Reviews
Discovery of a Vascular Endothelial Stem Cell (VESC) Population Required for Vascular Regeneration and Tissue Maintenance
2019 Volume 83 Issue 1 Pages 12-17
Details
2019 Volume 83 Issue 1 Pages 12-17
The roles that blood vessels play in the maintenance of organs and tissues in addition to the delivery of oxygen and nutrients are being gradually clarified. The maintenance of tissue-specific organ stem cells, such as hematopoietic and neuronal stem cells, is supported by endothelial cells (ECs), which represent an important component of the stem cell niche. The maintenance of organogenesis, for example, osteogenesis and liver generation/regeneration, is supported by molecules referred to as “angiocrine signals” secreted by EC. The mechanisms responsible for the well-known functions of blood vessels, such as thermoregulation and metabolism, especially removal of local metabolites, have now been determined at the molecular level. Following the development of single-cell genetic analysis, blood cell heterogeneity, especially of mural cell populations, has been established and tissue-specific blood vessel formation and function are now also understood at the molecular level. Among the heterogeneous populations of ECs, it seems that a stem cell population with the ability to maintain the production of ECs long-term is present in pre-existing blood vessels. Neovascularization by therapeutic angiogenesis yields benefits in many diseases, not only ischemic disease but also metabolic disease and other vascular diseases. Therefore, vascular endothelial stem cells should be considered to use in vascular regeneration therapy.
The fundamental role of blood vessels is clearly to maintain constant supplies of oxygen and nutrients to cells in tissues and organs, as well as the rapid deployment of immune cells to regions of infection or damage. In addition to these crucial roles, blood vessels are also involved in the formation of stem cell niches to support and maintain the immature phenotype and self-renewal capacity of organ-specific stem cells.1 Moreover, not only for stem cells, vascular endothelial cells (ECs) secrete several growth factors known as “angiocrine signals” to maintain the integrity of and assist in the regeneration of tissues/organs. One example of this angiocrine signaling is secretion of hepatocyte growth factor and Wnt2 from liver sinusoidal ECs to maintain hepatocytes.2 Another example is the function of Noggin from ECs in the bone marrow to support the turnover of osteocytes.3
In addition to the function of blood vessels for extravasation of oxygen and nutrients into tissues/organs, capillaries drain materials from organs into intraluminal cavities. Regulation of permeability is an important function of blood vessels throughout the body. Particularly in terms of their tissue-specific functions, control of intravasation and extravasation is extremely important in the brain, liver, lung and kidney. In the brain, removal of waste products such as amyloid β through the blood-brain barrier is critical for the maintenance of neuronal function and survival.4
Because regulation of blood vessel formation is directly related to the improvement of therapies, the development of agents targeting blood vessels is being vigorously pursued by many pharmacological companies. In order to develop drugs to induce or suppress new blood vessel formation, it is necessary to understand the cellular and molecular mechanisms of neovascularization. In the adult, it is well known that new blood vessel formation is mainly induced by sprouting angiogenesis from pre-existing blood vessels. Two decades ago, endothelial progenitor cells (EPCs) were identified in the bone marrow and found to promote neovascularization in ischemic regions.5 Although a direct contribution of EPCs to ECs is still controversial,6,7 injection of bone marrow cells into ischemic regions has been suggested to yield clinical benefit.8 My group recently discovered a vascular endothelial stem cell (VESC) population in pre-existing blood vessels and have documented its direct contribution to ECs in ischemia and tissue injury.9–12 By means of cell surface marker studies, a hierarchy of ECs traversing a differentiation pathway from immature VESCs to terminally-differentiated ECs has been identified,12 and here, I will review blood vessel formation and the role of these newly-discovered VESCs.
Blood pumped from the heart into the aorta flows through arteries to supply oxygen and nutrients through the capillaries, returning to the heart through venules and veins. Although according to the caliber of the vessel, the wall may be thinner or thicker, the structure is basically the same, consisting of 2 types of vascular cells. Most of the inner surface is lined with vascular ECs and mural cells adhering at the basal side of the ECs. In capillaries and some areas of venules, pericytes adhere to ECs, and vascular smooth muscle cells adhere to ECs in arteries and veins. Pericytes and vascular smooth muscle cells are collectively designated “mural” cells.
In capillaries, pericytes adhere to ECs sparsely in order to control permeability and allow ease of diffusion of materials through EC–EC junctions. On the other hand, venules act as gates for extravasation of leukocytes for immune responses to invasive challenges. In the venules, pericytes adhere densely to ECs, such that the ratio of ECs to pericytes is 1:1. It has been recently documented that leukocytes first migrate between ECs and subsequently enter the parenchyma through gaps between pericytes.13,14
Blood vessel formation is induced by 2 different processes. During embryogenesis, ECs developing from the mesoderm form tubes in situ where blood vessels are required by the body plan, and then mural cells are gradually recruited near the ECs. When the mural cells adhere to ECs, EC–EC junctions and EC–mural connections are formed and structurally stable blood vessels are created. This de novo vascular formation normally observed in embryos is called vasculogenesis.15
ECs developing from the mesoderm form vascular tubes through activation of vascular endothelial growth factor receptors (VEGFRs).16 Activation of VEGFRs on ECs plays an important role in proliferation, movement and matrix reconstitution during blood vessel formation; however, it also promotes vascular leakage. Vascular hyperpermeability is induced by phosphorylation of vascular endothelial (VE) cadherin. VEGFR activated by VEGF on ECs phosphorylates VE-cadherin through several signaling pathways initiated by src phosphorylation.17 Activated VE-cadherin is internalized into the cytoplasm of ECs, and results in weakening of the junctions between the cells. During physiological blood vessel formation, vascular leakage is inhibited by mural cell attachment to ECs. When these cells have weak junctions, they secrete platelet-derived growth factor (PDGF), mainly the PDGF-BB isoform, and recruit mural cells expressing the PDGF receptor β.16 During embryogenesis, mural cell populations produce angiopoietin-1, a ligand for Tie2 expressed on ECs. This promotes tight EC–EC junctions by cross-linking Tie2 with Tie2 on neighboring ECs and causing membrane localization of VE-cadherin by inhibiting activation of src.17,18 Finally, mural cells adhere to ECs for final maturation of blood vessels. The molecular mechanisms of mural cell adhesion to ECs have not been clearly identified; however, tight adhesion between ECs must be a trigger for this process.
In contrast to embryonic angiogenesis, new blood vessel formation after birth is usually induced by sprouting and extension of vascular branches from pre-existing blood vessels, so-called sprouting angiogenesis.16 This process is involved in the progression of many vascular diseases such as cancer, retinopathy and chronic inflammation. By controlling this process, therapeutic interventions in angiogenesis have been developed and are in clinical use.
Sprouting angiogenesis is the process by which ECs sprout from pre-existing blood vessels and extend to new branches.16 Initially, tip cells emerge from intraluminal lining ECs to guide the direction of migration of the new branch. Behind these tip cells, ECs (so-called “stalk cells”) with high proliferative capacity migrate to connect with tip cells and facilitate extension of the new branch. Finally, ECs (so-called “phalanx cells”) emerge and assist in the maturation of the new blood vessels by inducing tight adhesion between ECs and connections between ECs and mural cells.19,20
Stimulation of ECs in pre-existing blood vessels by VEGF induces secretion of delta-like 4 (DLL4), which activates Notch expressed on neighboring ECs. This inhibits the expression of Sox17, a transcriptional factor, resulting in transcriptional suppression of VEGFRs such as VEGFR2, VEGFR3, and neuropilin-1.21 Therefore, by this lateral inhibition, ECs strongly expressing DLL4 become tip cells and migration of neighboring ECs next to the tip cells is suppressed because they do not respond to VEGF. Here, the question arises in terms of the origin of stalk cells. ECs stimulated by DLL4 become non-responsive to VEGF because of decreased VEGFR expression; however, stalk cells behind tip cells proliferate in response to VEGF. This suggests that stalk cells come from another place where DLL4 cannot affect Notch expression by ECs and cannot adhere to tip cells for extension of new branches (Figure 1).
What is the origin of stalk cells? After stimulation with vascular endothelial growth factor (VEGF), tip cell candidates start to secrete DLL4 and stimulate neighboring endothelial cells (ECs). Notch activated in ECs by DLL4 reduces expression of Sox17, resulting in the suppression of VEGFR2, VEGFR3, and neuropilin-1 (NRP1) expression. Tip cell candidates alone respond to VEGF and become tip cells to guide the migration of new branches from pre-existing blood vessels. Behind tip cells, highly proliferating ECs emerge and adhere to tip cells. Neighboring ECs around tip cells may be non-responders to VEGF, which otherwise induces EC proliferation. In this case, where do stalk cells come from?
In the resting state, phenotypically different ECs such as arterial, venous, and lymphatic ECs are present. In addition, at least 3 types of different ECs (i.e., tip, stalk, and phalanx cells) emerge during angiogenesis; thus there is a great deal of heterogeneity among ECs. Moreover, ECs need plasticity because they must adapt to the requirements for tissue/organ-specific blood vessel formation.
To date, tissue/organ-specific stem cell populations such as hematopoietic, neuronal, gut and epithelial stem cells have been isolated and their contribution to tissue regeneration and integrity of tissue and organs has been elucidated.22 It is hypothesized that VESC populations may also be present in pre-existing blood vessels and could give rise to endothelial progenitors (stalk cells) connecting to tip cells for acute elongation of new blood vessels. Several lines of evidence have documented the existence of immature cells in the vascular wall that can give rise to vascular ECs and smooth muscle cells.
Adult (somatic) stem cells are cells that differentiate into progenitor cells with proliferative capacity and giving rise to terminally-differentiated tissue-specific cells.23 Several studies have reported that tissue-resident cells can differentiate into vascular lineage cells such as ECs and mural cells.24,25
The wall of a large blood vessel has 3 layers: the tunica intima, tunica media, and tunica adventitia. In mice, vascular progenitor cells derived from adult aorta have been identified in the tunica media.26 In that study, single-cell suspensions were found to contain side population (SP) cells, as determined by fluorescence-activated cell sorting (FACS) analysis. These were characterized as CD31−Sca1+ and could differentiate into ECs and smooth muscle cells in vitro. Other groups identified the presence of stem and progenitor cells in the adventitia. In one study, isolated Sca1+ cells in the adventitia of large and medium-sized arteries and veins from adult apoE-deficient mice were shown to be able to differentiate into smooth muscle cells, but not ECs, in vitro and in vivo.27 In another study, CD34+CD31−VEGFR2+ vascular wall cells were shown to form capillary sprouts in the region between the media and adventitia.28 Moreover, the presence of resident angiogenic mesenchymal stromal cells that can differentiate into ECs in vitro has also been documented in the same region.29 However, the contribution of these vascular cells to neovascular ECs in vivo has not been unequivocally demonstrated. Under certain specific culture conditions, some cells may transdifferentiate into the EC lineage from other cell types. From this point of view, identification of VESCs that are already committed to the EC lineage (and indeed proven to contribute to ECs in a pathological and physiological angiogenesis process) has long been and still is required.
Methods to analyze the cellular phenotype of ECs by flow cytometry are not well established. One reason for this is the difficulty in isolating single-cell suspensions from primary organs. In the case of the hematopoietic system, based on sorting bone marrow hematopoietic cells by different cell surface markers, stem cell populations defined by their bone marrow regenerative ability have been isolated. Thus far, cell surface markers such as c-kit, sca-1, and hematopoietic lineage markers have been generally used in this field of study.1 ECs and hematopoietic cells are derived from common progenitor (ancestor) cells, the so-called hemangioblasts; however, identification of stem cells committed to the endothelial lineage has not been accomplished using cell surface markers. Because of this lack of unequivocal stem cell markers for ECs, commonly used techniques for identification of stem cells need to be used (i.e., the SP method). SP cells were briefly mentioned before, but more detail is now provided.
SP cell isolation is a method of purifying stem cell populations using the universal property of high stem cell expression of drug efflux pumps such as ATP-binding cassette transporters. In this approach, cell suspensions are treated with the DNA-binding dye Hoechst33342. This fluorescent dye is taken up by most cells (main population [MP]), but some remain unlabeled (SP) and can be separated by FACS. Stem cells with high efflux capacity are enriched in the SP population. Hematopoietic stem cells have been enriched by this Hoechst efflux method using blue emission wavelength fluorescence,30 but subsequently, by using 2 different emission wavelengths, more enriched hematopoietic stem cell populations have been identified.31 Later, this method was used to isolate stem cells for which no cell surface markers had been identified (i.e., heart, mammary gland, liver, lung, testis, epidermis and muscle stem cells).32–38
In pre-existing blood vessels, a line of evidence previously indicated the presence of SP cells in the vascular wall and reported that they can differentiate into ECs. However, in that paper, the authors showed no living ECs in their single-cell suspensions, thus making it impossible to identify EC lineage-committed immature cells. Indeed, the technical aspects of manipulating blood vessels and making single-cell suspensions are challenging because ECs form such tight junctions with each other and adhere tightly to mural cells. Different methods of preparing single-cell suspensions have been explored and detection of SP cells in EC lineage-committed ECs of pre-existing blood vessels has been accomplished.9 Hence, the features of SP cells can be described in detail next.
CD31 is generally used as an EC marker, but it is expressed on some hematopoietic cells as well. Therefore, it is necessary to define ECs as CD31-positive and at the same time CD45 negative (CD31+CD45−). Among such CD31+CD45− ECs, approximately 1% are in the SP fraction in most organs that have been investigated, including liver, heart, lung, brain and muscle. The proliferative capacity of primary ECs from murine tissues and organs has been evaluated using OP9 osteoblastic stromal cells derived from OP/OP mice.39,40 When SP or MP cells (i.e., EC-SP cells or EC-MP cells) are cultured on OP9 cells, the latter rarely generate EC colonies, and those that form are small (the frequency of colony-forming units is approximately 1/100 cells), whereas the former generate very large EC colonies at a frequency of approximately 1 in 8 SP cells. Cells passaged from EC-SP cell primary cultures are able to generate secondary EC colonies on fresh OP9 cells but cells from EC-MP colonies are unable to do so.
On cell-cycle analysis, EC-SP cells are dormant, but they initiate RNA synthesis when exposed to ischemia. The high drug efflux capacity of EC-SP cells is consistent with their high levels of expression of ABC transporters (i.e., ABCG2, ABCB1a, ABCB2, and ABCA5). In contrast, expression of other endothelial-related molecules is similar between EC-SP cells and EC-MP cells. Similar levels of VE-cadherin, VEGFR2, and CD34 suggest that EC-SP cells are indeed EC lineage-committed ECs.
Occlusion of the femoral artery causing hindlimb ischemia is a frequently used mouse model of neovascularization. In this model, it was found that EC-MP cells could not generate functional blood vessels but ECs derived from EC-SP cells were able to generate entire tubular formations of new blood vessels containing large amounts of red blood cells and connecting with recipient blood vessels. Moreover, blood vessels in the hindlimb were gradually replaced by ECs from EC-SP cells over the long-term (≥6 months).
It has been reported that EPCs are located in the bone marrow and therefore it is possible that EC-SP cells originate from bone marrow. However, as far as my group could determine, there were no EC-SP cells in the bone marrow, as shown in a mouse model of bone marrow transplantation in neonates and adults. Localization of dormant EC-SP cells reveals that most of these cells localize to the intraluminal inner surface of the vessel but not distant from blood vessels (Figure 2).
Resident endothelial stem cell population localized on the intraluminal surface of blood vessels, where EC-SP cells produce EC progenitors and contribute to newly-developed blood vessels as terminally-differentiated ECs. EC, endothelial cell; SP, side population.
Analyzing EC-SP cells in tumors, taken as an ischemic tissue, an approximately 10-fold higher frequency of these cells relative to normal tissues (0.7% in normal lung vs. 7% in lung cancer) has been found.11 As ischemia induces RNA synthesis in EC-SP cells in the hindlimb ischemia model, ischemia must be one of the triggers promoting proliferation (self-renewal) of vascular EC-SP cells. Thus, it can be concluded that EC-SP cells may be the source of stalk cells during sprouting angiogenesis and contribute to acute proliferation of ECs.
Detection of stem cell populations by the Hoechst method is a promising approach, but it is context dependent. In order to prove the stemness of ECs, their hierarchy from stem cells to terminally-differentiated ECs and their self-renewal potential should ideally be proven. To do this definitively, injection of a single cell should be shown to cause a long-term vascular EC contribution and expansion as a stem cell population. The hindlimb ischemia model is not appropriate for demonstrating single-cell transplantation reconstitution ability. Using a liver injury model of monocrotaline to induce damage to liver sinusoidal ECs, it was found that a single EC-SP cell from the liver could reconstitute ECs long-term. Which molecules are expressed on EC-SP cells specifically has also been investigated and 2 markers, CD200 and CD157,12 which are highly expressed by VESCs have been found.
CD200 is an immunoglobulin superfamily membrane-anchored glycoprotein containing 2 immunoglobulin domains and 1 transmembrane domain. CD200 is expressed by some leukocytes, especially B cells,41 but its function has not been fully clarified. CD157 is a glycosylphosphatidylinositol-binding membrane glycoprotein with a 33% amino acid homology to CD38. This molecule has ADP-ribosyl cyclase activity and promotes pre-B cell growth.42 ECs from the liver could be fractioned into different populations according to their expression of these 2 molecules: CD200+CD157+, CD200+CD157− and CD200−CD157− ECs. It was found that 70% of SP cells express both CD200 and CD157. When cultured on an OP9 stromal layer, CD200+CD157+ ECs generate higher number of EC colonies than CD200+CD157− ECs. The latter do also generate EC colonies but fewer in number and smaller in size. In contrast, CD200−CD157− ECs were completely unable to form any EC colonies. The CD200+CD157+ ECs contributed long-term to ECs after massive monocrotaline-induced liver injury. CD200+CD157− ECs also formed blood vessels in the damaged liver, but at much lower efficiency. As observed in vitro, CD200−CD157− ECs were completely unable to contribute to new blood vessels in the liver. When a single CD200+CD157+ EC was transplanted into mice with liver injury, it generated CD200+CD157+ ECs and many differentiated CD200−CD157− ECs via a stage where cells were CD200+CD157−. Therefore, it can be concluded that CD200+CD157+ ECs are at the top of the EC hierarchy of self-renewing stem cells, and that they maintain the vasculature by supplying terminally differentiating ECs through a stage of CD200+CD157− endothelial progenitors (Figure 3).
Hierarchy of endothelial cell (ECs). CD157+CD200+ ECs are present as a stem cell population with self-renewal capacity, as demonstrated by the finding that a single cell can construct blood vessels that are sustainable long-term. CD157+CD200+ ECs differentiate into CD157−CD200− ECs through a CD157−CD200+ stage, possibly endothelial progenitor cells in the liver.
Based on the fact that CD200+CD157+ ECs constantly supply terminally-differentiated ECs and contribute to blood vessel formation long-term, it can be proposed that this type of VESC can be exploited for vascular regeneration therapy in ischemic patients. Moreover, when disease is caused by a lack of secretion of certain factors by ECs, transplantation of normal VESCs can restore the function of the blood vessels. It is well known that liver is the major source of blood coagulation factor VIII, a blood-clotting protein. Defects in the factor VIII-encoding gene (F8 gene) cause hemophilia A, a recessive X-linked coagulation disease. So far, it has been established that ECs in the liver or lymph nodes specifically produce factor VIII.43 Because it has been reported that EPCs localize to the bone marrow, whether factor VIII elevation is induced in leukemia patients with hemophilia A after bone marrow transplantation has been investigated. However, bone marrow cells may contribute to some vascular wall cells in the liver, and significant factor VIII elevation was not observed.44 On the other hand, it has been shown that transplantation of CD200+CD157+ VESCs from controls to mice with hemophilia A effectively improved bleeding episodes in the latter. Moreover, factor VIII elevation is continuously induced over the long-term because VESCs can maintain blood vessels induced by normal ECs.12 In the future, it is possible that VESC transplantation may become a curative treatment for hemophilia A.
The reason why bone marrow replacement therapy is useful in leukemia patients is because hematopoietic stem cells in the graft maintain an immature status with the ability to differentiate into all hematopoietic lineage cells. As observed in the hematopoietic system, in every tissue regeneration therapy, use of a stem cell population is critical for the continuous repopulation of newly-developed tissues. The vascular system is no exception and long-term sustainable blood vessel formation needs to be induced and maintained for therapeutic angiogenesis. In this context, the discovery of VESCs will significantly contribute to future therapeutic angiogenesis therapy.
Although the usefulness of VESCs in tissue regeneration is clear, the number of VESCs in pre-existing blood vessels is very low. For these cells to be used effectively, their in vitro or ex vivo expansion or induction of VESCs from ES cells or induced pluripotent stem cells may be required.
The study of VESCs is still in its infancy. After the identification of CD157+ VESCs in mice, many researchers are now focusing on vascular stem cell populations in pre-existing blood vessels. Recently, it was reported that existing EC stem cell-like cells in the aorta facilitate recovery from massive intraluminal vascular injury.45 Moreover, as reported, liver-resident EC lineage cells alone, but not bone marrow cells, can rescue vascular damage occurring in the liver.46 In addition to the heterogeneity of ECs during angiogenesis (i.e., tip, stalk, and phalanx cells), the concept of VESC populations needs to be considered when seeking to clarify the precise molecular mechanisms of angiogenesis.
I thank all people involved in the isolation of endothelial stem cell populations in pre-existing blood vessels. This work was partly supported by the Japan Agency for Medical Research and Development (AMED) under Grant number (18 cm0106508 h0002, 18 gm5010002s110), Japan Society for the Promotion of Science (JSPS) Grants-in-Aid for Scientific Research (A) (16H02470). This research was supported by Integrated Frontier Research for Medical Science Division, Institute for Open and Transdisciplinary Research Initiatives, Osaka University.
