Molecular regulation of vessel maturation


	Journal home

	Advance online publication

	Current issue

	Archive

	Press releases



	Supplements

	Focuses

Guide to authors

Reprints and permissions

About this site

For librarians


	NPG Resources

	Nature

	Nature Reviews

	Nature Immunology

	Nature Cell Biology

	Nature Genetics

	news@nature.com

	Nature Conferences

	Dissect Medicine

Molecular Cell Biology

Neuroscience

Pharmacology

Physics

Browse all publications

Review

Molecular regulation of vessel maturation

Rakesh K Jain

E.L. Steele Laboratory, Department of Radiation Oncology, Massachusetts General Hospital and Harvard Medical School, 100 Blossom Street, Boston, Massachusetts 02114, USA. jain@steele.mgh.harvard.edu

The maturation of nascent vasculature, formed by vasculogenesis or angiogenesis, requires recruitment of mural cells, generation of an extracellular matrix and specialization of the vessel wall for structural support and regulation of vessel function. In addition, the vascular network must be organized so that all the parenchymal cells receive adequate nutrients. All of these processes are orchestrated by physical forces as well as by a constellation of ligands and receptors whose spatio-temporal patterns of expression and concentration are tightly regulated. Inappropriate levels of these physical forces or molecules produce an abnormal vasculature—a hallmark of various pathologies. Normalization of the abnormal vasculature can facilitate drug delivery to tumors and formation of a mature vasculature can help realize the promise of therapeutic angiogenesis and tissue engineering.

Blood vessels deliver nutrients and other molecules, as well as blood and immune cells, to all tissues in our body. The vascular network branches in a hierarchical fashion and is organized spatially to provide adequate nutrients to the cells of all organs and supporting structures by diffusion and convection. The walls of vessels are composed of endothelial cells and mural cells, which are embedded in an extracellular matrix (ECM; Fig. 1). The origin, number, type and organization of mural cells, the composition of the associated matrix, and the connection between the vascular system and the nervous system depend on the location of the vessel and its function. To reach this level of complex organization, the immature vascular network, formed by vasculogenesis or angiogenesis, must mature at the level of the vessel wall as well as at the network level. Maturation of the wall involves recruitment of mural cells, development of the surrounding matrix and elastic laminae, and organ-specific specialization of endothelial cells (ECs), mural cells and matrix (such as interendothelial junctions, fenestrations, apical-basal polarity, surface receptors and foot processes; Fig. 1). Maturation of the network involves optimal patterning of the network by branching, expanding and pruning to meet local demands (Fig. 2). Here we will focus on our current understanding of the molecular and cellular players involved in blood vessel maturation. Recent studies have provided important new discoveries about the ways in which these players govern vessel maturation during embryogenesis and during physiological processes, as well as failures of maturation in various pathologies. Although key questions remain unanswered for both blood and lymphatic vessel development, these discoveries could lead to new approaches for both therapeutic angiogenesis and antiangiogenic therapy^1,².

Figure 1. Wall composition of nascent versus mature vessels.

	(a) Nascent vessels consist of a tube of ECs. These mature into the specialized structures of capillaries, arteries and veins. (b) Capillaries, the most abundant vessels in our body, consist of ECs surrounded by basement membrane and a sparse layer of pericytes embedded within the EC basement membrane. Because of their wall structure and large surface-area-to-volume ratio, these vessels form the main site of exchange of nutrients between blood and tissue. Depending upon the organ or tissue, the capillary endothelial layer is continuous (as in muscle), fenestrated (as in kidney or endocrine glands) or discontinuous (as in liver sinusoids). The endothelia of the blood-brain barrier or blood-retina barrier are further specialized to include tight junctions, and are thus impermeable to various molecules. (c) Arterioles and venules have an increased coverage of mural cells compared with capillaries. Precapillary arterioles are completely invested with vascular SMCs that form their own basement membrane and are circumferentially arranged, closely packed and tightly associated with the endothelium. Extravasation of macromolecules and cells from the blood stream typically occurs from postcapillary venules. (d) The walls of larger vessels consists of three specialized layers: an intima composed of endothelial cells, a media of SMCs and an adventitia of fibroblasts, together with matrix and elastic laminae. The advential layer has its own blood supply, known as vasa vasorum, that extends in part into the media. SMCs and elastic laminae contribute to the vessel tone and mediate the control of vessel diameter and blood flow. Additional control of blood flow is provided by arterio-venous shunts that can divert blood away from a capillary bed when necessary. (e) Lymphatic capillaries lack pericytes. Larger (collecting) lymphatic vessels are invested in a basement membrane and contain valves that permit lymph flow only in one (proximal) direction; the lymphatic capillaries (initial lymphatics) contain microvalves in their walls⁹³. The lymphatic endothelial cells are connected to the surrounding connective tissue through anchoring filaments. Full Figure and legend (70K)

Figure 2. Steps in network formation and maturation during embryonic (physiological) angiogenesis (a) and tumor (pathological) angiogenesis (b).

	(a) The nascent vascular network forms from an initial cell plexus by processes of vasculogenesis or angiogenesis. This is regulated by cell-cell and cell-matrix signaling molecules (Box 2) and mechanical forces, as is further growth and expansion of the network (by proliferating and migrating cells) alongside its normal remodeling by cell death (apoptosis). Ordered patterns of growth, organization and specialization (including the investment of vascular channels by mural cells) produce mature networks of arteries, capillaries and veins—networks that are structurally and functionally stable and appropriate to organ and location. (b) ECs and mural cells derived from circulating precursor cells or from the host vasculature form networks that are structurally and functionally abnormal. Continual remodeling by inappropriate patterns of growth and regression (cell apoptosis and necrosis) contribute to the instability of these networks. Full Figure and legend (87K)

Vessel maturation in embryonic development
During development, the vasculature forms by both vasculogenesis and angiogenesis. In either case, the components of the vessel wall—endothelial cells, mural cells and matrix—originate from multiple sources (Box 1). In its simplest embodiment, vascular development can be thought of as involving the following processes: formation; stabilization; branching, remodeling and pruning; and specialization. The timing of most of these processes overlaps, allowing the vasculature to evolve seamlessly to maturation. Furthermore, each of the molecules involved has multiple functions in the development of a mature vascular network (Box 2; for a detailed description of each molecule, see Supplementary Table 1 online and refs. 3, 4, 5, 6, 7).

Formation of immature vasculature. During embryonic development, the nascent vascular network is formed by vasculogenesis (de novo vessel formation from angioblasts or stem cells) as well as angiogenesis (sprouting, bridging and intussusceptive growth from existing vessels). Vascular endothelial growth factor (VEGF) signaling is an important aspect of this process^8,⁹. VEGFA not only initiates vessel formation, but also sets in motion a chain of molecular and cellular events that ultimately lead to a mature vascular network^3,^4,^5,^6,⁷. In brief, CD31⁺CD34⁺ VEGF receptor (VEGFR)-2-positive angioblasts form a vascular plexus that gives rise to the dorsal aorta, the cardinal vein and the embryonic stems of yolk sac arteries and veins. Sprouting angiogenesis is presumably facilitated by hypoxia, which upregulates expression of a number of genes involved in vessel formation, patterning and maturation, including nitric oxide synthase, VEGF, and angiopoietin-2. Existing vessels dilate in response to nitric oxide, a product of nitric oxide synthase, and become leaky in response to VEGF. The basement membrane and ECM dissolve in response to activation of proteases (such as matrix metalloproteinase (MMP)2, MMP3 and MMP9) and suppression of protease inhibitors (such as tissue inhibitor of metalloproteinase-2). Plasma proteins leaked from these nascent vessels serve as a provisional matrix. ECs migrate through interactions between integrins and the matrix, and proliferate in response to VEGF and other endothelial cell mitogens. Angiopoietin (Ang)2 facilitates sprout formation in the presence of VEGF. The sprouts anastomose to form vascular loops and networks.

Stabilization of immature vasculature. The nascent vessels are stabilized by recruiting mural cells and by generating ECM. At least four molecular pathways are involved in regulating this process: platelet-derived growth factor (PDGF) B−PDGF receptor (PDGFR)- beta

; sphingosine-1-phosphate-1 (S1P1)−endothelial differentiation sphingolipid G-protein-coupled receptor-1 (EDG1)); Ang1-Tie2; and transforming growth factor (TGF)- beta

. PDGFB is secreted by ECs, presumably in response to VEGF, and facilitates recruitment of mural cells. Although PDGFB is expressed by a number of cells, including ECs and mural cells, signaling through PDGFR- beta

, which is expressed on mural cells, is responsible for their proliferation and migration during vascular maturation¹⁰. Compelling support for this hypothesis comes from studies of Pdgfb knockout mice, which undergo embryonic lethality, lack pericytes in certain vessels and exhibit microvascular aneurysm (Supplementary Table 1 online).

The similarity between the phenotypes of Pdgfb-Pdgfrb and Edg1 knockout mice (failure of mural cells to migrate to blood vessels) indicates that signaling through the EDG1 receptor, which is expressed by mural cells, is another key pathway for mural cell recruitment¹¹. EDG1 receptor signaling may occur downstream of PDGF signaling, although this hypothesis has recently been questioned¹². Alternatively, the lack of EDG1 may alter the EC matrix production or EC−mural cell interaction, and interfere with vessel maturation. Targeted deletion of G alpha

_i13, a molecule downstream of EDG3, also yields a Edg1 knockout phenotype, and both PDGFB- and PDGFR- beta

-deficient mice exhibit markedly reduced expression of RGS5, a GTPase-activating protein for G alpha

_i, on their vascular plexi and small arteries¹³ (Supplementary Table 1 online).

Also critical for vessel formation and stabilization are the Tie receptors, Tie1 and Tie2, and two ligands for Tie2, Ang1 and Ang2 (refs. 4,14). Main sources of Ang1 and Ang2 are the mural cells and ECs, respectively. Ang1 is known to stabilize nascent vessels and make them leak-resistant, presumably by facilitating communication between ECs and mural cells. Notably, in the absence of mural cells, recombinant Ang1 restored a hierarchical order of the larger vessels, and rescued edema and hemorrhage, in the growing retinal vasculature of mouse neonates¹⁵. Thus, the mechanism of vessel maturation by Ang1 is far from clear. The role of Ang2 appears to be contextual. In the absence of VEGF, Ang2 acts as an antagonist of Ang1 and destabilizes vessels, ultimately leading to vessel regression. In the presence of VEGF, Ang2 facilitates vascular sprouting.

TGF- beta

1, a multifunctional cytokine, promotes vessel maturation by stimulating ECM production and by inducing differentiation of mesenchymal cells to mural cells^16,¹⁷. It is expressed in a number of cell types, including ECs and mural cells and, depending on the context and concentration, is both pro- and antiangiogenic¹⁸. Studies of knockout mice also underscore the importance of TGF- beta

1, its receptors (RI, RII and endoglin) and the downstream signaling molecules (ALK1, Smad5) in the initial phases of angiogenesis and vessel maturation^16,¹⁹ (Supplementary Table 1 online). Recent in vitro studies indicate that the TGF- beta

1−ALK1 pathway induces ECs and fibroblasts to express Id1, a protein required for proliferation and migration. On the other hand, the TGF- beta

1−ALK5 pathway induces the plasminogen activator inhibitor (PAI)1 in endothelial cells. PAI1 promotes vessel maturation by preventing degradation of the provisional matrix around the nascent vessel. Thus, the ratio of TGF- beta

signals through ALK1 versus ALK5 is likely to determine the pro- or antiangiogenic effect of TGF- beta

. One molecule that may orchestrate this balance is endoglin, a TGF- beta

−binding protein (type III receptor). Endoglin knockout mice exhibit normal vasculogenesis but undergo embryonic lethality as a result of defective vascular remodeling and smooth muscle cell (SMC) differentiation. Similarly, mutations in endoglin and ALK1 have been linked to human vascular disorders (hereditary hemorrhagic telengiectasia (HHT)-1 and HHT-2, respectively). Collectively, these studies indicate that TGF- beta

1, ALK1 and endoglin are positive regulators of endothelial cell migration and proliferation, whereas the TGF- beta

1−ALK5 pathway is a positive regulator of vessel maturation²⁰.

Branching, remodeling and pruning of vasculature. The final or optimal pattern of the vascular network for an organ is determined by the growth, branching, remodeling and pruning of its different segments. In addition to using signaling pathways that are involved in regulating branching in the nervous system (such as ephrins and neuropilins)^4,^21,²², various basement membrane and ECM components provide cues for these processes by regulating the proliferation, survival, migration and differentiation of ECs and mural cells (see accompanying review in Nature Reviews Cancer²³).

The matrix serves as a store for various growth factors and proenzymes involved in vessel development. The control of basement membrane and ECM degradation by proteases (such as MMP2, MMP3, MMP9, and urokinase plasminogen activator) and their inhibitors (such as tissue inhibitors of metalloproteinases and PAI1) influences EC and mural-cell migration. These proteases also release various proangiogenic growth factors (such as VEGF and basic fibroblast growth factor (FGF)) that are sequestered in the matrix, and generate antiangiogenic molecules by cleaving plasma proteins (such as angiostatin from plasminogen), matrix molecules (such as tumstatin from collagen type IV), or the proteases themselves (such as PEX from MMP2). The spatial and temporal concentration profiles of these growth factors and protein fragments, determined by their transport and binding to the matrix, presumably affect the branching pattern of vessels by regulating proliferation and apoptosis of endothelial cells and mural cells (Supplementary Table 1 online).

Advances in our understanding of integrins provide clues about the ways in which different matrix components influence EC survival and migration. For example, both pharmacological and genetic approaches indicate that fibronectin, its receptor alpha

₅

₁, collagen I, and collagen receptors alpha

₁

₁ and

₂

₁, are proangiogenic. Similarly, both pharmacologic and genetic approaches show that Tsp1 and Tsp2 are powerful inhibitors of angiogenesis and can exert this effect through intergrins and proteases²⁴. Paradoxically, unlike the pharmacologic interventions, the disruption of genes encoding alpha

₃ and

₅ (integrin receptors for fibronectin, vitronectin, fibrinogen, osteopontin, thrombospondin, endostatin and von Willebrand factor) does not block angiogenesis^25,²⁶. Also, unlike administration of endostatin (a fragment of the basement-membrane collagen XVIII), disruption of the gene encoding collagen XVIII does not affect angiogenesis²⁷. Further investigations are needed to provide a unified framework for the complex role of cell-matrix interactions in vessel formation and maturation^23,^26,²⁸.

Vessel specialization. The least understood step in the maturation process is the tissue- and organ-specific specialization of wall and network structure (Figs. 1 and 2)²⁹. This process includes arterio-venous determination, formation of homotypic and heterotypic junctions, and EC differentiation to form organ-specific capillary structures. Based on the observation that the veins grafted on the arterial side of the vascular network develop an arterial wall structure, it was initially assumed that flow (shear stress) was the determinant of arterio-venous specification. However, recent findings from ephrin knockout mice indicate that arterio-venous specification is genetically determined, and that an arterio-capillary-venous arrangement is completed before the heart starts pumping blood (Supplementary Table 1 online). Compelling evidence from knockout mice and zebrafish, as well as analysis of two human diseases, Alagille syndrome and cerebral autosomal-dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL), indicate that the Notch pathway determines the arterial-venous fate of ECs, possibly by committing angioblasts to these two lineages⁷. As the capillary plexus forms, bidirectional ephrin and ephrin receptor signaling repels the arterial and venous sides, and thus guides branching. Arterial differentiation may then be further promoted by TGF- beta

1−Alk1 signaling, and venous differentiation suppressed by Notch signaling. Continued arterial growth is then promoted by VEGF₁₆₄−VEGFR2−neuropilin (NRP)1 signaling. Lastly, expansion of larger arteries and veins occurs by acquisition of additional layers of mural cells, ECM and elastic laminae to provide the requisite viscoelastic properties and neural control.

Homotypic and heterotypic junctions, including EC−EC, EC−mural cell, and gap junctions, facilitate cell-to-cell communication and regulate vessel permeability. Vascular endothelial cadherin is an important component of EC-EC junctions, whereas neural (N)-cadherin facilitates EC−mural cell communication. Gap junctions made of connexins (such as Cx37, Cx40 and Cx43) also facilitate communication between ECs, and between ECs and perivascular cells. CD31, present on ECs and most leukocytes, is involved in angiogenesis as well as leukocyte extravasation through EC−EC junctions. Tight junctions made of occludins, claudins and zona occludens (ZO1, ZO2 and ZO3) contribute to the blood-tissue barrier in the brain and retinal capillaries. Finally, VEGF and endocrine gland−derived (EG)-VEGF induce fenestrations in ECs. How the local mechanical or biochemical microenvironment controls the formation of cell-cell junctions and leads to continuous, discontinuous, and fenestrated capillaries in different organs to meet local demands is yet to be determined.

Formation and maturation of lymphatic network
The vascular network in our body consists of not only blood vessels, but also of lymphatic vessels. The lymphatic vessels collect fluid, macromolecules, and immune cells that have extravasated from blood into the tissue. They also provide an important route of metastasis. Our understanding of the formation and maturation of lymphatic networks, however, is in its infancy in comparison to our understanding of blood vessels. Embryonic lymphatic vessels primarily originate from blood vessels. Lymphatic ECs originate from the cardinal vein during embryonic development. Molecular approaches have confirmed this notion in recent years, and have suggested additional sources of lymphatic ECs³⁰ which include lymphangioblasts³¹ and lymphatic endothelial precursor cells³². In the early embryo, ECs of the cardinal vein express lymphatic vascular endothelial receptor (LYVE)1 and VEGFR3. An as-yet-unknown signal triggers the expression of the homeobox gene Prox1, which commits these cells to the lymphatic lineage. These LYVE1-VEGFR3-Prox1−positive cells begin to sprout. Expression of secondary lymphoid cytokine and upregulation of VEGFR3 signals the formation of the lymphatic vessels. The Syk-SLP76 pathway triggers the separation of the embryonic lymphatic and blood vascular networks³³.Targeted deletion of Ang2 suggests that it is involved in the maturation and patterning of lymphatic vessels, and Ang1 can rescue this function of Ang2. Targeted deletion of NRP2 suggests that it is required for the formation of lymphatic capillaries but not large lymphatic vessels (Supplementary Table 1 online). How the nascent lymphatics mature, develop valves, anchor to surrounding matrix and form a functional lymphatic network is still enigmatic. A recent mouse model of lymphangiogenesis suggests that interstitial fluid flow guides the formation and pattern of lymphatic network³⁴. With the development of new animal models and the identification of new molecular players, our understanding of lymphangiogenesis and lymphatic maturation will advance rapidly.

Vessel maturation in physiological angiogenesis
Angiogenesis and the maturation of resulting vessels contribute to a number of physiological processes, including wound healing, reproductive cycling and ocular maturation. It is reasonable to assume that molecules involved in vessel formation and maturation during embryonic development are also involved in the postnatal period, but their precise role is not known because most knockout mice (Supplementary Table 1 online) die pre- or perinatally. Based on antibody blocking studies and gain-of-function studies, it appears that the spatio-temporal pattern of expression and concentration of various molecules involved may differ between the pre- and postnatal period. Changes in the local metabolic and mechanical microenvironment, such as the presence of hypoxia, low pH, abnormal hydrostatic pressure or shear stress, also profoundly influence the formation, maturation and remodeling of small and large vessels as a part of normal physiological processes. Information about the way in which these triggers alter the transcriptional profile of ECs and mural cells is beginning to shed light on the molecular pathways that underlie vascular remodeling and maturation in health and disease. (Further discussion of these pathways is beyond the scope of this review)^35,^36,^37,³⁸. Wound healing provides an example of the general principles involved in physiological vessel maturation.

After wound or tissue injury, activated platelets stimulate vessel growth by releasing a number of proteins, including TGF- beta

and PDGF³⁹. Formation of this granulation tissue is facilitated by chemotaxis of neutrophils, monocytes, fibroblasts, myofibroblasts and ECs⁴⁰. Fibroblasts initially secrete collagen III, followed by collagen I; once enough collagen is generated to allow wound closure, its synthesis is stopped. In the early stages of wound healing, a large number of immature vessels form. Later, some are pruned and the remaining vessels mature⁴¹. Formation of the lymphatic network follows that of the vascular network³⁴.

In cutaneous wound healing, intravital and immunohistochemical studies indicate that VEGF and Ang2 expression increase initially, and subsequently decrease to baseline levels after a stable vascular network is formed. A slight and transient decline in Ang1 expression is observed soon after wound formation, and a second decline is observed after vessel maturation⁴². During this process, sources of VEGF include keratinocytes, monocytes and fibroblast-like cells, whereas Ang1 is produced largely by pericytes. These data are consistent with the hypothesis that VEGF and Ang2 induce vessel formation, whereas Ang1 is involved in vessel stabilization by mediating EC−mural cell interactions. Surprisingly, the angiogenesis inhibitor endostatin, a collagen-XVIII fragment, impairs vessel maturation during wound healing without altering the expression of VEGF, Ang1 and Ang2 (ref. 42). Endostatin-treated wounds show a significant reduction in the number of functional vessels, as well as a lower expression of matrix molecules (collagens I and III and fibronectin). The reduced connective tissue density may improve the quality of the healed wound. This finding indicates that excessively large numbers of newly formed vessels may not be required for normal healing. In patients with diabetes, wound healing may be impaired due to deregulation of VEGF, PDGF, FGF and other growth factors⁴³. Further studies in conditional knockout mice are likely to provide mechanistic insights into physiological vessel formation and maturation.

Abnormal maturation in pathological angiogenesis
A large number of human diseases are characterized by abnormal vessels (see accompanying review in this issue⁴⁴). Here we will illustrate the general concepts using tumors as an example, as an abnormal vasculature is a hallmark of solid tumors (Fig. 2b)^45,⁴⁶. Tumor vessels are organized in a chaotic fashion and do not follow the hierarchical branching pattern of normal vascular networks (Fig. 3a,b). Whereas normal tissue maintains an equilibrium between vascular growth and cellular demands—no cells are farther from the nearest blood vessel than the distance the nutrients can diffuse before being completely consumed⁴⁷—the lack of such an equilibrium in tumors results in avascular, hypoxic voids of many sizes. The size and number of such voids, as measured by the fractal dimension, correspond to invasion percolation (a stochastic process in which a network expands around randomly placed obstacles; see Box 3)⁴⁸.

Figure 3. Vessel normalization and EC−mural cell interactions in tumors growing in dorsal windows in mice.

	(a) Normal capillary bed (dorsal skin and striated muscle). (b) Tumor vasculature (human tumor xenograft). (c,d) Anti-VEGFR2 therapy prunes immature vessels, leading to a progressively 'normalized' vasculature by days 1 and 2. (e) Further treatment leads to a vasculature that is inadequate to sustain tumor growth by day 5. (f) Perivascular cells expressing GFP (under the control of the VEGF promoter) envelope some vessels in the tumor interior. (g) A perivascular cell, presumably a fibroblast, leading the endothelial sprout (arrow). From ref. 53, with permission from Nature Medicine. Red, vessel lumens; green, perivascular cells. Scale bar (f,g) = 50 m. Images were obtained using a two-photon microscope. Full Figure and legend (97K)

The structure of the vessel wall is also abnormal in tumors. Vessel diameters are uneven, due in part to compression of the immature wall by proliferating tumor cells⁴⁹. Furthermore, ECs form an imperfect lining, with wide junctions at some locations and stacked layers of ECs at others. The ECs may contain a large number of fenestrations, vesicle vacuolar organelles (VVOs) or both. Some ECs do not express common endothelial markers (such as CD31) and may undergo apoptosis, thus exposing cancer cells to the lumen (so-called mosaic vessels)⁵⁰. The expression of adhesion molecules is also more heterogeneous than in normal tissue. In some tumor regions, tumor necrosis factor- alpha

or VEGF upregulates the expression of adhesion molecules such as intercellular adhesion molecule-1, vascular cell adhesion molecule-1 and E-selectin. In others, TGF- beta

, basic FGF or Ang1 downregulate adhesion molecule expression⁵¹. This heterogeneity makes targeting tumor vessels a challenge.

The data on the mural-cell coverage of tumor vessels is somewhat controversial, with some studies showing a paucity of mural cells and others showing their abundance⁵². Whether this is due to the different tumor types examined or the different molecular markers (such as alpha

-smooth muscle actin and desmin) used to identify these cells is unknown: also unknown is the origin of the mural cells. One possibility is that fibroblasts at the tumor-host interface are triggered by components of the tumor microenvironment (such as TGF- beta

) to differentiate first into myofibroblasts, and then into pericyte-like cells. Indeed, these pericyte-like cells have been proposed to guide the endothelial sprouts in tumors^52,⁵³. Both intravital and immunostaining studies show that the tumor-associated pericytes have abnormal morphology and form tenuous contacts with the ECs and the matrix (Fig. 3f,g). Intravital studies also show that these perivascular cells produce VEGF, which can serve as a survival factor for the ECs and make these vessels leaky^53,⁵⁴. These contrasting roles of the tumor-associated pericyte—stabilizing versus rendering tumor vessels leaky—are enigmatic and need further investigation.

As a result of the abnormal organization and ultrastructure of tumor vessels, the blood flow in tumor vessels is chaotic and the vessels are leaky^48,^55,⁵⁶. Furthermore, because of continuous remodeling of the vasculature, blood flow and permeability vary between tumors, between a tumor and its metastases, and within a given tumor from one location to another and from one day to the next in a single location. These abnormalities were initially attributed to overexpression of VEGF. Now it is widely accepted that these abnormalities result from an imbalance between levels of pro- and antiangiogenic molecules⁵⁷. More specifically, members of the VEGF family have been shown to increase permeability and vessel diameter, and Ang1 and Tsp1, in some contexts, are thought to decrease permeability and vessel diameter. The relative contribution of different cytokines to the formation and function of tumor vasculature is a fertile area of research⁵⁸.

The lymphatic vessels in tumors are also abnormal structurally and functionally, and contribute to interstitial hypertension in tumors⁵⁹. Although LYVE-1 positive structures are present within tumors, these do not necessarily function as lymphatics. The lymphatics in the tumor margin are functional and hyperplastic, however, presumably due to upregulation of VEGFC and VEGFD, and are adequate for metastasis through lymphatics⁶⁰.

Because an abnormal vasculature poses a formidable challenge to the delivery of nutrients and drugs to solid tumors, restoring the balance of pro- and antiangiogenic cytokines might 'normalize' the vasculature and facilitate the delivery of therapeutics⁶¹. Either hormone withdrawal from a hormone-dependent tumor (which lowers VEGF)⁶², herceptin treatment of a HER2-neu−overexpressing tumor (which impacts at least five angiogenic pathways)⁶³, or interfering with VEGF signaling⁶⁴ can prune immature vessels⁶⁵ and render remaining vessels more mature and efficient (Fig. 3c−e). Thus, judiciously applied antiangiogenic therapy may increase the efficacy of both conventional cytotoxic and newer specifically targeted therapies⁶¹. Alternatively, interfering with EC−mural cell interactions by targeting PDGFR- beta

or EDG1 signaling may destabilize tumor vessels and make them more vulnerable to anti-VEGF therapy⁶⁶. Thus, an improved understanding of EC−mural cell interactions in tumors will yield new antiangiogenic strategies.

Toward vessel maturation for therapeutic angiogenesis
The ability to form vascular networks in vivo by delivering angiogenic molecules as recombinant proteins (through controlled-release devices) or as genes (through plasmid DNA or viral vectors) offers hope for patients with various ischemic diseases^2,^67,⁶⁸. Clinical trials using members of the VEGF or FGF family are under way for the treatment of limb ischemia. Not surprisingly, some patients treated with VEGF for lower limb ischemia develop transient limb edema, presumably due to leakiness caused by VEGF⁶⁹. This is consistent with the preclinical data indicating that the vessels resulting from delivery of basic FGF or VEGF, through a controlled release device or an adenoviral vector, are stable for some time but eventually resemble immature tumor vessels^9,⁷⁰. Clearly, the formation of a mature vascular network requires precise spatial and temporal regulation of a large number of angiogenic stimulators and inhibitors. To this end, several molecular and cellular strategies are being tested in preclinical settings and Supplementary Table 2 online).

One approach is to deliver a cocktail of angiogenic molecules to a desired site in a carefully orchestrated spatio-temporal manner. Delivery of VEGF and PDGF in combination using a controlled-release polymeric device leads to vessels that appear more mature than those formed using VEGF alone⁷¹. Also, an adenoviral vector engineered to express Ang1 prevents VEGF-induced leakiness of vessels⁷². Unfortunately, these vessels have irregular diameters that may lead to impaired perfusion. To find the minimum number of molecules and to deliver them in an optimal fashion is therefore a major challenge for the field of therapeutic angiogenesis. One approach is to deliver a transcription or growth factor that regulates expression of the various growth factors required for mature vessel formation. As stated earlier, hypoxia activates the transcription factor hypoxia-inducible factor (HIF)-1 alpha

, which regulates the expression of a large number of pro- and antiangiogenic genes³⁶. Endogenous HIF-1 alpha

, however, is rapidly degraded in the presence of oxygen. Three strategies may overcome this problem. The first involves the delivery of PR39 peptide or the antimycotic ciclopirox olamine, both of which inhibit HIF-1 alpha

degradation and induce angiogenesis in ischemic mouse heart, mouse skin wounds and chicken chorioallantoic membrane^73,⁷⁴. Another approach is injection of naked DNA encoding HIF-1 alpha

−VP16, a stable fusion protein that induces angiogenesis and enhances perfusion in ischemic hind limb and myocardium^75,⁷⁶. Finally, transgenic expression of a mutant form of HIF-1 alpha

that lacks an oxygen degradation domain has been shown to promote the development of leak-resistant vessels in mouse skin⁷⁷. These promising results offer hope for therapeutic angiogenesis in patients with ischemic diseases.

Tissue engineering is another area of research where the formation of a functional, mature vascular network remains a challenge. Similar to normal tissue, engineered tissue requires blood vessels to grow and to remain viable. Initial efforts towards this goal used primitive vascular networks derived from endothelial cells growing in a matrix, in the presence or absence of an angiogenic factor (such as VEGF or basic FGF). When these gels were implanted in vivo, the engineered vessels connected with the host vessels. Unfortunately, the engineered vessels remained immature and did not survive for long. Two approaches have been used to overcome these problems. One involves the introduction of antiapoptotic (such as Bcl2) or antisenescence (such as human telomerase reverse transcriptase) genes into the ECs^78,⁷⁹. The resulting vessels last longer and appear to differentiate into arteries, capillaries and veins. It is not known to what extent these vessels continue to function as normal vessels, or how safe the introduction of these genes is over time.

An alternate approach is to form a primitive vascular network in vitro by coculturing ECs with mural cells or their precursors (such as 10T1/2 cells), and then to implant this network in vivo. This coculture system has been used extensively in vitro to gain molecular and functional insight into vessel maturation^80,⁸¹. Recently, these networks have been implanted in vivo and shown to integrate with the host vasculature, remain stable and function for more than one year (N. Koike et al., unpublished data). Furthermore, vascular grafts made of ECs and mural cells, cultured under pulsatile conditions, lead to functional arteries when implanted in miniature swine⁸². The possibility that such engineered vessels can be successfully incorporated into a regenerated tissue or into an organ is an exciting area of research for the field of regenerative medicine and tissue engineering.

A major challenge to any cell-based approach is the availability of an adequate number of ECs and mural cells to form mature vascular structures and networks. These cells can be derived from embryonic stem cells⁸³ or common progenitor cells^84,⁸⁵. However, these approaches present two problems: controlling differentiation and avoiding teratogenesis. Adult stem or progenitor cells, found in circulating blood, do not have teratogenic potential and maintain some degree of pluripotency^86,^87,^88,⁸⁹. Whether these adult EC and SMC progenitors can be expanded in a reproducible manner to a large number, and whether the resulting cells can form a functional vasculature, remains to be established. Another possibility is to expand and mobilize these cells in vivo to the site of angiogenesis. It is tempting to speculate that placental growth factor (PGF) may mobilize hematopoietic, perivascular and endothelial cells to the desired site. The recent finding that viral delivery of PGF leads to the formation of relatively stable, mature and functional vessels in an ischemic limb model supports this hypothesis^90,⁹¹.

Conclusions and unanswered questions
Research over the past decade has provided unprecedented insight into the molecular determinants of new vessel formation. Gene targeting approaches coupled with intravital microscopy allow us to dissect steps in vessel formation, maturation, remodeling and regression (Box 2). As a result we can create a primitive vascular network in vitro and in vivo and destroy a pathological vascular network—at least in part. Now the challenge is to create a mature (both structurally and functionally) network in vivo and to destroy a pathological network completely. The former will benefit regenerative medicine and treatment of various ischemic diseases; the latter the treatment of various angiogenic diseases, including cancer.

Meeting these challenges will require an integrative understanding of the cross-talk between various molecular and cellular players during pre- and postnatal vasculogenesis and angiogenesis. Mathematical modeling, bioinformatics and proteomics will aid this integration⁹². We will need to determine the minimum number of angiogenic stimulators and inhibitors, their dose and the sequence of delivery required to form mature vessels. We will also need to design viral or nonviral delivery systems to target the cocktail (or a master transcription factor) to the appropriate site, and to develop noninvasive imaging approaches to monitor the assembly and function of the engineered network in situ (both micro- and macroscopically). In cell-based approaches, we will need to isolate, differentiate and expand the appropriate progenitor or stem cell population in a reproducible way, and to test the structure and function of the resulting vessels. As our knowledge of the formation and maturation of lymphatic vessels increases, we will have to address similar issues to make functional lymphatic networks.

Likewise, in order for antiangiogenic therapy to be effective, we need to develop genomic and proteomic approaches to determine the minimum number of angiogenic pathways that must be targeted to slow tumor growth. Additionally, we need to develop noninvasive imaging technology to monitor the response to therapy. When antiangiogenic therapy is used in combination with cytotoxic therapy, we need to follow the dynamics of pruning of vessels, and find the normalization window during which cytotoxic therapy can be delivered most efficiently. With the rapid pace of research and development in the areas of vascular biology, proteomics and imaging, the future looks exciting.

Supplementary information is available on the Nature Medicine website.

	Top

REFERENCES

Folkman, J. Angiogenesis. in Harrison's Principles of International Medicine (eds. Braunwald, E. et al.) 517-530 (McGraw-Hill, New York, 2001).
Jain, R.K. & Carmeliet, P.F. Vessels of death or life. Sci. Am. 285, 38-45 (2001). | PubMed | ChemPort |
Carmeliet, P. Mechanisms of angiogenesis and arteriogenesis. Nat. Med. 6, 389-395 (2000). | Article | PubMed | ChemPort |
Yancopoulos, G.D. et al. Vascular-specific growth factors and blood vessel formation. Nature 407, 242-248 (2000). | Article | PubMed | ChemPort |
Carmeliet, P. & Jain, R.K. Angiogenesis in cancer and other diseases. Nature 407, 249-257 (2000). | Article | PubMed | ChemPort |
Nguyen, L.L. & D'Amore, P.A. Cellular interactions in vascular growth and differentiation. Int. Rev. Cytol. 204, 1-48 (2001). | PubMed | ChemPort |
Rossant, J. & Howard, L. Signaling pathways in vascular development. Annu. Rev. Cell Dev. Biol. 18, 541-573 (2002). | Article | PubMed | ChemPort |
Ferrara, N. VEGF and the quest for tumour angiogenesis factors. Nat. Rev. Cancer 2, 795-803 (2002). | Article | PubMed | ChemPort |
Nagy, J.A. et al. VEGF-A induces angiogenesis, arteriogenesis, lymphangiogenesis, and vascular malformations. Cold Spring Harbor Symposium on Quantitative Biology 67, 227-237 (2002).
Hellstrom, M. et al. Lack of pericytes leads to endothelial hyperplasia and abnormal vascular morphogenesis. J. Cell Biol. 153, 543-553 (2001). | Article | PubMed | ChemPort |
Kluk, M.J. & Hla, T. Signaling of sphingosine-1-phosphate via the S1P/EDG-family of G-protein-coupled receptors. Biochim. Biophys. Acta 1582, 72-80 (2002). | Article | PubMed | ChemPort |
Kluk, M.J., Colmont, C., Wu, M.T. & Hla, T. Platelet-derived growth factor (PDGF)-induced chemotaxis does not require the G protein-coupled receptor S1P1 in murine embryonic fibroblasts and vascular smooth muscle cells. FEBS Lett. 533, 25-28 (2003). | Article | PubMed | ChemPort |
Cho, H., Kozasa, T., Bondjers, C., Betsholtz, C. & Kehrl, J.H. Pericyte-specific expression of Rgs5: implications for PDGF and EDG receptor signaling during vascular maturation. FASEB J. 13, 440-442 (2003).
Loughna, S. & Sato, T.N. Angiopoietin and Tie signaling pathways in vascular development. Matrix Biol. 20, 319-325 (2001). | Article | PubMed | ChemPort |
Uemura, A. et al. Recombinant angiopoietin-1 restores higher-order architecture of growing blood vessels in mice in the absence of mural cells. J. Clin. Invest. 110, 1619-1628 (2002). | Article | PubMed | ChemPort |
Pepper, M.S. Transforming growth factor-: vasculogenesis, angiogenesis, and vessel wall integrity. Cytokine Growth Factor Rev. 8, 21-43 (1997). | Article | PubMed | ChemPort |
Chambers, R.C., Leoni, P., Kaminski, N., Laurent, G.J. & Heller, R.A. Global expression profiling of fibroblast responses to transforming growth factor-(1) reveals the induction of inhibitor of differentiation-1 and provides evidence of smooth muscle cell phenotypic switching. Am. J. Pathol. 162, 533-546 (2003). | PubMed | ChemPort |
Gohongi, T. et al. Tumor-host interactions in the gallbladder suppress distal angiogenesis and tumor growth: involvement of transforming growth factor 1. Nat. Med. 5, 1203-1208 (1999). | Article | PubMed | ChemPort |
Weinstein, M., Yang, X. & Deng, C. Functions of mammalian Smad genes as revealed by targeted gene disruption in mice. Cytokine Growth Factor Rev. 11, 49-58 (2000). | Article | PubMed | ChemPort |
Goumans, M.J. et al. Balancing the activation state of the endothelium via two distinct TGF- type I receptors. EMBO J. 21, 1743-1753 (2002). | Article | PubMed | ChemPort |
Mukouyama, Y.S., Shin, D., Britsch, S., Taniguchi, M. & Anderson, D.J. Sensory nerves determine the pattern of arterial differentiation and blood vessel branching in the skin. Cell 109, 693-705 (2002). | PubMed | ChemPort |
Neufeld, G. et al. The neuropilins: multifunctional semaphorin and VEGF receptors that modulate axon guidance and angiogenesis. Trends Cardiovasc. Med. 12, 13-19 (2002). | Article | PubMed | ChemPort |
Kalluri, R. Basement membranes: structural features, assembly, cellular interactions and role in cancer angiogenesis. Nat. Rev. Cancer (in the press).
Lawler, J. The functions of thrombospondin-1 and-2. Curr. Opin. Cell Biol. 12, 634-640 (2000). | Article | PubMed | ChemPort |
Hynes, R.O. A reevaluation of integrins as regulators of angiogenesis. Nat. Med. 8, 918-921 (2002). | Article | PubMed | ChemPort |
Stupack, D.G. & Cheresh, D.A. Get a ligand, get a life: integrins, signaling and cell survival. J. Cell Sci. 115, 3729-3738 (2002). | Article | PubMed | ChemPort |
Fukai, N. et al. Lack of collagen XVIII/endostatin results in eye abnormalities. EMBO J. 21, 1535-1544 (2002). | Article | PubMed | ChemPort |
Drake, C.J., Cheresh, D.A. & Little, C.D. An antagonist of integrin _V₃ prevents maturation of blood vessels during embryonic neovascularization. J. Cell Sci. 108, 2655-2661 (1995). | PubMed | ChemPort |
Ruoslahti, E. Specialization of tumour vasculature. Nat. Rev. Cancer 2, 83-90 (2002). | Article | PubMed |
Alitalo, K. & Carmeliet, P. Molecular mechanisms of lymphangiogenesis in health and disease. Cancer Cell 1, 219-227 (2002). | Article | PubMed | ChemPort |
Wilting, J., Tomarev, S. & Christ, B. Lymphangioblasts in embryonic lymphangiogenesis. Lymphatic Res. Biol. 1, 33-44 (2003). | Article |
Salven, P., Mustjoki, S., Alitalo, R., Alitalo, K. & Rafii, S. VEGFR-3 and CD133 identify a population of CD34⁺ lymphatic/vascular endothelial precursor cells. Blood 101, 168-172 (2003). | Article | PubMed | ChemPort |
Jain, R.K. & Padera, T.P. Development. Lymphatics make the break. Science 299, 209-210 (2003). | Article | PubMed | ChemPort |
Boardman, K.C. & Swartz, M.A. Interstitial flow as a guide for lymphangiogenesis. Circ. Res. 92, 801-808 (2003). | Article | PubMed | ChemPort |
Topper, J.N. & Gimbrone, M.A., Jr. Blood flow and vascular gene expression: fluid shear stress as a modulator of endothelial phenotype. Mol. Med. Today 5, 40-46 (1999). | Article | PubMed | ChemPort |
Semenza, G.L. HIF-1 and mechanisms of hypoxia sensing. Curr. Opin. Cell Biol. 13, 167-171 (2001). | Article | PubMed | ChemPort |
Xu, L., Fukumura, D. & Jain, R.K. Acidic extracellular pH induces vascular endothelial growth factor (VEGF) in human glioblastoma cells via ERK1/2 MAPK signaling pathway: mechanism of low pH-induced VEGF. J. Biol. Chem. 277, 11368-11374 (2002). | Article | PubMed | ChemPort |
Helisch, A. & Schaper, W. Arteriogenesis - the development and growth of collateral arteries. Microcirculation 10, 83-97 (2003). | Article | PubMed |
Coussens, L.M. & Werb, Z. Inflammation and cancer. Nature 420, 860-867 (2002). | Article | PubMed | ChemPort |
Tomasek, J.J., Gabbiani, G., Hinz, B., Chaponnier, C. & Brown, R.A. Myofibroblasts and mechano-regulation of connective tissue remodelling. Nat. Rev. Mol. Cell Biol. 3, 349-363 (2002). | Article | PubMed | ChemPort |
Zawicki, D.F., Jain, R.K., Schmid-Schoenbein, G.W. & Chien, S. Dynamics of neovascularization in normal tissue. Microvasc. Res. 21, 27-47 (1981). | PubMed | ChemPort |
Bloch, W. et al. The angiogenesis inhibitor endostatin impairs blood vessel maturation during wound healing. FASEB J. 14, 2373-2376 (2000). | PubMed | ChemPort |
Altavilla, D. et al. Inhibition of lipid peroxidation restores impaired vascular endothelial growth factor expression and stimulates wound healing and angiogenesis in the genetically diabetic mouse. Diabetes 50, 667-674 (2001). | PubMed | ChemPort |
Carmeliet, P. Angiogenesis in health and disease. Nat. Med. 9, 653-660 (2003). | Article |
Jain, R.K. Angiogenesis and lymphangiogenesis in tumors: insights from intravital microscopy. Cold Spring Harbor Symposium on Quantitative Biology 67, 239-248 (2002).
Jain, R.K., Munn, L.L. & Fukumura, D. Dissecting tumour pathophysiology using intravital microscopy. Nat. Rev. Cancer 2, 266-276 (2002). | Article | PubMed | ChemPort |
Helmlinger, G., Yuan, F., Dellian, M. & Jain, R.K. Interstitial pH and pO₂ gradients in solid tumors in vivo: high-resolution measurements reveal a lack of correlation. Nat. Med. 3, 177-182 (1997). | PubMed | ChemPort |
Baish, J.W. & Jain, R.K. Fractals and cancer. Cancer Res. 60, 3683-3688 (2000). | PubMed | ChemPort |
Helmlinger, G., Netti, P.A., Lichtenbeld, H.C., Melder, R.J. & Jain, R.K. Solid stress inhibits the growth of multicellular tumor spheroids. Nat. Biotechnol. 15, 778-783 (1997). | PubMed | ChemPort |
Chang, Y.S. et al. Mosaic blood vessels in tumors: frequency of cancer cells in contact with flowing blood. Proc. Natl. Acad. Sci. USA 97, 14608-14613 (2000). | Article | PubMed | ChemPort |
Melder, R.J. et al. During angiogenesis, vascular endothelial growth factor and basic fibroblast growth factor regulate natural killer cell adhesion to tumor endothelium. Nat. Med. 2, 992-997 (1996). | PubMed | ChemPort |
Morikawa, S. et al. Abnormalities in pericytes on blood vessels and endothelial sprouts in tumors. Am. J. Pathol. 160