Recently we reported that macrophage colony-stimulating factor (M-CSF) can mobilize endothelial progenitor cells (EPCs) from the bone marrow into the peripheral blood, resulting in an increase in the number of blood vessels and augmentation of blood flow in the ischemia-induced legs. M-CSF accelerates neovascularization of ischemic lesions resulting from the mobilization of EPCs. In the present paper, we analyze the mechanisms underling the mobilization of EPCs by M-CSF. M-CSF augments the production of vascular endothelial growth factor (VEGF) from the bone marrow cells, especially from myeloid lineage cells. In vivo administration of anti-VEGF antibody abrogates both the acceleration of the recovery of blood flow in the ischemia-induced limbs by M-CSF and the augmentation of the mobilization of EPCs induced by M-CSF. These results suggest that the M-CSF contributes to rapid recovery of blood flow in ischemic lesions by mobilization of EPCs from the bone marrow through augmentation of VEGF production in the bone marrow and that the VEGF is mainly produced by myeloid lineage cells.
Asahara et al. have reported that endothelial progenitor cells (EPCs) exist in the bone marrow (BM) and the peripheral blood (PB) and that the EPCs can differentiate into the endothelial cells of the blood vessels [1, 2]. Based on these findings, Tateishi-Yuyama et al. clinically used bone-marrow mononuclear cells to augment the blood flow in ischemic limbs of the patients with arteriosclerosis obliterans or Bürger's disease, and they have shown that the injection of own bone marrow cells (BMCs) into ischemic muscles of the patients is effective not only in increasing the blood flow in ischemic limbs but also in decreasing the ulceration and the leg pain . It has also been reported that the injection of BMCs is effective in ischemic heart disease, resulting from the augmentation of blood flow . Thus, usage of EPCs in the BM is a novel and desirable therapy for ischemic diseases.
Three kinds of colony stimulating factors (CSF) are known at present: granulocyte colony-stimulating factor (G-CSF), macrophage colony-stimulating factor (M-CSF), and granulocyte-macrophage colony-stimulating factor (GM-CSF). It has been already reported that both G-CSF and GM-CSF can activate myeloid lineage cells and can mobilize hematopoietic precursor cells [5–8]. At present, G-CSF has been tried clinically for ischemic heart diseases, but the effects of G-CSF on ischemic lesions are controversial [9–12].
Recently, we have shown that M-CSF, as well as G-CSF, mobilizes EPCs from the BM into the PB, resulting in the acceleration of neovascularization in the ischemia-induced legs . It has also been reported that the downregulation of CXCR4 and SDF-1 is related with the mechanisms underlying the mobilization of precursor cells from the BM to the PB by G-CSF . However, the mechanisms underlying mobilization of EPCs by M-CSF are unclear. In this paper, we focus on clarifying the mechanisms underlying the mobilization of EPCs by M-CSF. We demonstrate that M-CSF can mobilize EPCs through vascular endothelial growth factor (VEGF) mainly produced by myeloid lineage cells in the BM, resulting in the augmentation of blood flow and the increased number of blood vessels in ischemia-induced legs.
Anti-CD3, anti-CD19, anti-CD11b, anti-CD11c, anti-Ter119, anti-CD45, anti-Flk-1, and anti-NK1.1 antibodies (Abs) were purchased from BD Biosciences (San Diego, http://www.bdbiosciences.com). M-CSF was kindly donated by Morinaga (Zama, Japan, http://www.morinagamilk.co.jp/menu/english.html). G-CSF was kindly donated by Chugai Pharmaceutical Co. (Tokyo). Anti-CD115 (receptor for M-CSF) Ab was purchased from eBioscience (San Diego). Anti-VEGF Ab was purchased from Lab Vision (Fremont, CA, http://www.labvision.com).
Detection of CD115+ Cells in BM
To examine the expression of CD115 in various lineages of BMCs, BMCs were stained with anti-CD115 Ab + lineage-specific Abs. To detect the expression of CD115 on erythroid lineage cells, myeloid lineage cells, and B lymphocytes, BMCs were incubated with biotin-labeled anti-Ter119 + PE-labeled anti-CD115 Ab followed by staining with fluorescein isothio-cyanate (FITC)-labeled avidin, FITC-labeled anti-Gr-1 Ab + PE-labeled anti-CD115 Ab, FITC-labeled anti-CD11b Ab + PE-labeled anti-CD115 Ab, and FITC-labeled anti-CD19 Ab + PE-labeled anti-CD115 Ab. The stained cells were analyzed with a FACScan (Becton, Dickinson and Company, San Diego, http://www.bd.com). Each lineage of cells was gated, and the expressions of CD115 in the cells were analyzed. To analyze the expression of CD115 in lineage-negative (Lin−) cells, BMCs were stained with biotin-labeled anti-Ter119 Ab + biotin-labeled anti-Gr-1 Ab + biotin-labeled anti-Mac-1 Ab + biotin-labeled anti-CD19 Ab + biotin-labeled anti-CD3 Ab + biotin-labeled anti-NK1.1 Ab + biotin-labeled anti-CD11c Ab + PE-labeled anti-CD115 Ab followed by staining with PE Cy5-labeled avidin. Lin− cells were gated, and the expression of CD115 in the cells was analyzed. To examine the expression of CD115 of EPCs, BMCs were stained with FITC-labeled anti-CD45 Ab, PE-labeled anti-Flk-1 Ab, and biotin-labeled anti-CD115 Ab followed by staining with PE Cy5-labeled avidin. CD45− /Flk-1+ cells were gated, and the expression of CD115 of the cells was analyzed.
Preparation of Hind Limb Ischemia
On day 0, unilateral hind limb ischemia was induced by ligation of the right femoral arteries and veins of B6 mice, as previously described .
Administration of G-CSF, M-CSF, and anti-VEGF Ab
Human recombinant M-CSF (250 μg/kg) or human recombinant G-CSF (250 μg/kg) was administered to the mice intraperitoneally for 3 days (day 0, day 1, and day 2). The dosage of M-CSF and G-CSF was determined according to our previous experiments [13, 16]. Anti-VEGF Ab (250 μg/mouse) was intraperitoneally administered on day 0, simultaneously with injection of M-CSF or G-CSF.
Laser Doppler Perfusion Image
The hind limbs of mice were shaved using a razor. The mice were anesthetized with 160 mg/kg pentobarbital and fixed supine on a cork plate. We next measured the blood flow of limbs using a laser Doppler perfusion image (LDPI) analyzer (Moor Instruments, Millway Devon, U.K., http://www.moor.co.uk), as described previously . LDPI indexes were shown as the ratio of the blood flow of the ischemic (right) to normal (left) limb.
BMCs and spleen cells were adjusted to 1 × 106/ml in RPMI 1,640 containing 10% of fetal calf serum with or without M-CSF (10 ng/ml).
For purification of Terr119+ cells, Gr-1+ cells, Mac-1+ cells, or CD19+ cells, BMCs were incubated with PE-labeled anti-Terr119, Gr-1+ cells, Mac-1+ cells, or CD19+ Abs, and then positive cells for each Ab were sorted using an EPICS ALTRA cell sorter (Beckman Coulter, Fullerton, CA, http://www.beckmancoulter.com). For preparation of Lin− cells, BMCs were cultured with PE-labeled anti-Terr119, anti-Gr-1, anti-Mac-1, anti-CD19, anti-CD11c, and anti-NK1.1 Abs, and cells negative to these Abs were collected using the EPICS ALTRA. These cells were cultured with or without M-CSF (10 ng/ml) for 3 days.
Enzyme-Linked Immunosorbent Assay (ELISA) for VEGF
Supernatants of cultured cells were collected, and ELISAs for VEGF were performed using the Quantikine Mouse VEGF immunoassay kit (R&D Systems Inc., Minneapolis, http://www.rndsystems.com) following the manufacturer's instructions.
Estimation of Number of EPCs in Peripheral Blood
B6 mice were injected with M-CSF (250 μg/kg) for 3 days. Some of the mice were intraperitoneally injected with anti-VEGF Ab (250 μg/mouse) simultaneously with the injection of cytokines on day 0. Four days after the first injection of M-CSF (day 3), PBs of the mice were collected, and numbers of nuclear cells were analyzed using an SF-3000 autoanalyzer for the peripheral blood (Sysmex, Kobe, Japan). The cells were then stained with PE-labeled anti-Flk-1 Ab, biotin-labeled anti-CD45 Ab, and biotin-labeled anti-Terr119 Ab, followed by staining with PE Cy5-labeled avidin. The cells were analyzed with a FACScan. The nuclear cells were gated using development of forward scatter and side scatter, and the percentages of EPCs that show Flk-1+/CD45−/Terr119− were analyzed, followed by calculation of absolute number of EPCs in PB from the numbers of nuclear cells and percentages of EPCs .
Statistical analyses were performed with Student's t test. Values of p < .05 were considered to be statistically significant.
Anti-VEGF Ab Abrogates Acceleration of Blood Flow by M-CSF but Not Acceleration by G-CSF
We have reported that M-CSF and G-CSF can accelerate neovascularization of ischemia-induced legs and that both cytokines can induce the production of VEGF in the bone marrow cells but not spleen cells . Therefore, we examined the role of VEGF in the acceleration of the recovery of blood flow of the ischemia-induced legs by the cytokines using anti-VGEF Ab. Representative images of LPDI and means and standard deviations of the data are shown in Figure 1A and 1B, respectively. As we previously described, the administration of M-CSF and G-CSF augments blood flow in ischemia-induced legs (Fig. 1). When anti-VEGF Ab was injected into the mice, the recovery of blood flow induced by M-CSF was abrogated, whereas anti-VEGF Ab had no effect on the recovery of the blood flow induced by G-CSF on day 3. On day 7, the blood flow of the ischemia-induced legs in the mice without either cytokine or Ab rose to the same level as the blood flow in the cytokine-injected mice. However, the ischemia-induced legs of the mice with both M-CSF and anti-VEGF Ab still showed low blood flow in the ischemic legs. These results suggest that M-CSF accelerates the recovery of blood flow through an increase in VEGF production and that G-CSF does not use VEGF. These results are compatible with previous reports indicating that G-CSF downregulates the expressions of SDF-1 and CRCX4 in the BM, resulting in augmentation of the mobilization of precursor cells from the BM to the PB . These results also suggest that M-CSF effects acceleration of blood flow through VEGF.
Expression of Receptor for M-CSF in BMCs
Next, we analyzed the expression of receptor for M-CSF (CD115) in various lineage cells in the BM. As shown in Figure 2, mainly Gr-1+ cells and Mac-1+ cells express CD115, whereas only a very low number of EPCs (less than 5%) express CD115, suggesting that M-CSF does not work directly on the EPCs but, rather, works on the EPCs via another factor. These results are compatible with the abrogation of the effects of M-CSF by the administration of anti-VEGF Ab.
Myeloid-Lineage Cells in the BM Produce VEGF by Stimulation of M-CSF
We measured serum VEGF levels using ELISA, but we detected VEGF neither in the serum of nontreated mice nor in the serum of M-CSF-administered mice (data not shown; ). Next, we examined the precise production of VEGF in BMCs induced by M-CSF. We previously performed a similar experiment and showed that M-CSF augments the production of VEGF by BMCs, but in that experiment, we measured only the optical density of the culture supernatants . In the present study, we measured the concentrations of VEGF on an absolute scale and with the time course. As shown in Figure 3A, BMCs produced VEGF, and the production of VEGF was enhanced by M-CSF, whereas spleen cells produced only a small quantity of VEGF even when cultured with M-CSF.
Next, we examined the production VEGF in various populations of BMCs in the presence of M-CSF. As shown in Figure 3B, Mac-1+ cells, Gr-1+ cells, and Lin− cells, especially Mac-1+ cells and Gr-1+ cells, produced VEGF without M-CSF, and they produced much more VEGF in the presence of M-CSF. These results suggest that myeloid lineage cells mainly produce VEGF naturally but that erythroid cells or B cells cannot produce VEGF and that M-CSF augments the production of VEGF by myeloid lineage cells and Lin− , especially myeloid cells, resulting in the mobilization of endothelial progenitor cells.
Anti-VEGF Ab Abrogates Mobilization of EPCs Induced by M-CSF
We examined the effect of anti-VEGF Ab on the EPC mobilization induced by M-CSF. As shown in Figure 4, M-CSF augmented the number of EPCs in the peripheral blood, as we previously reported. However, when anti-VEGF Ab was injected into the mice, the number of EPCs in the PB decreased to the level without M-CSF. These results suggest that M-CSF stimulates the production of VEGF in the BM, resulting in the mobilization of EPCs, and that mobilized EPCs differentiate into endothelial cells in the ischemic lesions to augment the blood flow.
The belief has been that angiogenesis contributes to postnatal neovascularization, whereas both angiogenesis and vasculogenesis contribute to fetal neovascularization [18, 19]. However, in 1997, Asahara et al. reported that EPCs exist even in adult PB and contribute to neovascularization, suggesting that vasculogenesis also contributes to postfetal neovascularization, as well as angiogenesis . Their paper has had an impact not only on basic science but also on clinical medicine. They also demonstrated that the BMCs contain EPCs and that EPCs in the PB are derived from the BM . Based on these findings, BMCs have been used clinically for purposes of recovery of blood flow in ischemic limbs due to arteriosclerosis obliterans or Bürger's disease . In their experiments, Tateishi-Yuyama et al. injected BMCs into ischemic areas of the muscles and showed an increase in the number of blood vessels in the area and recovery of blood flow . However, although the mobilization of EPCs from the BM using G-CSF has been attempted recently, there is concern that the administration of G-CSF accelerates arteriosclerosis and may give rise to thrombosis , since G-CSF is well known as a cytokine that induces inflammation . It has even been reported that G-CSF could accelerate ischemic heart disease . However, other clinical trials of G-CSF for augmentation of neovascularization have shown no relationship between G-CSF and ischemic diseases, and authors have stated the benefits of the use of G-CSF for neovascularization without prominent side effects [10–12]. It has also been reported that not only G-CSF but GM-CSF can also mobilize EPCs , and we have recently shown that M-CSF mobilizes EPCs from the BM into the PB, resulting in an acceleration of angiogenesis of ischemic limbs (Figs. 1, 4) . In our previous work, we have also shown that M-CSF does not augment the number of white blood cells in the PB but augments the number of EPCs in the PB . In the present work, we investigated the mechanisms underlying acceleration of angiogenesis by M-CSF, and we have shown that M-CSF can mobilize EPCs from the BM into the PB through VEGF.
VEGF is one of the most effective cytokines for angiogenesis, not only under normal conditions, but also in ischemic conditions and for development of tumors, and many tissues or cells (tumor cells, skin, brain, pituitary gland, kidney, ovary, uterus, heart, lung, liver, skeletal muscle and hematopoietic cells) can produce VEGF (Fig. 3) [23–29]. VEGF can mobilize EPCs from the BM into the PB as well as GM-CSF, G-CSF, M-CSF, and esotrogen do [13, 22, 30, 31]. However, systemic administration of VEGF involves side effects such as the induction of edema . The administration of M-CSF augments the production of VEGF from the BM but does not augment either serum levels of VEGF or production of VEGF from the spleen cells (Fig. 3). These results suggest that the effect of systemic administration of M-CSF is similar to local administration of VEGF into the BM as far as mobilization of EPCs. Therefore, the administration of M-CSF is effective in mobilizing EPCs without major side effects such as the systemic edema induced by VEGF. M-CSF administration could be a new strategy for neovascularization for ischemic organs.
In the present paper, we have shown the mechanisms underlying neovascularization by M-CSF. M-CSF induces production of VEGF from BMCs, especially myeloid lineage cells, and the VEGF mobilized EPCs from the BM.
The authors indicate no potential conflicts of interest.
We thank S. Miura, M. Murakami-Shinkawa, Y. Tokuyama, K. Hayashi, and A. Kitajima for expert technical assistance and Hilary Eastwick-Field and K. Ando for the preparation of the manuscript. This study was supported by grants from the “Millennium,” “Science Frontier,” and “21st Century” (project leader) programs of the Ministry of Education, Culture, Sports, Science and Technology; grants-in-aid for scientific research on priority areas from the Ministry of Education, Culture, Sports, Science and Technology (11470062, 10181225, 11162221); and Health and Labour Sciences research grants (Research on Human Genome, Tissue Engineering Food Biotechnology). It was also supported by grants from the Department of Transplantation for Regeneration Therapy (sponsored by Otsuka Pharmaceutical Company, Ltd.), Molecular Medical Science Institute, Otsuka Pharmaceutical Co., Ltd., and Japan Immunoresearch Laboratories Co., Ltd.