Neurovascular unit (NVU) consisted of cerebral microvessel endothelial cells, pericytes, astrocytes, neurons, and microglia is a structural and functional basis of the blood-brain barrier (BBB). NVU/BBB is actively involved in the regulation of crucial physiological processes within the brain, but is dramatically compromised in almost all the types of brain pathology (neurodevelopmental disorders, neurodegeneration, trauma, ischemia, neuroinflammation, neuroinfection etc.).

Brain microvascular endothelial cells (BMEC) represent one of the most interesting subpopulations of endothelial cells (EC) due to their specific properties required for the appropriate functioning of these cells: (i) regulation of cerebral blood circulation, (ii) selective and controlled permeability of the blood-brain barrier (BBB). In contrast to endothelial cells of non-cerebral localization, BMEC are characterized by high expression of tight junctions, high electrical resistance, low fenestration, small perivascular space, high prevalence of insulin and transferrin receptors, relatively big number of mitochondria (Stamatovic et al., 2008; Salmina et al., 2014). As it was demonstrated (De Bock et al., 2013), metabolic plasticity of endothelial cells allows rapid switching to active growth upon stimulation, i.e., under the conditions supporting establishment of tip or stalk cells phenotypes with the corresponding proliferative and migratory activity. Thus, metabolism of endothelial cells is considered as a major driver of angiogenesis and might serve as a marker of endothelial dysfunction seen in cardiovascular and cerebrovascular diseases (Eelen et al., 2015). However, another question arises on the metabolic activity of endothelial progenitor cells (EPC) that are responsible for vessel development and (re)endothelialization.

Currently, there is considerable interest to the assessment of BBB-related angiogenesis in developing and adult brain. Development of NVU critically depends on the availability of EPC that originate from embryonic hemangioblasts and hematopoietic stem cells and form the primary vascular plexus (Rae et al., 2011a). Embryonic vasculogenesis and establishment of BBB is driven by newly formed EPC migrating from the sites of hematopoietic stem cells (HSC) development. Bone marrow starts to function as a source of HSC just before birth whereas in embryogenesis, multi-lineage hematopoietic progenitors exist in the extraembryonic yolk sac at E8.25, in the placenta and embryonic aorta-gonad-mesonephros region at E10, and in the fetal liver at E11.0 in mice (Coşkun et al., 2014). Later, in the postnatal brain, EPC may also come from the bone-marrow hematopoietic niches, but their role in the endothelialization, maturation and maintenance of the structural and functional integrity of BBB is not clear yet.

The term “endothelial progenitor cells” initially covered various subsets of circulating progenitors derived from bone-marrow pluripotent stem cells and hemangioblasts. EPC are involved in vessel development and regeneration, being recruited from the bone-marrow to the peripheral blood, displaying immunopositivity for CD34, CD45, CD133, VEGFR2, c-kit, and, presumably existing as hematopoietic cells with pro-angiogenic activity in vitro and in vivo (Richardson and Yoder, 2011; Yoder, 2012). In adults, there are two origins of EPC: (i) hematopoietic origin (EPC derived from bone marrow multipotent hemangioblasts [VEGFR2(+)VE-cadherin(+)CD45(−) and mesenchymal stem cells (MSC) (CD73(+)CD90(+)CD105(+)CD34(−)CD45(−)]; (ii) and non-hematopoietic origin (EPC found at the sites of extensive angiogenesis but demonstrating no signs of hematopoietic origin, being, probably, derived from tissue multipotent cells) (Chao and Hirschi, 2010; Leeper et al., 2010; Boxall and Jones, 2012).

Bone-marrow-derived MSC possess the ability to migrate though the BBB in vivo and in vitro models without evident alterations in the barrier's integrity (Matsushita et al., 2011). Interestingly, there are the reciprocal effects of BMEC and MSC in hypoxic conditions: BMEC stimulate differentiation of MSC into EC, whereas MSC stimulate proliferation and migration of BMEC, thereby contributing to local angiogenesis associated with high BBB permeability (Liu et al., 2008).

Brain tissue multipotent cells express markers for mesenchymal stem cells (i.e., CD13, CD105) and pericytes (i.e., PDGFR-β/CD140b, RGS5, Kir 6.1, NG2) and demonstrate strong multilineage potential (Paul et al., 2012). Relative similarity of MSC and pericytes is very well-known. In the adult brain, pericytes originate from the pre-existing pool or from some bone-marrow progenitors, may express wide spectrum of MSC markers in culture (CD44, CD73, CD90, CD105), contribute to the maintenance of BBB integrity being in the tight contact with EC (Pombero et al., 2016). Pericytes and endothelial cells are under the control of perivascular astrocytes that induce their differentiation needed for effective angiogenesis, thereby astroglial dysfunction may affect angiogenesis via dysregulation of EPC/EC/pericytes interactions in cerebral microvessels. In several tissues, aberrant angiogenesis may be caused by the loss of pericytes number and inadequate proliferative response of EC (Ergul et al., 2015).

Thus, vasculogenesis (establishment of new vessels) is provided by EPC differentiated from embryonic hemangioblasts, or adult EPC, multipotent stem cells, and vessel wall-associated mesenchymal-like cells (mesoangioblasts) (Schmidt et al., 2007). Also, adult hemangioblasts have been detected in the CD133+-population of peripheral blood (Loges et al., 2004). Sprouting or splitting angiogenesis (adult vascular growth) is provided by EC pre-existing in the vessel wall, but acquiring new phenotype (tip and stalk cells) and acting with the support of EPC coming from bone marrow or non-hematopoietic sources (Rae et al., 2011b), and pericytes.

In pathological conditions, angiogenesis and vascular remodeling are usually considered as significant components of brain tissue repair program after injury (hypoxic, ischemic, traumatic, inflammatory, toxic etc.), thus, EPC-mediated mechanisms should be of great importance. As an example, in stroke models, mobilization of EPC from bone marrow correlates to the severity of cerebral alterations (Arai et al., 2009). In Alzheimer's type of neurodegeneration, accumulation of amyloid results in excessive angiogenesis and leaky BBB whereas the levels of circulating EPC is dramatically reduced (Lee et al., 2009; Biron et al., 2011). In autism, persistent remodeling of brain microvessels with the characteristics of splitting angiogenesis due to predominant proliferation of pericytes but not EC may affect neuronal connectivity (Azmitia et al., 2016). In diabetes mellitus, cerebral angiogenesis is exclusively enhanced and is associated with the appearance of non-functioning microvessels and decreased ratio of pericytes to EC (Prakash et al., 2012), but hypoxic injury of diabetic brain is characterized by delayed angiogenesis impeding brain tissue repair (Poittevin et al., 2015). In depression, insufficient angiogenesis is caused by the decreased number of EPC in the peripheral blood, low VEGF effects, and elevated levels of anti-angiogenic factors, whereas stimulation of cerebral angiogenesis is a marker of functional recovery (Boldrini et al., 2012; Yamada, 2016). In sum, it is clear that in almost all the types of brain pathology, deficits of circulating EPC attribute to the aberrant angiogenesis, thus suggesting impaired mobilization, migration of these cells to the brain.

Chemokines-, SDF1-, MMP9-, VEGF-, NO-dependent mechanisms are responsible for the mobilization of EPC from the bone marrow, whereas homing of the recruited cells is provided by molecules with high pro-angiogenic potential (i.e., VEGF, IGF, angiopoietins, cytokines, integrins) at the sites of developmental or pathological angiogenesis (Tilling et al., 2009; Caiado et al., 2011). Also, in addition to bone marrow-derived EPC, non-bone marrow-derived cells (tissue resident) could transform to EC and take part in the re-endothelizlization and angiogenesis (Balaji et al., 2013).

According to the current view, circulating EPC may differentiate into EC to restore the endothelial layer, may directly incorporate into injured endothelium, or may secrete pro-angiogenic factors (VEGF, SDF-1, PDGF etc.) or release microparticles to stimulate tip and stalk cells (Li et al., 2015b). Arrival of EPC to the sites of angiogenesis results in the establishment of functional connections between EPC and EC through the coordinated expression of adhesion proteins in a TNFα-dependent manner (Prisco et al., 2015). EPC provide paracrine signaling to facilitate angiogenesis and phenotype changes in the pre-existing resident endothelial (or endothelial progenitor) cells in the tissue-specific manner at the sites of endothelial injury or high metabolic demand (Zhang et al., 2014). Either in embryonic and adult EPC, secretion of pro-angiogenic chemokines is up-regulated by hypoxia simulating in greater expression of CXCR4, CXCR2, and release of CXCL1, CXCL12, macrophage migration inhibitory factor MIF, VEGF. As a result, adhesive capacity of EPC and local tube formation are greatly improved (Kanzler et al., 2013). Thus, EPC whose recruitment from the bone marrow is stimulated by hypoxia or cytokines, serve as cellular carriers of angiogenic regulatory factors instructed to promote angiogenesis or re-endothelialization. Such carrier function of EPC is compromised in aging as it was demonstrated on the reduced secretion of VEGF, IL-8, IL-17, and granulocyte-colony stimulating factor (G-CSF) from elderly human EPC (Kushner et al., 2008).

Recent data suggest that this mechanism seems to be supplemented with the complementary ones. Firstly, there is a transfer of organelles (i.e., mitochondria, lysosomes) through active tunneling nanotubes from circulating EPC [that should be very prone to initiate nanotubes activity similarly to other stem cells (Yang et al., 2016)] to the resident EC (de Cavanagh et al., 2014). Such organelle donation may rescue damaged endothelial cells from apoptosis or, conversely, facilitate active cell death and establishment of a platform for further endothelial replacement. As an example, in the senescent stressed endothelium, lysosomal transfer from EPC improves EC viability, normalizes endothelial-regulated vasorelaxation, and reduces programmed cells death (Yasuda et al., 2011). Secondly, EPC-released membrane-derived microvesicles are responsible for mRNA transfer to EC. To do that, the microvesicles incorporate in EC due to intermolecular interaction of EC proteins with α4 and β1 integrins (Very Late Antigen-4, VLA-4) expressed in microvesicles, thereby promoting EC proliferation, tube formation, and reducing EC apoptosis (Deregibus et al., 2007). It should be mentioned that α4β1 integrins are the receptors for fibronectin and VCAM-1, and the latter is important for BBB functioning in neuroinflammation (O'Carroll et al., 2015).

General population of EPC is very heterogenous and dynamic. Bone marrow hemangioblasts-derived early and late EPC might be determined in vitro being different in their ability to form new vessels and to incorporate into growing vascular networks: Late CD31(+)VE-cadherin(+)CD34(+)CD14(−)CD45(−) EPC with high expression of eNOS and VEGFR2 but not early CD31(+)VE-cadherin(−)CD34(−)CD14(+)CD45(+) EPC with low expression of eNOS and VEGFR2 (Fadini et al., 2012; Cheng et al., 2013; Minami et al., 2015). Expression of eNOS is critical for EPC proliferation and migration (Lu et al., 2015) and may contribute to controlling viability of EPC at the sites of induced angiogenesis (Dong et al., 2016). Transcriptomic data confirmed that early EPC are hematopoietic cells with monocytic phenotype and low angiogenic potential, whereas late EPC have greater proliferative potential and high expression of VEGFR2 (Medina et al., 2010) that is ultimately involved in various steps of angiogenesis and is up-regulated in EPC expansion (Smadja et al., 2007). Interestingly, differential expression of VEGFR2 and another universal EPC marker—Tie2—correlates to the ability of EPC to promote angiogenesis: high Tie2 expression is required for re-endothelialization, whereas greater number of low Tie2/high VEGFR2 cells better incorporate into CD31(+) capillaries (Adamcic et al., 2013).

Cell adhesion glycoprotein CD146 is a pan-endothelial marker found in EPC, circulating EC, and in murine brain blood vessels (Schrage et al., 2008; Flores-Nascimento et al., 2015). CD146(+) EPC belong to the group of late EPC with high pro-angiogenic potential. These EPC could be easily distinguished from CD146(+) circulating mature EC: CD146+ CD34+ CD45+ CD133+ or CD117+, and CD146+ CD34+, CD45– CD133– or CD117–, respectively (Delorme et al., 2005). Moreover, proteolytically generated soluble form of this molecule (sCD146) possesses pro-angiogenic activity (Stalin et al., 2013). The exact role of CD146 expressed in cerebral endothelium remains to be evaluated, however, it takes part in the mechanism of lymphocyte extravasation through the BBB in neuroinflammatory conditions (Duan et al., 2013).

Some bone marrow-derived EPC express stem cells marker CD117 (c-kit) (Beaudry et al., 2007) which is involved in MMP9-mediated mobilization of EPC from the bone marrow in a response to high concentrations of VEGF (Heissig et al., 2003), and CXCR4 which is a receptor for CXCL12 chemokine (stromal cell-derived factor-1, SDF-1) involved in the process of EPC mobilization to the peripheral blood. Bone marrow-derived and umbilical cord blood-derived EPC differ in CXCR4 expression even they demonstrate compatible angiogenic properties (Finney et al., 2006).

Increased expression of CXCR4 on EPC is required for adenosine-induced mobilization of these cells from the bone marrow (Rolland-Turner et al., 2013). The remarkable fact is that mature BMEC are also regulated by adenosine which is known as a mediator of BBB structural and functional integrity: acting at A2A adenosine receptors it increases the barrier permeability for some drugs and immune cells for the meaningful time window (Carman et al., 2011; Kim and Bynoe, 2015), therefore being suggested as a potent agent for improving drug delivery to the CNS (Gao et al., 2014). Adenosine in high concentrations is produced in various tissues from ATP by CD39/CD73 ectonucleotidases or, alternatively, via CD38 (or CD157)/CD203a/CD73 pathway sensitive to the local concentrations of NAD⁺ (Horenstein et al., 2013). Bone marrow clonogenic niches are enriched in such enzymatic activities (Quarona et al., 2015), however, it remains to be assessed whether similar mechanism is active at NVU/BBB. At least, very recent hypothesis announced in (Panfoli et al., 2016) stated that elevated levels of adenosine in the brain and blood due to endothelium-driven metabolism of ATP to adenosine in premature infants might be a biomarker of prematurity risk.

Circulating EC might be found in the peripheral blood due to endothelial injury, anoikis, or apoptosis. These cells express the markers of mature differentiating EC (i.e., CD31 and vWF but not CD133), and could also contribute to vascular repair (Tenreiro et al., 2016). Genomic studies revealed that expression patterns of RNAs and miRNAs in EPC and EC are different and reflect significant changes in the functional role of these two types of cells in development and maintenance of endothelium competence (Cheng et al., 2013; Chang et al., 2014).

Pericytes surrounding and supporting endothelial cells share some common properties with EPC (or with bone marrow stromal cells, BMSC): expression of adhesion molecules and VEGFR2, angiopoietin signaling, and prominent pro-angiogenic potential (Bagley et al., 2005; Winkler et al., 2014; Cantoni et al., 2015). That is not surprising due to bone marrow origin of pericytes (Lamagna and Bergers, 2006) and their potential to differentiate into various cell types or to originate from early EPC in vitro (Cantoni et al., 2015). Moreover, co-culture of EPC and bone marrow-derived stem cells in vitro (mesenchymal stem cells) results in the appearance of pericyte-like cells due to ability of bone marrow-derived stem cells to differentiate into pericytes with the paracrine regulatory activity of EPC (Loibl et al., 2014). In the context of BBB, pericytes origin and functioning are of great importance due to higher (10–30-fold) pericyte/endothelial cell ratio in the central nervous system (CNS) comparing to other tissues (Winkler et al., 2014).

Aberrant physiology of EPC is tightly implicated in the pathogenesis of various types of pathologies associated with endothelial dysfunction and vasculopathy including diabetes mellitus, Alzheimer's disease, cerebrovascular and cardiovascular diseases etc. (Lee et al., 2010; Brea et al., 2011; Yiu and Tse, 2014; Berezin and Kremzer, 2015). Circulating CD146(+) EPC as well as BMEC have been proposed as novel biomarkers of BBB impairment in neuroinfection and neurotoxicity (Huang et al., 2013a) applicable for diagnostic purposes. At the same time, there is a growing evidence that EPC might serve as a therapeutic tool for a number of CNS pathologies including stroke, neurodevelopmental disorders, and neurodegeneration (Castillo-Melendez et al., 2013; Fukuda et al., 2013; Zhao et al., 2013). Very promising data have been obtained in the application of EPC for BBB repair in vivo (Huang et al., 2013b). However, application of EPC into routine neurological clinical practice is hampered by poor understanding the mechanisms underlying EPC homing at CNS and EPC-mediated cell-to-cell communications in cerebral microvessels within the NVU.

Migration of EPC to the BBB is dictated by numerous factors contributing to elevated permeability of BBB that might serve as signals for EPC mobilization and moving toward the brain tissue. Studying brain angiogenesis and BBB establishment and maturation (so-called barriergenesis) offers some unique opportunities to distinguish a role for EPC in either developmental or pathological angiogenesis since both these processes have different mechanisms and functional significance for the developing, mature and aging brain (Vallon et al., 2014). Metabolic activity of EPC and related cells with pro-angiogenic potential is not well-studied yet. There is no doubt that metabolism of EPC and metabolic microenvironment at the sites of their destination affect efficacy of angiogenesis. As an example, remodeling of primary capillary plexus and vasculogenesis during embryogenesis depend on the actual metabolic demands of the defined tissue (Yoder, 2012). Particularly, EPC metabolic activity in the resting state and upon stimulatory conditions (i.e., expansion in the bone marrow, mobilization, and targeted migration) differs. The same should be true for BMEC activated in angiogenesis and converted to tip or stalk phenotype.

Thus, it is reasonable to propose that local microenvironment in the bone marrow supporting EPC maintenance, expansion, and mobilization would have some similarities with the microenvironment providing within the NVU to support BMEC functional activity, BBB establishment and repair. Since elevated permeability of BBB is associated with the establishment of sites of active neurogenesis (Lin et al., 2015b), migration of EPC from the hematopoietic tissue to the adult BBB is a transfer from one clonogenic niche to another one. Upon coming to the brain microvessels, EPC incorporate in the endothelial layer or serve as a source of pro-angiogenic molecules in a way similar to pericytes and perivascular astrocytes, thereby establishing appropriate conditions for branching angiogenesis, restoration of BBB structural and functional integrity, and reparative neurogenesis (Figure 1).

Angiogenesis consists of cell proliferation, vessel sprouting, establishment of anastomosis, pruning and remodeling, acquisition of endothelial quiescence (Ehling et al., 2013). Antigenic and functional heterogeneity of endothelial progenitors determines the mechanisms of EPC-supported angiogenesis. In the brain, vascularization mainly occurs through angiogenesis (Grant and Janigro, 2006), therefore different populations of cells involved into angiogenic events contribute to the initiation of vessel growth, destabilization of extracellular matrix, establishment of novel microvessels and their maturation (Figure 2). In the context of BBB, the latter stage is corresponded to the maturation of the barrier and acquisition of adequate selective permeability and transporting activity. Presumably, all the earlier steps of angiogenesis are coupled with the elevated permeability of the BBB. Therefore, a reasonable question is how all these processes are linked to the metabolic plasticity of endothelial progenitors and other cellular components of the NVU (neurons and glia)?

Developmental angiogenesis and, probably, neuroplasticity-associated angiogenesis, are mainly regulated by local production of pro-angiogenic molecules (VEGF, TGFβ) and corresponding changes in the metabolism, proliferation, and differentiation status of migrating EPC governed by Wnt/β-catenin- and Notch-signaling (Vallon et al., 2014). In pathological angiogenesis associated with brain injury, the main stimuli provoking re-endothelialization are neuroinflammatory mediators (i.e., cytokines, growth factors) as well as hypoxia/ischemia-induced changes in the tissue metabolism that mobilize EPC from bone marrow and attract them to the brain tissue. As an example, membrane-bound Kit-ligand expressing on microvascular EC at the sites of inflammation provide effective homing of EPC to the activated endothelium (Dentelli et al., 2007).

It is reasonable to assume that neuroplasticity-associated angiogenesis might critically depend on the local branching provided by pre-existing BMEC acquiring tip or stalk phenotype, whereas developmental and pathological angiogenesis would require extensive mobilization and homing of bone marrow-derived endothelial progenitors. It is still under the debates which local metabolic factors in the brain are attributed to homing of EPC to the brain and their integration into the developing cerebral microvessels, or how it is corresponded to the metabolic pattern of EPC.

Development of relevant blood-brain models in vitro is one of the topic problems in modern neurobiology and neuropharmacology. The “ideal” model should match the following criteria: (i) combination of minimally required cell types critical for structural and functional BBB integrity; (ii) reconstruction of the key mechanisms responsible for selective permeability of the BBB; (iii) long-term preservation of functional and structural integrity during the model culturing in vitro; (iv) reproduction of specific properties of the defined cell population within the model in (patho)physiological conditions; (v) relative simplicity of assembling the barrier and assessment of its permeability.

Among all the approaches proposed to achieve the above-mentioned goals, using the stem cell-derived components of the BBB in the models in vitro appears as very promising solution, particularly, for some special tasks (i.e., modeling the developing BBB, assessment of BBB permeability corresponding to the neonatal period etc.). This direction is further actualized while thinking about microphysiological systems reflecting basic properties of biological tissues and organs in the miniature scale.

It is now possible to see the increasing significance of the knowledge of EPC contribution to BBB development and functioning in the context of microphysiological systems (MPS). MPS are the complex in vitro models of tissues and organs (combination of so-called “organoids”) aimed to establish their functional interrelations. In most the cases, it is the next step after the application of microfluidic technologies to the reconstruction of “organ-on-chip” or “body-on-chip” in the microenvironment close to the real one. Moreover, “physiome-on-chip” concept (Stokes et al., 2015) could be more beneficial in obtaining relevant physiological data using MPS approach.

It should be mentioned that microfluidics has influenced BBB modeling in vitro a lot: several successful attempts to establish functional BBB on the microfluidic platform have been reported (Booth and Kim, 2012; Griep et al., 2013; Cho et al., 2015). At the same time, there is also increasing interest in the utilization of stem cells-derived material for BBB constructing in vitro: development of astrocytes and neurons from neural progenitor cells (Lippmann et al., 2011; Khilazheva et al., 2015), development of BMEC from induced pluripotent stem cells (Lippmann et al., 2012), or even establishment of the barrier from BMEC and neural stem cells in vitro (Chou et al., 2014).

Complementary properties of both these approaches will make possible the reconstruction of stem cells-derived cell components of NVU/BBB in the improved microenvironment like it was recently suggested for “brain-on-chip” technologies (Alcendor et al., 2013; van der Helm et al., 2016) applicable for effective in vitro drug screening. In this context, reconstruction of brain microvasculature in a flow-directed MPS is one of the topic question which might be solved using EPC as not only a source of BMEC, but also as informative monitoring system to study angiogenesis/barriergenesis “on-line.”

So, what are the possible advantages in the application of EPC for NVU/BBB modeling in vitro and MPS designing? Establishment of vessel networks at the microfluidic platform and in MPS is performed by seeding EC (usually HUVEC, or BMEC if BBB model is reconstructing) as well as accessory cells, (i.e., pericytes or astrocytes) within the chamber channels. To provide efficient endothelial proliferation and vessel sprouting, chemical gradient of pro-angiogenic factors (i.e., VEGF) can be settled in the close vicinity to the growing vessel (Sakolish et al., 2016). Dynamic fluid flow mimics the conditions achieved in the real BBB and allows controlling in vivo-like vessel barriergenesis. In a case of BMEC, BBB-specific phenotype of the cells is one of the critical factors for proper modeling the barrier or its functional connection with other tissues/organoids within the MPS (Alcendor et al., 2013). Therefore, the following physiological and metabolic parameters of EPC or BMEC should be taken into the consideration in BBB models or MPS: (i) ability to establish functionally competent monolayer with BBB-specific properties (selective permeability, tight junction connections, expression of specific receptors and transporters) reproducible in either static or microfluidic conditions; (ii) high sensitivity to the action of pro-angiogenic stimuli (VEGF, MMP, lactate etc.) produced by surrounding cells; (iii) dynamic changes in cell metabolism upon action of glia-, pericyte-, or neuron-derived signals; (iv) ontogenesis-related changes in cell metabolism; (v) expression of molecules critical for CNS homing and acquisition of BBB-specific phenotype; (vi) ability to support neurogenesis within the neurogenic niches in vitro.

Translational prospects of BBB models in vitro or BBB-on-chip technologies dictate the urgent interest in selecting the most appropriate EC whose growth and functional activity would be relevant for in vitro testing and drug candidates and assessment of BBB disruption in various CNS disorders. Thus, application of EPC would be beneficial for the controlled growth of brain microvessel-like structures in the conditions close to the BBB microenvironment in order to facilitate further studies of NVU/BBB or cerebrovascular (patho)physiology (Cucullo et al., 2013), particularly, at early stages of ontogenesis. Also, EPC would overcome the problem of human brain microvascular cells availability for BBB modeling or MPS in vitro. Currently, few BBB models in vitro utilize human endothelial cells, and this problem hampers progress in preclinical pharmacological studies (Pamies et al., 2014). Human induced pluripotent stem cells are tested as a source of NVU components not only in the static BBB models (Lippmann et al., 2012) but also for the microfluidics-based technological solutions (Brown et al., 2015). Microfluidic approach might be also used for specific and efficient capture of EPC from the peripheral blood as it was demonstrated in cardiovascular bioengineering (Plouffe et al., 2009; Lin et al., 2015a). Then, development of “user friendly” protocols for in vitro differentiation of EPC obtained from the bone marrow or peripheral blood into BMEC, and for the establishment of NVU microenvironment supporting EPC integration in the growing microvessels would give us an opportunity to establish fully functioning endothelial layer with BBB characteristics. Solving these technological questions would provide novel engineering approaches toward controlling brain developmental and pathological angiogenesis, improved microfluidically-supported BBB models and/or brain multicompartment MPS suitable for translational studies in neurology and neuropharmacology.

Metabolic and functional plasticity of EPC controls their timely recruitment, precise homing to the brain microvessels, and efficient support of brain angiogenesis. Being under the control of numerous regulatory factors, EPC serve as a source of BMEC with BBB specific characteristics in the developing brain. In the adult brain, EPC contribute (directly or as a source and carrier of pro-angiogenic factors) to BBB re-endothelialization upon injury or to the cerebral angiogenesis associated with physiological conditions (i.e., activity-induced neurogenesis) or pathological conditions (i.e., tumor progression). A key challenge in the field of EPC physiology is the current shortage of the knowledge on the most efficient ways to manipulating their activity both in vitro and in vivo. However, solving this problem would be beneficial for the therapy of cerebrovascular diseases, brain trauma, neurodegeneration, brain tumors, and neuroinfection. Metabolic activity and functionality of EPC differs from those of BMEC and is tightly regulated by numerous systemic and local factors at all the steps of EPC development. Accumulating evidence suggests that further progress in studying BBB (patho)physiology, establishment of new therapeutic methods for reversible and controlled BBB opening, or designing the new in vitro assays for drug development (BBB models or microphysiological constructions) critically depends on deciphering molecular and biochemical mechanisms underlying EPC functional role in the developmental, pathological, and plasticity-driven cerebral angiogenesis.

NM Drafting the work for important intellectual content; YK Substantial contributions to the conception or design of the work; VS Final approval of the version to be published; AM Drafting the work for important intellectual content; ANS Substantial contributions to the conception or design of the work; YP Substantial contributions to the conception or design of the work; EB Drafting the work for important intellectual content; ABS Substantial contributions to the conception or design of the work, final approval of the version to be published.