Vascular Biology and the Skeleton


  • The authors have no conflict of interest


BLOOD VESSELS ARE organized in a hierarchical fashion to deliver oxygen, soluble factors, and various types of cells to all tissues in our body in a carefully regulated manner. The vascular network forms by both vasculogenesis (de novo vessel formation from angioblastic stem cells) and angiogenesis (sprouting from existing vessels). (1–3) Both processes are essential during skeletal embryogenesis or repair, but otherwise represent pathological events linked to adverse consequences.

The endothelium is one of the most important components of the vasculature, not only because it functions as an essential barrier that limits the movement of cells and molecules between the circulation and tissues, but also because it is a dynamic organ actively capable of directly communicating in a rich language with adjacent tissue and circulating blood cells. This appears evident when considering the remarkable heterogeneity of endothelial cells and their vessel size-specific, tissue-specific, and age-specific differences. (4, 5) What is the underlying cause of endothelial diversity? Certainly during development and ongoing tissue maintenance in the adult, circulatory and tissue components must continually communicate with endothelial cells. (5) However, the signaling between cells and endothelium within a tissue is, in fact, bidirectional, and endothelial cells just as readily communicate with surrounding cells through a host of humoral and growth factors, cytokines and chemokines, reactive metabolites, and polarized surface-associated molecules. (5)

Vasculature plays a major role also in the fabrication and homeostatic “health” of living bone, without which bone tissue dies and cannot be repaired or rejuvenated. (6, 7) The development of various “in vitro” models of bone endothelial cells has helped us to gain a better understanding of why the vasculature is so important in bone, what unique features characterize the bone endothelium, and how it functionally interacts with bone cells of different nature. (8–24) In particular, bone endothelium has been found to actively recruit circulating cells, direct hematopoietic cell homing to the bone marrow, and collaborate with various metastatic cells to selectively target them into bone, largely through the preferential display on bone endothelium of particular attractant signals that orchestrate such movement. (19, 20, 25–30) In addition, modifications in the blood supply correlate with numerous skeletal pathologies, including osteoporosis or osteopetrosis, inflammatory bone loss, and tumor-associated osteolysis. (31–33)

This overview is intended to highlight the potential roles of various paracrine interactions between bone endothelium and differentiated bone cells during organogenesis or in the mature skeleton, with brief consideration given to bone vascularization in pathological conditions. Potential implications of this growing body of information for medicine are noted.


The vasculature in bone is important for skeletal development and growth by affecting both bone modeling and remodeling processes. The majority of bones in the vertebrate skeleton are formed by endochondral ossification, a process initiated in the embryo by mesenchymal condensation, that gives rise to cartilage elements which, in turn, are invaded by blood vessels that bring along perivascular osteoprogenitor cells and recruit osteoclasts, leading to these elements being resorbed and gradually replaced by bone tissue. (34–36) The developmental pattern of the vascular network is an important component coordinating or directing limb morphogenesis. (37) Early in osteogenesis, the cartilage appears to produce anti-angiogenic factors that inhibit vessel penetration. (38–45) However, vascularization is required for osteogenesis to occur, and this is accomplished by a combination of factors, including adequate oxygen tension, compression forces, nutrients, and growth and differentiative factors. During or after vascular invasion, the hypertrophied cartilage core is degraded by chondroclasts/osteoclasts and replaced by bone marrow and later by bone initially deposited on calcified cartilage spicules. The vasculature may direct such new bone formation by serving as a scaffolding for bone-forming cells. Indeed, the intimacy of endothelial cells and osteoblasts has led in the past to the theory that the endothelial cell itself is an osteoblast precursor. (46) More recently, it has been proposed that bone marrow stromal cells (which are known to be capable of differentiating into osteoblasts) are actually a bone-specific type of microvascular pericyte, cells that are tightly affiliated with capillary endothelial cells and exhibit multilineage (osteogenic, chondrogenic, adipogenic, and fibroblastic) potential. (47–52) Thus, pericytes may be supplied by the invading vasculature and constitute a cellular reserve for use in bone tissue formation, remodeling, or repair. (47) Consistent with this, only cells that are situated near blood flow capillaries give rise to bone tissue, as mineral deposition is observed in the immediate vicinity of vessels in close correlation with the onset of cartilage vascularization. (53) Evidence also exists that endothelial progenitors, skeletal stem cells, and mesenchymal progenitors may be found in the mammalian circulation, from which they might be delivered to angiogenic sites. (54–58) Endothelial cells of the advancing capillaries also may participate both directly and indirectly in resorption during vascular invasion of the hypertrophied cartilage core by producing proteases and other regulatory or degradative molecules and by recruiting osteoclast precursors from the circulation. The invading vasculature therefore serves as both a reservoir and a conduit for the recruitment of essential cell types involved in cartilage/bone resorption and bone deposition/repair, and it regulates the functional activities of such cells, provides key signals necessary for bone morphogenesis, and seems responsible for the overall synchronization of endochondral ossification processes. (59–64)

The growth plates at the distal end are responsible for most longitudinal growth in some bones, such as the radius and ulna. The growth plate at the distal end is more active in the humerus and the tibia. Growth plates are composed mainly of chondrocytes organized in a proliferative zone, a prehypertrophic zone, and a hypertrophic apoptotic zone with a partially calcified matrix and invading vessels contained within canals emanating from the perichondral plexus surrounding the growth cartilage. (65) In its most complex form, the vascular structure within bone consists of an arteriole branching out into an anastomosing network of capillaries. (66) The capillaries circle back along the arteriole and rejoin into a single venule that follows the course of the arteriole back to the perichondral plexus. (67) The vascular canals serve three purposes to nourish chondrocytes, play a role in the induction and maintenance of the secondary ossification center, and potentially supply cartilage and bone tissue with mesenchymal cells. (68–70) As the animal grows, the ossification front of the secondary center of ossification advances toward the articular cartilage, making the layer of growth cartilage thinner. Concurrently, the cartilage canals undergo a physiological regression process known as chondrification, leaving the cartilage in this area avascular and ready for osteogenesis to ensue. (71) Vasculature invasion from the metaphysis coincides with apoptosis of hypertrophic chondrocytes(72) and their hypoxia-inducible factor 1 (HIF-1) driven production of vascular endothelial growth factor (VEGF), a critical regulatory signal stimulating vessel recruitment, chondrogenesis, and osteogenesis. (73) Hypertrophic chondrocytes also produce other important factors including RANKL, (74, 75) which not only promotes osteoclast formation, survival, and resorption activity, but has been shown recently to also stimulate angiogenesis(76) and endothelial cell survival. (77) Moreover, VEGF causes an upregulation of the receptor RANK in endothelial cells, thereby increasing their biological responsivity to RANKL. (78) Thus, these key signals of bone development seem intimately linked. The efficacy of VEGF antagonists to inhibit vascularization associated with bone growth(73) has potential clinical ramifications, opening the possibility of treating bone-associated pathologies such as osteosarcoma, osteoarthritis, rickets, and other disorders. (79, 80)

Angiogenesis is crucial for bone formation, remodeling, and healing. In onlay bone transfers, vascularized bone grafts retain their mass to a greater degree than nonvascularized tissue. (81) Moreover, electrical stimulation promotes bone formation with a parallel increase in the number of capillaries, (82, 83) whereas shock wave stimulated osteoblasts produce VEGF and stimulate angiogenesis, thereby contributing to bone regeneration. (84) Hypoxia (present in hypertrophic cartilage zones and inflammatory or bone fracture sites) also regulates osteoblast production of key modulators that influence endothelial proliferation, including VEGF and members of the TGF-β, IGF, and fibroblast growth factor families. (85–88) In fracture repair, osteoclasts have been reported to produce heparinase, an enzyme that releases VEGF from heparin in an active form, thereby contributing to local angiogenesis and osteoclast formation and activity that promotes callus resorption in the endochondral ossification stages of fracture healing. (89) Overall, angiogenesis is a critical feature of bone fracture healing and an adequate vascular supply is fundamentally required for the regenerative process to succeed, (6, 7, 90, 91) during which the early vascular response is exquisitely sensitive to initial micromovement. (92) Consistent with this, distraction osteogenesis is a vascular-dependent process accompanied by the robust induction of factors associated with angiogenesis, including VEGF. (93–96) Conversely, fracture healing is prevented by inhibiting angiogenesis. (97) Anoxia, a major component of the fracture microenvironment, has been reported to downregulate osteogenesis-associated bone morphogenic protein-2 (BMP-2) and Runx2 expression in osteoblastic cells and their precursors. (98) On the other hand, recent data suggest that endothelial cells subjected to mechanical, VEGF, or hypoxic treatments may play an osteogenic role through their production of BMP-4 and BMP-2, thereby potently stimulating osteoblast differentiation, bone formation, and bone fracture healing. (99–101) Such findings may provide not only an opportunity for a better understanding of vascular biology and fracture healing, but may also point the way toward new strategies for improving bone fracture repair. Attention directed at characterizing the humoral regulation of bone morphogenesis in the growth plate has led to the discovery in recent years of numerous other highly important or even critical local and systemic regulators of endochondral bone formation and vascularization in the growth plates. (68, 73, 102–111) Of particular note, bone repair contrasts with cartilage repair, in that the former requires angiogenesis and the latter is inhibited by it. (112) These and other studies have only begun to clarify how angiogenic events promote bone repair and regeneration. Hence, further studies will be necessary to uncover the mechanisms by which these codependent processes occur and to define the molecules, signaling pathways, and spatiotemporal relationships involved.


Despite the difficulties in separating bone endothelial cells from a tissue as hard as bone, a number of cellular models have been developed from different animal species, (8, 9, 18–24, 113) making it possible to characterize unique features of bone endothelium. For example, clonal bovine endothelial (BBE-1) cells, an immortal cell line generated from fetal bovine bone marrow cells, are distinctive in their ability to respond to PTH, a response not seen with endothelial cells obtained from other tissues. (8) Similarly, bone endothelial cells may respond to or produce systemic or local signal molecules in a tissue-specific manner. For instance, amongst six different human endothelial cell lines, only those of bone marrow origin secreted high levels of the hormone B-type natriuretic peptide. (114) Bone endothelial cells are also unique in their high surface expression of the chemoattractant stromal cell-derived factor-1 (SDF-1), their tissue-specific heparan sulfate proteoglycan patterns, and their adhesion molecule profiles, properties that functionally collaborate to preferentially cause hematopoietic progenitor cell homing to the bone marrow in an essential homeostatic mechanism that ensures a constant influx of such cells into bone. (25–27, 115–117) In pathological conditions, metastatic prostate and breast cells are selectively targeted to bone through specialized recruitment mechanisms where bone endothelial cells may very well have a role. (29, 30, 118–123) These data provide additional support for the concept that significant functional differences exist between endothelial cells from different tissues, and even from different microanatomical sites within a tissue. (35, 124–131) In addition to structural and functional diversity, endothelial cells also evidence antigenic and transcriptional heterogeneity. (5) An obvious question arises: are these differences the underlying cause or consequence of specific interactions between endothelium and each tissue microenvironment? Even without knowing the precise answer to this question, mounting evidence points to a direct role for bidirectional paracrine signaling between blood vessel cells and surrounding target organ cells during embryonic development and cell differentiation.


Understanding the nature of the signaling between endothelial and other cells in bone, including osteoclasts, osteoblasts, and their progenitors, is important not only to provide better insights into the mechanisms governing normal bone physiology, but also to suggest new avenues for the development of novel targeted therapeutic approaches for a variety of bone diseases. Through a combination of methodologies spanning in vivo research, isolated or reconstituted tissue model systems, and especially the use of well-characterized pure cellular models, advances will surely be forthcoming that will elucidate intercellular signaling pathways, mechanisms of cell-to-cell cross-talk, and establish key communication networks between vascular and other cells residing within bone.

An example of these latter types of studies is offered by studies into the “in vitro” interactions between bone endothelial cells and cells of the osteoclastic lineage. Bone-resorbing osteoclasts derive from hematopoietic precursors present in both the bone marrow and the peripheral circulation. (132) In both compartments, endothelial cells separate osteoclast precursors from the bone surface. To reach future sites of bone resorption, osteoclast precursors therefore need to adhere to and migrate through endothelium. This process is likely to be tightly regulated, similar to mechanisms governing the transendothelial migration of leukocytes and monocytes. (133) It has been suggested that, to correctly accomplish this, endothelial cells in cutting cones in cortical bone may express a specific “area code” that allows osteoclast precursors access to bone in a highly time- and space-dependent manner. (134) Evidence newly emerging from current research suggests that this may, in fact, be true. Moreover, consistent with their putative roles in regulating osteogenesis, osteoclastogenesis, and bone remodeling, endothelial cells synthesize and display or secrete many factors known to influence bone cells, including macrophage-colony stimulating factor (M-CSF), RANKL, various growth factors, cytokines, chemokines, prostanoids, free radicals, small peptides, adhesion molecules, and matrix constituents. (8–10, 13, 16, 36, 135–137) Conversely, bone endothelial cells also are capable of responding to bone modulators, including PTH, sex steroids, and proinflammatory cytokines. (8, 11) In vitro coculture and transmigration models have been used to mimic aspects of the osteoclast-vascular relationship during osteoclastogenesis and therefore to characterize structural and functional relationships between pre-osteoclasts and vascular endothelial cells. (15, 138–140) Results from these studies support the close linkage witnessed in vivo between angiogenesis and osteoclast formation and activity. (141–144) Future studies in this area may have important implications for the prevention and treatment of bone diseases associated with excessive bone resorption, in that they may identify previously unrecognized pathways that prove amenable to clinical interventions.

Another area of exciting investigation is that of the close relationship between bone formation and angiogenesis. Owing to the fact that microvascular endothelium is an integral and vital part of bone tissue, it is perhaps not surprising that endothelial cells may be directly involved in bone formation. Deficiencies in vascular supply or inappropriate bone vascularity have been linked to decreases in bone formation and bone mass, (6, 31) and coupling between vascular ingrowth and osteoblast differentiation is considered important in de novo bone formation, repair, and regeneration, as well as in the osseointegration of uncemented orthopedic and dental implants. (6, 7, 34, 49, 62–64, 73, 145–152) Numerous regulatory molecules that can be synthesized and secreted by vascular endothelial cells (growth factors, BMPs, cytokines, endothelins, free radicals, prostanoids) are known to exert major effects in controlling the differentiation, metabolism, survival, and function of bone-forming cells. For example, one interesting mechanism proposed to mediate interactions between bone endothelium and cells of the osteoblastic lineage is through the stimulatory effect of a product of endothelial cells, endothelin-1, on osteoprogenitor cell differentiation. (136) Furthermore, endothelial-derived angiotensin II has been shown to elicit anabolic effects in osteoblastic cells and to stimulate VEGF expression and angiogenesis. (153) Besides targeting endothelial cells, VEGF also acts directly on osteoblasts and osteoclasts, thereby potentially coordinating their actions. (154) Finally, the identification of mechano- and oxygen-sensitive BMP production (as well as other molecules) by endothelial cells(98–101) provides additional routes for interactive cross-talk between the bone endothelium and cells of the osteoblastic lineage. Thus, endothelial cells may respond to developmental or fracture site stimuli such as hypoxia or VEGF and, in turn, produce BMPs and other substances that can act reciprocally on osteoblasts as osteogenic signals. These positive interactions between endothelial and osteoblast lineage cells deserve further study, particularly in view of their potential relevancy for mesenchymal/stromal stem cells present in the bone marrow and circulation.


Abundant evidence indicates that the vasculature plays an active and important role in numerous skeletal pathologies, from vascular necrosis, (155–157) to pachydermoperiostosis, (158–161) osteopetrosis, (162–165) rickets, (62–64, 166–171) osteoporosis, (31–33) inflammatory bone loss, (138, 141, 142, 172–178) multiple myeloma, (179–182) Paget's disease of bone, (183) metastatic bone disease, (19, 29, 30, 184) melorheostosis, (185) and Gorham-Stout disease. (186) In such cases, pathological perturbations in the normal bone vasculature are associated (in a correlative or causative fashion) with profound changes in skeletal physiology. For example, angiogenesis is a prime factor underlying the pathology of various diseases of inflammatory bone loss, including rheumatoid arthritis and periodontal disease. (172–178) Unlike normal bone healing, in which angiogenesis (and inflammation) is appropriately stimulated and then halted, angiogenesis is persistently induced in chronic inflammatory conditions, leading to excessive recruitment of inflammatory cells, including osteoclast precursors, that engage in cartilage erosion and bone destruction. (138, 175) Because angiogenesis and inflammation behave as codependent reciprocal regulators of one another, inhibitors of either process produce beneficial effects to dampen the other as well. (177, 178)

The emerging likelihood that specific paracrine endothelial cues have important roles in coordinating bone morphogenesis and remodeling underscores the necessity for much further research into basic bone vascular biology. The reports briefly highlighted in this review provide examples of communication pathways and cross-talk between bone endothelial cells and adjacent cells or tissue, although these represent only the beginnings of our understanding of potentially far wider roles for bone endothelial cell signaling. At the moment, however, we understand little overall about the regulatory mechanisms that direct bone endothelium paracrine signaling, either in the developing vasculature or in adult vessels, and we are only in the nascent stages of identifying downstream events in responding cells. Characterizing the nature of this signaling by bone endothelial cells will undoubtedly unveil key signaling events heretofore unknown to be involved in bone modeling and remodeling and may offer new strategies for the development and delivery of targeted therapies. Broadening our understanding of bone endothelial cell biology will also provide a firmer foundation on which to rely for devising suitable approaches to build new vascular networks, in engineered tissues, for example. Our understanding of stem and progenitor cell recruitment, and their differentiation into bone-resorbing or bone-forming cell lineages, is likely to be significantly advanced through determining the endothelial signaling routes that govern their relocalization or cellular transmigration through the endothelial cell layer into bone tissue (Fig. 1). Finally, the discovery or generation of specific agonists (angiogenic) or antagonists (anti-angiogenic) of vessel growth and expansion in bone may prove to be beneficial for either inducing formative pathways or for targeting destructive conditions, respectively. Current preclinical and clinical trials are already taking advantage of the specialized properties of angiogenic vessels and their unique tissue-specific characteristics to locally incorporate vascular or mesenchymal progenitors and stem cells (with or without prior genetic modification), to perform selective gene targeting and gene therapy procedures, to deliver particular bioactive or toxic molecules into bone or other sites in a tissue-selective fashion, and to promote or prevent organ vascularization and function in a restricted manner. (187, 188) No doubt, our expanding knowledge of bone vascular biology eventually may be exploited to a similar beneficial end. With the rapid pace of research and development now taking place in the areas of vascular and bone biology, the future looks to be both exciting and productive.

Figure FIG. 1..

Vascular endothelial cells, OB/stromal cells, and OC and their precursors (pre-OC) all produce multiple regulatory substances (highlighted in boxes) that may act in autocrine/paracrine fashion to control the recruitment, proliferation, differentiation, function, and survival of these vascular and bone cell types. Thus, their close proximity to one another and responsiveness to such factors readily enables complex multidimensional communication pathways (dashed arrows) to coordinate the biological activities of such cells. Numerous conditions are capable of modulating such communication on a short-term or long-term basis (central Modulation box), thereby leading to shifts in the balance of regulators that mediate bone growth or development, bone repair or remodeling, and various skeletal pathologies. EC, endothelial cell; OB, osteoblast; BMSC, bone marrow stromal cell; OC, osteoclast; pre-OC, OC precursor; TEM, transendothelial migration; MSC, mesenchymal stem cell; HSC, hematopoietic stem cell; SDF-1, stromal cell-derived growth factor-1; IL-, interleukin; MCP-, monocyte chemoattractant protein-; CKβ-8, a chemokine; FKN, fractalkine; RANTES, regulated on activation of normal T cell expressed and secreted; MIP-1α, macrophage inhibitory protein-1α; Col, collagen; LM, laminin; FN, fibronectin; TSP, thrombospondin; GAGs, glycosaminoglycans; VEGF, vascular endothelial growth factor; PDGF, platelet-derived growth factor; FGFs, fibroblast growth factors; OPG, osteoprotegerin; BMPs, bone morphogenetic proteins; TGFβ, transforming growth factor β; M-CSF, macrophage-colony stimulating factor; G-CSF, granulocyte-colony stimulating factor; GM-CSF, granulocyte/macrophage-colony stimulating factor; Ang II, angiotensin II; PGE2, prostaglandin E2; ROS, reactive oxygen species; ET-1, endothelin-1; OCN, osteocalcin; OPN, osteopontin; ON, osteonectin; MGP, matrix Gla protein; BGP, bone Gla protein; BG, biglycan; IGFs, insulin-like growth factors; BLC, B-lymphocyte chemoattractant; TECK, thymus-expressed chemokine; ECF-L, eosinophil chemotactic factor-L.20


This work was supported by the Fondazione Ente Cassa di Risparmio di Firenze (to MLB) and NIH Grant DK46547 (to PCO).