|Home | About | Journals | Submit | Contact Us | Français|
Increasing evidence suggests the importance of bone marrow-derived cells for blood vessel formation (neovascularization) in tumors, which can occur in two mechanisms: angiogenesis and vasculogenesis. Angiogenesis results from proliferation and sprouting of existing blood vessels close to the tumor, while vasculogenesis is believed to arise from recruitment of circulating cells, largely derived from the bone marrow, and de novo clonal formation of blood vessels from these cells. Although bone marrow-derived cells are crucial for neovascularization, current evidence suggests a promotional role of these cells on the existing blood vessels rather than de novo neovascularization in tumors. This is believed to be due to the highly proangiogenic features of these cells. The bone marrow-derived cells are heterogeneous, consisting of many different cell types including endothelial progenitor cells, myeloid cells, lymphocytes, and mesenchymal cells. These cells are highly orchestrated under the influence of the specific tumor microenvironment, which varies depending on the tumor type, thereby tightly regulating neovascularization in the tumors. In this review, we high-light some of the recent findings on each of these cell types by outlining some of the essential proangiogenic cytokines that these cells secrete to promote tumor angiogenesis and vasculogenesis.
Endothelial progenitor cells (EPCs) are defined as precursor cells that have the ability to differentiate into mature endothelial cells when recruited to angiogenic sites. They are believed to arise from hemangioblasts, hematopoietic stem cells that are common precursors of EPCs . Asahara and colleagues initially identified EPCs by isolating and culturing mononuclear blood cells from human peripheral blood using magnetic beads against CD34, an endothelial cell surface antigen . In this study, EPCs were characterized by their spindle shape, incorporation of DiI-labeled acetylated low-density lipoprotein, expression of endothelial cell surface molecules such as CD31, Flk-1 (vascular endothelial growth factor receptor-2; VEGFR2), and Tie-2, and release of nitric oxide. The recent availability of transgenic mice where β-galactosidase (Lac-Z) or green fluorescent protein (GFP) are expressed either in endothelial cells only  or in all tissues  has led to improved understanding of the role of bone marrow-derived cells on blood vessel formation. Using these mouse models, numerous studies have reported that EPCs incorporate into the vasculature of solid tumor xenografts [5–7] and ischemic tissues [8–10].
However, the extent to which EPCs are incorporated into tumor vasculature has been controversial. While some reports have suggested that EPCs can comprise 50–100% of the tumor vasculature [6, 11], the majority of studies, including ours, showed only rare or undetectable levels of EPCs in the tumor vasculature [12–17]. Table 1 summarizes some of the findings from these studies. The reason for this discrepancy is not clear, although, the characterization of the EPCs phenotype, different tumor models, and stages of tumor growth have been suggested as possibilities [12, 18, 19]. A recent study by Nolan and colleagues, for example, reported that EPCs are incorporated into the tumor vasculature at 20–35% efficiency on days 4–6 following tumor implantation, but then becomes diluted away by local non-bone marrow-derived endothelial cells resulting in only 1% of EPCs to be detected in the tumor vasculature after 4 weeks of growth . Further studies are needed to address many questions on the fundamental characteristics of EPCs some of which include: whether EPCs lose proliferating abilities once recruited to the tumor vasculature despite their high self-renewal capabilities , whether the proliferating capabilities of EPCs are controlled by the tumor microenvironmental factors, whether local (non-bone marrow-derived) endothelial cells proliferate relatively faster than EPCs in tumors, and whether EPCs lose their identifying markers in daughter cells?
Despite the controversy regarding EPC incorporation into the tumor vasculature in animal models, a clinical study demonstrated the presence of EPC incorporation into the tumor vasculature of cancer patients, who had received gender-mismatched bone marrow transplantation . The levels of EPCs were generally low (1–12%) and again dependent on the type of cancers (spindle cell sarcoma, Hodgkin’s lymphoma, mucoepidermoid carcinoma, thyroid carcinoma, and osteogenic carcinoma) . Recent clinical studies have also reported that EPCs are detected more frequently in the peripheral blood and in the tumors of patients with more invasive disease  or recurrence , indicating a possible involvement in tumor progression. Consistent with this, recent preclinical studies have shown that EPCs promote metastasis  and determine sensitivity of tumors to chemotherapeutic  or vascular disrupting agents .
Genetic mouse models have demonstrated that impaired EPC mobilization inhibits growth of some tumor models [6, 26]. EPCs are proangiogenic secreting many angiogenic cytokines including VEGF [27, 28], stromal-derived factor-1 (SDF-1) , angiopoietin-1 , angiopoietin-2 , and erythropoietin . The proangiogenic activity of EPCs has been shown to be further enhanced by tumor cell- or stromal cell-secreted VEGF and placental growth factor (PlGF) in a paracrine manner . However, further studies are needed to determine whether these effects are due to bona fide EPCs or via other populations of bone marrow-derived cells such as myeloid cells.
Myeloid cells are VEGFR1 positive hematopoietic cells that eventually give rise to monocytes, macrophages, and granulocytes/neutrophils. The importance of myeloid cells in promoting tumor growth was first underscored by a study by De Palma and colleagues, who showed that Tie2-expressing monocytes are recruited to perivascular regions of the tumor vasculature and that genetically targeting Tie2-expressing monocytes significantly inhibited tumor growth in mice . Other investigators have reported similar findings that myeloid cells promote tumor angiogenesis  and progression [31, 32].
Myeloid cells have features that are similar to EPCs in that they may express (under some circumstances) endothelial specific markers such as CD34, CD31, and VEGFR2 [33, 34]. They are also functionally similar to EPCs or endothelial cells by integrating lectin and acetylated low-density lipoprotein . In addition, myeloid cells are often observed in closely adjacent to blood vessels  and have been shown to improve blood flow in ischemic models [36, 37]. Although some studies have suggested that myeloid cells may differentiate into endothelial cells [38, 39], this remains controversial and further studies are needed to define the precise relationship between myeloid and endothelial cells.
Many of the proangiogenic factors in myeloid cells that promote angiogenesis and vasculogenesis have been identified. For example, myeloid cells have shown to express matrix metalloprotease-9 (MMP-9) which induces the angiogenic switch by releasing VEGF from its membrane-bound form [40, 41]. MMP-9 expressing myeloid cells have also been shown to be crucial for angiogenesis in multistep carcinogenesis of cervical [42, 43] and pancreatic  mouse models. Another class of proteases, cathepsin cysteine protease has been reported to be expressed in myeloid cells regulating tumor growth, vascularity, and invasion . A recent study by Shojaei and colleagues demonstrated that Bv8, prokineticin 2, is expressed in myeloid cells and this induces angiogenesis in tumors . Further, this study showed that granulocyte colony-stimulating factor mobilizes myeloid cells and positively regulates the expression level of Bv8 in the myeloid cells . Other recent studies indicate that Stat3 and its downstream effectors (such as VEGF, basic fibroblast growth factor, interleukin-1β, and MMP-9)  or semaphorin 4D, a proangiogenic ligand that interacts with plexin B1 receptor on endothelial cells , is involved in myeloid cell-mediated angiogenesis in tumors.
Myeloid cells are recruited to tumors and other angiogenic sites via VEGF and SDF-1 gradients. By using a mouse model in which VEGF levels could be conditionally modulated, Grunewald and colleagues have shown that activation of VEGF results in retention of the myeloid cells and other bone marrow-derived cells and that this is via VEGF-induced SDF-1 release from perivascular regions . Similar findings have been reported by other investigators  including the fact that VEGF and SDF-1 can be controlled by their upstream regulator, hypoxia-inducible factor-1 (HIF-1) . Recently, Du and colleagues have demonstrated that HIF-1 levels in tumors modulate the recruitment of these bone marrow-derived cells into glioblastomas, thereby regulating angiogenesis in these tumors .
Mast cells and platelets are also subpopulations of myeloid cells that have been shown to play an important role in tumor angiogenesis. Coussens and colleagues have shown that mast cells promote tumor progression in a mouse model of skin carcinogenesis by first infiltrating into the tumor, degranulating and activating dermal fibroblasts, and then by activating pro-MMP-9 (inactive form) to MMP-9 (active form) thereby promoting angiogenesis . In this study, the authors showed that mast cells release chymase and tryptase, the mast cell-specific serine proteases . Mast cells have also been shown to be a rich source of VEGF, basic fibroblast growth factor, and MMP-9 [52, 53], and their functions are crucial for maintaining endothelial cell proliferation . Platelets have been shown to be recruited to the activated endothelium providing VEGF [54, 55] and SDF-1 . Platelets have also been shown to release platelet-derived microvesicles to tumors, thereby stimulating expression of proangiogenic factors such as MMP-9, VEGF, and interleukin-8 in the tumor cells to promote angiogenesis .
We and others have shown that myeloid cells in the tumors are therapeutically important. We have recently shown that myeloid cells restore tumor re-growth after irradiation by supporting immature blood vessel development . Shojaei and colleagues have demonstrated that these cells are responsible for resistance to anti-angiogenic therapy . Other investigators have shown that these cells facilitate tumor metastasis [31, 59, 60], and there is strong clinical evidence that the abundance of macrophage infiltration correlates with poor prognosis in breast, prostate, ovarian, and cervical cancers .
Thus, the myeloid cell appears to be an excellent target to inhibit tumor growth, angiogenesis, and metastasis. However, given the efficiency with which myeloid cells are rapidly replenished from stem/progenitor cells in the blood and bone marrow, it is difficult to target them efficiently from the blood stream. Despite this, we have obtained encouraging early results in potentiating the efficacy of cancer therapy by inhibiting myeloid cells by using antibodies against CD11b, a cell surface molecule expressed on monocytes and macrophages.
CD4 or CD8 lymphocytes are capable of rejecting tumors by immune surveillance . However, some studies suggest that these lymphocytes are in certain cases associated with tumor promoting abilities. For example, Daniel and colleagues showed that infiltration of CD4 helper T cells enhance tumor progression in a skin cancer mouse model possibly by promoting the infiltration of MMP-9 producing myeloid cells . In a similar multistep carcinogenesis mouse model, CD4 cells have also shown to promote angiogenesis and the progression of pancreatic tumors but only in the absence of tumor necrosis factor receptor-1 or interferon-γ (IFN-γ) signaling . Although the evidence for a direct involvement of CD8 cytotoxic T cells in angiogenesis/vasculogenesis is rather weak, a recent study indicated that anti-angiogenic therapies disrupting the tumor vasculature can potentiate the immune response by increasing CD4 and CD8 lymphocytes and releasing IFN-γ in the tumor . It has been suggested that the tumor microenvironment plays a crucial role in suppressing T cell activities which could in turn promote stability of the tumor vasculature .
Mesenchymal cells derived from the bone marrow are adherent stromal cells characterized by their ability to differentiate into mesenchymal tissues such as bone, cartilage, and fat . Increasing evidence indicates that bone marrow-derived mesenchymal cells support tumor angiogenesis by providing a supportive role as carcinoma-associated fibroblasts [67, 68] or perivascular mural cells . These cells express α-smooth muscle actin , Tie-2 , and other pericyte markers . Mesenchymal cells secrete SDF-1  and VEGF  and are recruited to the sites expressing VEGF, platelet-derived growth factor (PDGF), and epidermal growth factor (EGF) . By exploiting this feature of being recruited to tumors, Studeny and colleagues showed that intravenous injection of human mesenchymal stem cells expressing interferon-β to MDA231 metastases bearing mice significantly prolonged the overall survival of the animals .
In this review, we have highlighted some of the recent findings on the role of bone marrow-derived cells and their proangiogenic cytokines for new blood vessel formation in tumors. Convincing evidence indicates a critical role played by bone marrow-derived cells in promoting tumor angiogenesis and progression. This suggests that solid tumors have the capability of modulating their tumor environment such that infiltrating immune cells favor, rather than inhibit tumor growth. Molecular mechanisms by which the bone marrow-derived cells promote tumor angiogenesis are not well-understood. This is largely because of the heterogeneity of the bone marrow-derived cells and the complex nature of these cells including the diverse interactions occurring among them. But, this is an exciting time for this field because of the development of powerful genetic mouse models as wells as the availability of specifically targeted therapies against each population of the bone marrow-derived cells. This will enable future studies to dissect out the critical players promoting tumor angiogenesis as well as develop new anticancer strategies.
This work was supported by National Institutes of Health grants CA118202 and CA128873 awarded to JMB. GOA is a recipient of Gary Slezak/American Brain Tumor Association Translational Grant.