental white matter diseases provide some guide. White matter transplants of a human ES cell product into children with Pelizaeus-Merzbacher disease produced some evidence of clinical improvement and improvement in measures of MRI signal in white matter, with no evidence of toxicity in the circumscribed immunosuppression regimen [89]. Long-term follow-up continued to show improvement in MRI imaging and evidence for some donor immune recognition of the transplant [90]. These studies offer a guide for future human trials and a framework for the overall platform of transplantation and immunosuppression that might accompany a human stem cell trial for white matter disease.

Concluding Remarks

Recent research has revealed subcortical WMI as the major factor for vascular dementia (VaD) and a significant contributor to Alzheimer’s disease (AD). Many patients with WMI or other cerebrovascular disease have cognitive impairment, but this is often not recognized. Community-based clinicopathological studies revealed that the largest proportion of dementia cases have mixed pathology, comprising features of AD (amyloid plaques and neurofibrillary tangles) as well as ischemic lesions [91, 92]. Without current treatment, it has been estimated that 75.6 million people will have dementia worldwide by 2030 and 131.5 million by 2050. The total estimated worldwide medical cost of dementia in 2018 was a trillion dollars. Recent studies have demonstrated that treatments aimed at preserving or regenerating white matter integrity, such as specialized glial cell–based therapies are promising candidates for reducing cognitive impairment and dementia. Further research that takes into account that neurovascular dysfunction and glial damage are drivers of cerebrovascular disease, white matter lesions, and cognitive impairment is needed.

Supplementary Information

Below is the link to the electronic supplementary material.

Funding

This research was supported by Dr. Miriam and Sheldon G. Adelson Medical Research Foundation, National Institute of Health (R01 NS103788 and R37NS102185) and California Institute of Regenerative Medicine (TRAN1-12891, DISC2-09018 and DISC2-12169).