Microglial Interaction with the Neurovascular System in Physiology and Pathology

Abstract

Microglia as immune cells of the central nervous system (CNS) play significant roles not only in pathology but also in physiology, such as shaping of the CNS during development and its proper maintenance in maturity. Emerging research is showing a close association between microglia and neurovasculature that is critical for brain energy supply. In this review, we summarize the current literature on microglial interaction with the vascular system in the normal and diseased brain. First, we highlight data that indicate interesting potential involvement of microglia in developmental angiogenesis. Then we discuss the evidence for microglial participation with the vasculature in neuropathologies from brain tumors to acute injuries such as ischemic stroke to chronic neurodegenerative conditions. We conclude by suggesting future areas of research to advance the field in light of current technical progress and outstanding questions.

[1] Introduction

Microglia are innate immune cells of the brain that are fully engaged in central nervous system (CNS) functions in normal and pathological conditions (Hanisch and Kettenmann, 2007; Casano and Peri, 2015). Depending on various methods used and regions examined, microglia make up 5–15% of brain cells (Pelvig et al., 2008; Lyck et al., 2009) and many of their properties distinguish them as a unique cell type from other brain cells as confirmed by recent transcriptome studies (Hickman et al., 2013; Zhang et al., 2014; Bennett et al., 2016). Neurons and macroglia like astrocyte and oligodendrocytes are derived from the neuroectoderm, while microglia originate from the embryonic yolk sac, then migrate and colonize the neuroepithelium (Prinz and Priller, 2014; Reemst et al., 2016). In the mature brain, microglia are exquisitely tiled and share non-overlapping territories with a small central soma and multiple elaborate processes (Davalos et al., 2005; Nimmerjahn et al., 2005; Wu et al., 2007). Unlike peripheral macrophages and circulating monocytes, microglia keep a fairly low turnover rate (Lawson et al., 1992; Askew et al., 2017; Tay et al., 2017) without contributions from peripheral sources such as circulating monocytes or blood-derived macrophages in the healthy condition (Ajami et al., 2007; Ajami et al., 2011; Gu et al., 2016).

One of the most notable features of microglia is their robust morphological plasticity by which they perform CNS surveillance (Davalos et al., 2005; Nimmerjahn et al., 2005). With remarkably dynamic processes, they frequently interact with neuronal elements including somata (Li et al., 2012; Eyo et al., 2014), axons (Baalman et al., 2015), dendrites and synapses (Wake et al., 2009; Tremblay et al., 2010; Eyo et al., 2015; Eyo et al., 2017) by which they sense and monitor neural activity and thus modulate neural circuit processing (Chen et al., 2010; Paolicelli and Gross, 2011; Parkhurst et al., 2013; Schafer et al., 2013; Squarzoni et al., 2014). Microglia are also the primary phagocytes of the brain and therefore engage and engulf excess neural material during early development (Marin-Teva et al., 2004; Petersen and Dailey, 2004; Wakselman et al., 2008; Svahn et al., 2013; Eyo et al., 2016) and in select neurogenic niches in the mature brain (Sierra et al., 2010; Sierra et al., 2014; Abiega et al., 2016; Fourgeaud et al., 2016). Thus, microglial roles in shaping brain development and function by regulating neuronal activity and phagocytic clearance have been well established and continue to be explored.

In addition to functioning in the healthy brain, microglia participate in aberrant neurological conditions from acute injury to neurodegenerative diseases (Ransohoff and Perry, 2009; Nayak et al., 2014; Peng et al., 2016; Eyo et al., 2017). Here, they are hotly debated to promote either neuroprotective or neurotoxic functions. They can ameliorate pathology as is the case with acute experimental seizures where seizure behaviors are worsened with microglial elimination or microglial P2Y12 receptor depletion (Mirrione et al., 2010; Eyo et al., 2014) and ischemic stroke where they are presumed to reduce infarct size and promote functional recovery (Lalancette-Hebert et al., 2007; Narantuya et al., 2010; Faustino et al., 2011). Conversely, they can also promote disease progression as in the case of neuropathic pain (Peng et al., 2016). It is noteworthy to point out that microglial functions are not always straight-forward in diseases. For example, while studies may suggest neuroprotective microglial functions in ischemia there is also evidence for neurotoxic activity in the disease (Wu et al., 2012; Tian et al., 2016). Interestingly, microglia have also been implicated in the pathogenesis of psychiatric disorders such as pathological grooming in mice which is suggested to mimic obsessive compulsive disorder (OCD) in humans (Chen et al., 2010). Together, these studies and many others indicate that microglia are significant cellular components in the brain orchestrating critical events during normal development, the maintenance of the mature CNS and during pathology.

It has been long recognized that microglia physically interact with the neurovasculature and interest in the physiological significance of these interactions are increasing (Arnold and Betsholtz, 2013; Dudvarski Stankovic et al., 2016). The brain vasculature, which includes a complex arrangement of arteries, veins and capillaries, distributes essential substances like oxygen and glucose to the brain while eliminating waste products like CO₂ (Daneman and Prat, 2015). The brain, though constituting only about 2% of the body’s weight, demands about 20% of the body’s energy (Magistretti and Allaman, 2015). This fact implies a need for efficient and controlled delivery of blood to the brain, which is especially critical for optimal synaptic and thus cognitive function. While astrocytes are now established as critical regulators of the neurovasculature (MacVicar and Newman, 2015), microglia have increasingly been implicated in neurovascular development and complexity and thus indirectly regulate brain function. This review will summarize the current literature on the role for microglia in neurovascular function. Our assessment will begin with the relevant literature on the current conclusions regarding microglial roles in developmental angiogenesis and conclude with the insights provided by studies investigating microglial function in pathological conditions including cerebral tumors and neurodegenerative diseases.

While the focus of this review is not on microglial-astrocyte or microglial-pericyte interactions, it should be noted that there is growing evidence for interactions between microglia and these cells as components of the neurovascular unit. For example, microglia promote astrocyte differentiation (Nakanishi et al., 2007). However, astrocytes also regulate microglial phenotypes by paracrine signaling of released molecules (Rezaie et al., 2002; Bohlen et al., 2017) indicating that bidirectional regulation between microglia and astrocytes could have consequences on vascular development and integrity. Pericytes also serve as cytokine responsive cells and amplify microglia activation (Matsumoto et al., 2014) and are suggested to function as immune suppressors (Hurtado-Alvarado et al., 2014). This might imply a regulation by pericytes of the microglial phenotype on the uninjured brain. They have been recently reported to be a source of microglia especially following ischemia (Ozen et al., 2014; Sakuma et al., 2016). Moreover, activated microglia promote pericytes cell death via ROS production (Ding et al., 2017) and may thus compromise blood vessel integrity since pericytes make up the basement membranes of the vasculature. Therefore, although we don’t highlight the various interactions between these cellular components of the neurovascular unit and microglia in the following pages, these interactions cannot be ignored.

[2] Microglia and Blood Vessels in Normal Physiology

[2.1] Microglial Colonization of the CNS Precedes the Development of, but then Closely Associates with Blood Vessels

Several studies have documented that microglial emergence precedes the carefully orchestrated formation of the cerebral blood vessel network in the developing brain. This is the case in the mouse retina (Rymo et al., 2011) as well as hindbrain (Fantin et al., 2010). Consistent with these results from mice, avian data using chick-quail chimeras also revealed that yolk sac derived macrophages that give rise to microglia invade the CNS prior to and independent of neurovascular development (Cuadros et al., 1993; Kurz and Christ, 1998).

Although microglia and the neurovasculature do not appear in the developing CNS at the same time, when they do appear, they often display physical interactions. Early studies documented the localization of microglial cells with the vasculature in the developing rodent brain (Ashwell et al., 1989) and were sometimes referred to as “pericytic macrophages” (Thomas, 1999). This has been repeatedly confirmed in subsequent studies in the brain proper (Arnoux et al., 2013) and suggests relevant (perhaps bi-directional) communication between microglia and the vasculature. Moreover, these observations provided support for an early hypothesis that microglia migrate into the brain via blood vessels (Perry et al., 1985). Subsequent results, however, suggest that circulating cells do not infiltrate the brain parenchyma in the healthy CNS (Ajami et al., 2007) dampening enthusiasm for this early hypothesis. The observation of microglial proximity to blood vessels is especially noticeable during development. For example, in both the developing human and mouse retina, microglia are closely opposed to blood vessels (Checchin et al., 2006; Rymo et al., 2011). Furthermore, “juxtavascular” microglia were observed by real time confocal imaging to migrate along the walls of cerebral blood vessels in brain slices (Grossmann et al., 2002). These microglia exhibited a greater likelihood of migration than non-juxtavascular microglia. Although such observations were determined in an excised tissue context, whether the observations are an artifact of tissue excision or representative of the native developmental brain environment will have to be elucidated using more robust in vivo imaging approaches. Moreover, the precise function of this migration along blood vessels is not clear.

Less work has been done to provide details on microglial-vascular interactions in the mature brain. Nevertheless, interest in functional interactions between microglia and the neurovasculature continues to mount. Our initial studies using in vivo two photon imaging in adult mice has revealed that microglia maintain robust physical contact with elements of the neurovasculature (Fig. 1). Work is underway to adequately characterize these interesting microglial interactions with the neurovascular system and determine their functional significance in physiology and pathology.

Figure 1. Microglia and blood vessels exist in close proximity in the adult brain.

Z-stack images from the somatosensory cortex of a transgenic CX3CR1-GFP^+/− mouse injected with SR101 (red) to label the neurovasculature. a–b, Low magnification images of ramified microglia and blood vessels in 50 μm z-stack images. Scale bar: 50 μm. c–d, High magnification images of ramified microglia and blood vessels in 10 μm z-stack images of the corresponding boxed regions in (a) and (b). Physical contact between microglia somata and capillaries are indicated with arrows. Scale bar: 20 μm.

[2.2] Microglia Promote Angiogenesis

One of the increasingly recognized functions of microglia is that they participate in the formation of new blood vessels or angiogenesis. This has especially been documented in pathological contexts (see next section) and presumes prior mechanistic similarities in earlier development. We will now consider some of the evidence that have been provided for this from both genetic and pharmacological approaches.

Support for the involvement of microglia in developmental angiogenesis from genetic approaches was first provided by the genetic depletion of macrophage colony stimulating factor, otherwise known as mCSF. This gene controls the development and survival of cells of the monocytic lineage including brain microglia (Cecchini et al., 1994). In mice genetically deficient of mCSF, the development of microglia and other monocyte-derived cells are lacking. Interestingly, the complexity of the developing retinal vasculature in mCSF knockout mice was also reduced suggesting roles for microglial mCSF signaling in angiogenesis in the retina. The deficiency in vasculature complexity was transient as differences were only observed in development but not in adulthood. Therefore, there might be complementary non-mCSF or microglia-independent mechanisms for retinal angiogenesis (Kubota et al., 2009).

Similar to mCSF, the PU.1 gene controls hematopoietic cell differentiation and genetic interruption of its function prevents microglial development (Scott et al., 1994; McKercher et al., 1996). Interestingly, PU.1 mutant mice also exhibit less elaborate blood vessel complexity (Fantin et al., 2010). Of course, given the limitation that these genetic approaches target non-microglial cells, the deficiency in vascular complexity could result from contributions from non-microglial population. Therefore, more selective microglial elimination techniques will have to be employed in future studies to adequately determine microglia-specific roles in developmental angiogenesis. However, it is worth noting that the challenge of ablating microglia in early development is not trivial and has yet to be accomplished at the time of this writing.

Pharmacological evidence suggesting microglial contributions to developmental angiogenesis has been multiply documented. First, consistent with a role for mCSF, its pharmacological inhibition using neutralizing antibodies to mCSF also reduced blood vessel complexity (Kubota et al., 2009). This pharmacological approach also suffers from the limitation that mCSF could target non-microglial cells that express mCSF receptors. Second, selective depletion of microglia using clodronate liposomes by which microglia die upon their uptake liposomes, resulted in reduced vascularization in the developing mouse retina (Checchin et al., 2006). Yet, as with the genetic approach, future studies will have to specifically target microglia selectively and do so especially in an in vivo context. Together, these results consistently suggest that microglia promote developmental angiogenesis.

[3] Microglia and Blood Vessels in Pathology

Although microglial roles in blood vessel development and maintenance have not received intense research attention as discussed above, more research has been directed to its function in various pathologies from brain tumors to acute brain injuries like ischemia to chronic neurodegenerative diseases. Current data on these findings will be discussed in this section. The relevance of findings from such studies is that understanding microglial roles in these pathological contexts could (i) inform future therapies that can be developed in the treatment directed towards ameliorating the progression of these pathologies and (ii) inform researchers on candidate mechanisms to identify factors by which microglia may regulate developmental angiogenesis and neurovascular physiology.

[3.1] Microglia and Blood Vessels in Brain Tumors

In the last two decades of angiogenesis research, it has become clear that vascular endothelial growth factor (VEGF) is a predominant regulator in the development and patterning of blood vessels (Thomas, 1996). This has been confirmed in multiple systems including the mouse brain (Ruhrberg et al., 2002; Haigh et al., 2003; Raab et al., 2004), the mouse retina (Stalmans et al., 2002; Haigh et al., 2003) and the quail neural tube (James et al., 2009). As with developmental angiogenesis, the tumor environment is pro-angiogenic and thus seems to promote the development of new blood vessels (otherwise termed “neovascularization”), which is important for (i) the delivery of blood and its accompanying nutrients to the tumor and (ii) the metastasis of tumor cells from the original tumor site to novel sites (Lorger, 2012). On this account, brain tumors possess a characteristically increased density of blood vessels (Lopes, 2003) some of which could be malformed and potentially leaky. Factors that promote the neovascularization of tumors are therefore generally considered detrimental to the outcome for the patient while those that conversely limit tumorigenic neovascularization are considered beneficial.

While astrocytes are considered to be a predominant cell type that secrete VEGF for angiogenesis (Pierce et al., 1995), selectively abrogating astrocyte-derived VEGF did not significantly alter the developmental angiogenesis (Scott et al., 2010; Weidemann et al., 2010) suggesting that either compensatory mechanisms are in play or astrocyte-derived VEGF is not critical for blood vessel development. Although microglial VEGF roles in developmental angiogenesis are not clear, microglia are now known to express some VEGF isoforms (Zhang et al., 2014) suggesting the capacity for regulating vascularization.

In human tumors, microglial/macrophage density was increased with a corresponding increase in tumorigenic neovascularization (Nishie et al., 1999; Brandenburg et al., 2016) and macrophages were increasingly associated with blood vessels (Leek et al., 1996; Brandenburg et al., 2016). Since microglia are the macrophages of the CNS, the implication that macrophages induce angiogenesis in tumor environments (Kobayashi et al., 1994; Sunderkotter et al., 1994; Polverini, 1997; Wang et al., 2013; Qin et al., 2015) could also apply to microglia (Wyckoff et al., 2004). Interestingly, despite the fact that it is difficult to distinguish brain resident microglia from infiltrated macrophages molecularly, recent evidence suggest that the pro-angiogenic function is predominantly carried out by microglia rather than macrophages in brain tumors (Brandenburg et al., 2016). Furthermore, factors derived from microglia are known to facilitate tumor progression (Sliwa et al., 2007). Specifically, microglia release tumor necrosis factor-alpha (TNFα) in the tumorigenic environment (Hattermann et al., 2014; Hwang et al., 2016), which in turn regulates the release of VEGF from glioma cell lines (Ryuto et al., 1996) that is critical for neovascularization. Thus either by the direct release of VEGF or the indirect release of other factors that increase its expression, microglia participate in the promotion of tumor angiogenesis. Finally, since VEGF serves as a chemoattractant to microglia (Forstreuter et al., 2002), its release in tumors could serve as an attractive signal to microglia towards blood vessels in an autocrine (if microglia are the source) or paracrine (if other cells are the source) manner.

In addition to microglial VEGF signaling, microglia release matrix metalloproteinases (MMPs) in vascularizing the tumor environment. Since MMPs function to degrade and remodel the extracellular matrix, their function is pro-angiogenic and thought to be recruited by the tumor for the promotion of tumor expansion (Egeblad and Werb, 2002; Rao, 2003). Microglia express several MMPs along with other cells of the brain (Hickman et al., 2013; Zhang et al., 2014; Holtman et al., 2015). MMP expression such as MT1-MMP (or MMP14), has been detected on human and murine microglia while tumor cells fail to express the protein (Markovic et al., 2009). Moreover, microglial depletion using either clodronate liposomes which eliminated MMP expression or genetic ablation both resulted in a reduction in tumor invasion (Markovic et al., 2009) confirming a pro-tumorigenic role for the microglial-derived MMP. Consistent with a role for microglial release of MT1-MMP in promoting tumor-induced angiogenesis, upregulation of the gene during tumor progression was correlated with increased neovascularization (Gabrusiewicz et al., 2011). These results indicate that microglial MMP activity promotes tumor progression by facilitating vascularization in the tumor environment.

Other microglial factors have been identified that promote tumor growth. For example, the microglial Na(+)/H(+) exchanger isoform 1 (NHE1) was recently identified as a respectable target to ameliorate tumor progression (Zhu et al., 2016). The mechanisms investigated suggested a regulation of tumor cell migration and proliferation. However, it would be interesting to determine whether the exchanger promotes tumor invasiveness by angiogenic mechanisms as well. Recently, CXCL2, a cytokine which is predominantly expressed by microglia in the brain parenchyma (Hickman et al., 2013; Zhang et al., 2014; Holtman et al., 2015), was shown to promote tumor-induced angiogenesis (Brandenburg et al., 2016). These results indicate that microglia facilitate blood vessel formation in a tumorigenic environment and some of the regulatory factors include growth factors, proteases, transporters and cytokines.

[3.2] Microglia and Blood Vessels in Ischemic Stroke

Stroke is the fifth leading cause of mortality in United States and a leading cause of disability (Talwalkar and Uddin, 2015). Many aspects of the role of microglia in ischemic contexts have been extensively reviewed (Ma et al., 2016). Microglia respond earliest following an ischemic insult and serve as the first line of defense to the injury (Morioka et al., 1991; Weinstein et al., 2010). Microglial accumulation is also one of the earliest cellular signatures in cerebral ischemia (Gelderblom et al., 2009). Ischemia induced by photothrombosis revealed that microglial dynamic activity is closely associated with capillary blood flow around its cell body (Masuda et al., 2011). The dynamics of microglial processes is suppressed around the capillary with decreased blood flow, suggesting microglial surveillance is inhibited during ischemia, which is consistent with evidence from the developing brain (Eyo and Dailey, 2012).

Microglia become closely associated with blood vessel after ischemia by forming perivascular clusters and phagocytic structures (Jolivel et al., 2015). The accumulation of microglia around the vasculature subsequently led to the disintegration of the vessels which included their upregulation of phagocytic CD68 expression in the penumbra area. Accumulation of microglia with blood vessels also correlated with the invasion of blood-borne molecules during reperfusion (Jolivel et al., 2015). Interestingly, a selective inactivation of microglial CX3CR1 that has been reported to regulate microglial migration (Cardona et al., 2006; Liang et al., 2009) and significantly reduce blood extravasation (Jolivel et al., 2015). This may be a mechanism for neuroprotection in stroke since several studies have indicated reduced stroke pathology in CX3CR1-deficient mice (Denes et al., 2008; Jolivel et al., 2015).

Furthermore, perivascular microglia might also contribute to cerebral ischemia by releasing microglia-specific cytokines that are known to compromise vascular integrity in ischemia (Sprague and Khalil, 2009). For example, IL-1β and TNFα, both known to increase the permeability of the blood brain barrier (BBB) (Tsao et al., 2001; Mayhan, 2002; Sibson et al., 2002; Wang et al., 2014; Richter et al., 2017) are released by microglia early during ischemia (Lambertsen et al., 2012) to promote the compromise of the BBB. Later in the progression of ischemia, microglia also release VEGF (Xie et al., 2013) known to promote angiogenesis in stroke (Zhang et al., 2000) suggesting that they could play some reparative functions by inducing neovascularization in later stages of stroke that contrasts with their earlier function. Of note, even IL-1β has been reported to have pro-angiogenic functions as well (Giulian et al., 1988) which may also be recruited in the latter repair following ischemic injury. Together, these results suggest that in the stroke context, microglia breakdown extant blood vessels, partially through cytokine insults and partially through phagocytic engulfment early on following the insult, and then contribute to building new vasculature later on. Future work will have to more precisely and adequately test this hypothesis.

[3.3] Microglia and Blood Vessels in Neurodegeneration

[3.3.1] Alzheimer’s disease

Alzheimer’s disease (AD) represents the most common neurodegenerative disease and is especially fatal in the aging population (Ballard et al., 2011). Characterized by amyloid beta (Aβ) deposits in the brain as a histopathological hallmark, microglial reactivity in the AD brain is also well known (Heneka et al., 2015; Yeh et al., 2017) but the specific contribution of microglia remain hotly debated (Gold and El Khoury, 2015; Malm et al., 2015). Here, we focus on some of the evidence that among other things, microglia in the AD context participates in vascular abnormalities that occur in the disease.

As with previous discussions, VEGF expression is increased in AD in response to Aβ deposition (Kalaria et al., 1998; Tarkowski et al., 2002). Since Aβ is deposited in and/or around blood vessels in addition to the parenchyma (Okamoto et al., 2009; Hickman and El Khoury, 2010), VEGF, among other chemoattractant, may mobilize microglia to surround blood vessels. This is consistent with a robust perivascular accumulation of microglia in AD (Ryu and McLarnon, 2009; Giannoni et al., 2016). In the 5xFAD model of AD, a longitudinal assessment of AD pathology using intravital two-photon microscopy revealed overlapping regions of neurovascular defects and Aβ plaques, which correlated with increased microglial activation (Giannoni et al., 2016). Whether the microglial reactivity in perivascular regions was a cause, consequence or an independent correlating factor with the histopathological vascular defects was not determined.

However, in an experimental model of AD where Aβ is injected into the hippocampus, there is a corresponding increase in neovascularization and microglial activation (Zand et al., 2005). Furthermore, in this condition, the BBB becomes leaky and this has been correlated with BBB-associated astrocytes and microglia (Ryu and McLarnon, 2006). Indeed, application of minocycline to inhibit microglial activation in this context significantly reduced the Aβ-induced defect in BBB integrity (Ryu and McLarnon, 2006). This result was further supported by an alternative approach using CD11b antibodies to block microglial function, which resulted in reduced neurovascular deficits from Aβ injection (Ryu and McLarnon, 2009). Although the precise mechanisms for microglial action remain to be elucidated, one of the mechanisms that may be proposed to be employed by microglia in promoting vascular pathology in Aβ pathology is through the purinergic P2X7 receptor (P2X7R) signaling (Ryu and McLarnon, 2008). Microglia are recognized as a predominant cell that expresses the P2X7R in the brain (Hickman et al., 2013; Zhang et al., 2014; Holtman et al., 2015) despite the lingering controversy of neuronal P2X7R expression (Illes et al., 2017; Miras-Portugal et al., 2017). In this light, since pharmacological inhibition of P2X7Rs improved defects induced by Aβ treatment such as aberrant vascular function (Ryu and McLarnon, 2008), it is tempting to speculate that a primary cellular component of action occurred through microglia.

MMPs expressed by microglia and released in AD pathology have also been documented to contribute to the progression of AD (Kim and Joh, 2012). However, whether MMP action is specifically through microglia or has any specific effect on AD-induced vascular aberrations remains to be determined. Finally, microglia also release several cytokines including TNFα and IL-1β during AD pathology (Cameron and Landreth, 2010; Mandrekar-Colucci and Landreth, 2010; Wang et al., 2015) that can also compromise the vasculature as discussed above (see section 3.2 above). On the basis of the above, microglia are thus thought to promote vascular pathology including a breakdown of the BBB in Aβ pathology like AD.

[3.3.2] Multiple Sclerosis and Parkinson’s disease

Multiple sclerosis (MS) is a progressive neurodegenerative autoimmune disease characterized by demyelination, brain atrophy and chronic inflammation (Lassmann et al., 2001; Vos et al., 2005). As with other conditions discussed above, angiogenesis has been reported in animal models of MS (Seabrook et al., 2010; Girolamo et al., 2014). In addition, BBB integrity is compromised in early stages of the disease leading to increased leukocyte infiltration and accumulation of blood products such as fibrinogen into the brain (Vos et al., 2005; van Horssen et al., 2007). Real time in vivo imaging revealed that leaked fibrinogen serves to attract microglia towards blood vessels in the early stages of animal models of MS. This attraction continues into more severe stages of the disease (Davalos et al., 2012). However, although microglia in general are thought to promote MS pathology (Heppner et al., 2005), it’s precise role in vascular abnormalities of the disease such as BBB compromise and angiogenesis have not been clarified and should be a focus of future studies.

Parkinson’s disease (PD) is a progressive neurodegenerative motor disease and is characterized by the loss of dopaminergic neurons in the substantia nigra (Kalia and Lang, 2015). Activation of the immune system and especially microglia is well known for disease progression (Whitton, 2007; Luo et al., 2017). As with the AD and MS, neovascularization (Barcia et al., 2005; Desai Bradaric et al., 2012) and a compromised BBB (Brochard et al., 2009; Gray and Woulfe, 2015) have been reported in PD/animal models of PD. However, whether and how microglia may be involved in regulating neurovascular changes that occur during PD has not been adequately explored.

[4] Outstanding Questions and Future Direction

The previous sections have documented the progress in understanding microglial engagement with the neurovascular system in health and disease (Fig. 2). However, at least three avenues of research should be employed to address outstanding questions. First is the question of molecular and mechanistic details of microglial regulation of vascular function in the developing and mature brain. While compelling evidence support an important role for microglia in maintaining vasculature function in the brain, the basis of communication between microglia and the neurovasculature is still elusive at this moment. More basic research would lead to a better understanding of microglia-vascular interaction that can be harnessed to treat vascular pathologies of the brain. Precise roles of cytokines, free radicals, purines and proteases should be investigated in these various contexts. Relevant intracellular signaling pathways by which microglia receive signals from and send signals to the vasculature remain to be identified. Furthermore, future research should be directed to reveal the extent and details of microglial physical and dynamic interactions with elements of the neurovasculature and how they may differ between for example, capillaries, arteries and/or veins. The advent of real time two photon imaging can be used to accomplish this.

Figure 2. Schematic diagram of microglial interaction with the neurovascular system in physiology and pathology.

Resting microglia, as depicted in normal condition, are characterized with ramified cellular processes and close proximity of blood vessel. Activated microglia under diseased conditions are featured with larger somata and enriched in the dysfunctional core of associated diseases. Molecules, like chemokines, cytokines & growth factors, included in interaction of microglia and neurovascular system are indicated in each panel as normal condition, brain tumor, ischemic stroke, Alzheimer’s disease and multiple sclerosis.

A second aspect that should guide future research is the question of specific microglial populations that may orchestrate neurovascular development and maintenance (or disintegration) in health and disease. RNA sequencing and transcriptional profiling data suggests a rich heterogeneity of microglia with distinctive molecular profiles (Zhang et al., 2014; Grabert et al., 2016; Keren-Shaul et al., 2017). In light of these results, it would be interesting to determine whether there is a microglial subtype specifically responsible for the maintenance/modulation of the neurovascular system in various brain regions. Morphological and molecular characterizations of such region specific microglia-vascular interaction would have to be determined. In addition, perivascular microglia have been identified in the literature but whether they are molecularly distinct from other resident microglia will need to be clarified to better understand their function.

Finally, while evidence is mounting for promising roles of microglia in angiogenesis and the maintenance of the BBB integrity, a lot of work remains to be done to determine precise microglial involvement in these processes especially in pathology. Prior work did not adequately ascertain microglial involvement because of the lack of specificity in the approaches. Approaches that potentially targeted microglia and peripheral monocytes/macrophages indiscriminately may have led to faulty conclusions in microglia-neurovascular interactions. Minocycline was mostly used to inhibit microglial activation in studying microglia-vascular interaction, however, it also has other non-microglial and perhaps direct neuromodulatory effects (Huang et al., 2010). Moreover, future studies are needed to differentiate the direct microglial interaction with neurovascular system from the indirect effect of microglia-neuron or microglia-astrocyte communication. Continual advancements in microglial-specific genetic and pharmacological tools in general cell ablation and specific protein deletion has now set the stage for better studies to adequately address these concerns.