INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

ARTERIOGENESIS (collateral artery growth) is potentially able to restore blood flow after coronary, cerebral, renal, or peripheral arterial occlusions under both clinical or experimental conditions (1, 7, 9, 10, 15, 25, 27). There is increasing evidence that collateral arteries, in contrast to capillary growth during angiogenesis (24), originate from preexisting arterial anastomoses that are composed of the endothelium and the surrounding smooth muscle cell layers. It has been suggested that these processes of adaptive growth and tissue remodeling are initially induced by a shear force-induced expression of adhesion molecules (i.e., selectins, ICAM-1, and VCAM-1) by the endothelial cells (31). According to this model, circulating mononuclear blood cells adhere, migrate into deeper parts of the vessel wall, and/or populate the adventitial space and subsequently stimulate collateral growth by the release of cytokines, growth factors, and proteases (1, 26). Several previous studies (1, 26, 31) based on electron microscopical and immunohistochemical experiments showed that monocytes accumulate within the vessel wall of growing collaterals, suggesting that these cells contribute to the processes of arteriogenesis. Furthermore, a study in which collateral artery growth was stimulated by a local infusion of monocyte chemotactic protein-1 (MCP-1) (15) showed that under the influence of chemoattractants monocytes are effectively recruited from the circulation to sites of the activated collateral endothelium. Arras et al. (1) hypothesized that the speed of arteriogenesis depends on the mitogenic activity released by monocyte-derived macrophages such as basic fibroblast growth factor. However, the evidence for an essential role of monocytes in collateral artery growth is based on correlative and descriptive data. Therefore, the aim of this study was to functionally link collateral artery growth with a controlled manipulation of the number of monocytes and their progenitors.

With the use of two animal models of acute femoral artery occlusion (1, 5, 6, 15), the improvement of collateral artery growth was investigated either during the 5-fluorouracil (5-FU)-induced monocytopenia or the subsequent reactive monocytosis (monocyte rebound). We present for the first time functional data indicating that the amount of circulating blood monocytes is critical for the speed of collateral artery growth. Furthermore, we identified the CD14⁺ and CD11b⁺ (Mac-1⁺) cells as being responsible for this effect.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Reagents, antibodies, and animals. Ficoll-Paque Plus was purchased from Pharmacia Biotech (Freiburg, Germany). The cytostatic agent 5-FU (cat. no. 1000) was obtained from Lederle (Wolfertshausen, Germany), adenosine was from Fluka (Steinheim, Germany), and gelatin and bismuth were purchased from Merck (Frankfurt, Germany). Phycoerythrin (PE)-conjugated monoclonal antibodies (MAb) against mouse CD11b (clone M1/70) and human CD14 (MY4), FITC-conjugated MAb against mouse CD45 (clone IOT-45-1-2), granulocytes (RB6-8C5), and flow-count fluorespheres were purchased from Beckman-Coulter (Krefeld, Germany). Mouse MAb against rabbit macrophages (RAM11) and FITC-conjugated MAb against mouse IgG were from DAKO (Hamburg, Germany). Tetramethylrhodamine isothiocyanate-conjugated phalloidin and 4',6-diamidino-2-phenylindole were from Sigma-Aldrich (Munich, Germany). The rabbits were obtained from Elevage Scientifique De Dombes (Chatillon sur Charlonne, France), and mice were purchased from Harlan-Winkelmann (Borchen, Germany). All cell culture media and supplies were bought from Invitrogen (Karlsruhe, Germany). Recombinant human MCP-1 was bought from Peprotech (London, UK). Osmotic minipumps were obtained from Alzet (Cupertino, CA).

Animal models. The present study was performed with the permission of the State of Hessen, Regierungspraesidium Darmstadt, according to Section 8 of the German Law for the Protection of Animals. It also conforms to the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85-23, Revised 1996). No animals were lost during or after femoral artery ligation. We also did not observe any gangrene or gross impairment of hindlimb function after femoral artery occlusion. All studies were blinded to the investigator.

Rabbit model. New Zealand White rabbits were randomly assigned to receive either one of the following treatments: The antimetabolite 5-FU was diluted to 4.6 mg/ml in isotonic saline and applied to anesthetized animals (20 mg/kg body wt; day 0; n = 6) within 30 min via an ear vein. The ligation of the femoral artery, the delivery of MCP-1 (positive controls) and postmortem angiography were performed as described by earlier studies (13, 14). In brief, animals were anesthetized again 3 days after 5-FU application. The femoral artery of the right hindlimb was exposed and ligated. Control animals received MCP-1 (0.2 µg · kg¹ · day¹; n = 5) via an osmotic minipump, which was connected to the proximal stump of the ligated femoral artery by a catheter. The animals were also treated with tetracyclin (Tridox, 30 mg/kg sc). For negative control, rabbits (n = 7) did not receive further treatment except from femoral ligation. For postmortem angiography, the descending aorta was dilated with adenosine (1 mg/kg) and infused with contrast medium based on bismuth and gelatin. Angiograms of each hindlimb were taken in a Balteau radiography apparatus (Maclett Laboratories). Collaterals as defined by the Longland classification (19) were counted.

Quantification of collateral conductance in rabbit hindlimb. Seven days after ligation of the femoral artery, the animals were again anesthetized and maintained, as described by Wright et al. (40). Systemic pressure was measured from the right carotic artery, and peripheral pressures were measured from both saphenal arteries each connected to P32 DC pressure transducers (Statham, Spectramed) by catheters (13). Hindlimb blood flow was measured with perivascular ultrasonic flowprobes (Transonic Flowprobe, 2.5 mm) positioned around both external iliac arteries at increasing adenosine concentrations (30 to 600 µg · kg¹ · min¹, infused through an aortic catheter positioned 3 cm proximal to the iliac bifurcation). All data were recorded on a computer with the use of commercially available software (MacLab, Macintosh). The maximal collateral conductance was calculated from the systemic and peripheral pressures and the maximal iliac blood flow.

Tissue sampling and histology. An additional six rabbits were operated as described. Three days after ligation of the femoral artery, the animals were euthanized and tissue from the vastus intermedius muscle was harvested for histological examination after perfusion fixation (4% paraformaldehyde for 15 min). For immunohistochemistry, frozen sections (10 µm thick) were placed on gelatin-coated slides, fixed for 10 min in 4% paraformaldehyde, and blocked in 0.1% bovine serum albumin. RAM11 was incubated over night at 4°C, followed by incubation of a FITC-conjugated MAb anti mouse IgG for 1 h at room temperature. After being washed with PBS, the slides were incubated for 10 min with a mixture of rhodamine-conjugated phalloidin (to stain F actin) and 4',6-diamidino-2-phenylindole dimeric thiazole blue (TOTO-3, nuclear staining, Molecular Probes). Prepared slides were examined on a Leica TCS NT confocal scanning laser microscope equipped with argon-krypton and helium-neon lasers, as previously described (18).

Mouse model. 129S2/SvHsd mice randomly received a single injection of 5-FU (150 mg/kg ip) or buffer only. Four or twelve days after 5-FU treatment, the mice were anesthetized, and the right femoral artery was exposed and ligated distally of the arteria profunda femoris. Solvent-treated mice served as controls. Seven mice were included in every group.

In vivo perfusion measurements. The relative blood flow in the mouse foot was quantified via erythrocyte motion detection with the use of a laser Doppler imager (model MLDI 5063, Moor Instruments; Devon, UK). Mice were anesthetized and maintained at 37°C in a custom-designed air-conditioned chamber. The measurements were performed before, immediately after the surgery, and on postoperative days 3, 7, 14, and 21 at a pixel resolution of 256 × 256 and a scan rate of 4 ms/pixel for a 1.7 × 3 cm area, including both feet. Dead mice were used to determine the background level for subsequent subtraction. The right-to-left ratio was calculated for each mouse.

Oxygen saturation of hemoglobin. Oxygen saturation of hemoglobin was determined by the hemoglobin absorption spectrum with an AbTisSpec spectrometer at a wavelength range from 500 to 620 nm by using a circular probe placed on the mouse foot. Measurements, performed at 37°C in an air-conditioned chamber, were done before and immediately after the right femoral artery occlusion as well as at days 3, 7, 14, and 21. Right to left ratios were calculated for each mouse and statistically evaluated.

Postmortem angiography and tissue sampling. At the end of the study, the mice were euthanized, the thoracic aorta was cannulated, and the circulatory system was rinsed under 100 mmHg pressure with Tris-buffered saline containing 0.1% adenosine. After achieving maximal vasodilatation, the contrast agent (bismuth chloride in 5% gelatin) was infused. Hindlimbs were stored on ice until roentgenography was performed with a Balteau radiography apparatus (20 kV, 8 mA; Maclett Laboratories). In addition, superficial collateral arteries visible under the stereomicroscope were photographed.

Murine monocyte isolation and transplantation. Blood was collected from heparinized SV129 mice. Red blood cells were removed by density gradient centrifugation by using Ficoll-Paque according to the manufacturer's instructions (Pharmacia). The residual cell layer was harvested and washed three times with Hanks' buffered saline. Cells were stained with a monoclonal antibody against mouse CD11b (PE conjugated). Monocytes were detected and sorted in a high-speed cell sorter (EPICS Altra, Coulter; Miami, FL) by their CD11b expression (fluorescence) and their scatter light characteristics. As recipients for the monocytes, mice were used whose femoral artery had been occluded four days after 5-FU treatment. Freshly isolated monocytes were injected into the tail vein immediately after ligation of the femoral artery.

Quantitative flow cytometry. Flow cytometry was performed as previously described (12). To quantify cell concentrations in blood, fluorescent standard beads were used. Whole blood (30 µl) was incubated with a PE-conjugated monoclonal antibody against the monocyte marker CD14 (clone MY4, in the case of rabbit blood) or against mouse CD11b (clone M1.70, in the case of mouse blood). Monocyte populations were identified by fluorescence and scatter light characteristics. Absolute monocyte counts were calculated from the numbers of analyzed monocytes and fluorescent standard beads.

Statistics. Results are presented as means ± SE. Mean values were compared by ANOVA among different groups, with the Tukey's post hoc test used for pairwise multiple comparisons. P < 0.05 was considered to be statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Hematological changes in response to 5-FU treatment. In rabbits, a pronounced increase of the monocyte concentration in peripheral blood was observed 3-5 days after 5-FU (675.75 ± 104.7 CD14⁺ cells/µl; P < 0.05) (Fig. 1A). The maximal increase in monocyte counts compared with untreated rabbits was 9.6-fold (4.4- to 17.1-fold increase) (Fig. 1B). In mice, the application of 5-FU reduced the monocytes in the peripheral blood more efficiently (Fig. 1C): after 4 days, the peripheral blood monocyte concentration was decreased by 86% from 207 ± 10.73 to 28 ± 9.8 cells/µl (P < 0.01). The monocyte rebound effect between day 10 and day 12 led to a 5- to 11-fold increase of circulating blood monocytes to a maximal mean of 2,405 ± 185.6 cells/µl. Whereas the number of peripheral blood granulocytes was also increased during the rebound reaction, the lymphocyte concentration did not show a significant change (data not shown). In conclusion, rabbits displayed a marked increase in CD14⁺ blood monocytes, whereas in mice both depletion and a subsequent increase of CD11b⁺ blood monocytes were observed.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Effects of 5-fluorouracil (5-FU) on blood monocyte concentration assessed by quantitative flow cytometry. A: kinetic of monocyte counts in peripheral venous blood after infusion of 20 mg/kg 5-FU in New Zealand White rabbits. Arrow indicates the day of ligation of the right femoral artery. B: maximal achieved increase of blood monocyte counts in rabbits during the monocyte rebound after 5-FU treatment. ** P < 0.001. C: monocyte concentration kinetic in 5-FU-treated (150 mg/kg) SV129 mice. Arrows indicate the day of ligation of the right femoral artery in two different mouse groups [depletion (dep) or rebound group (reb)].

Effects of 5-FU treatment on collateral formation. The growth of visible collateral arteries in both animal models was macroscopically studied using angiograms from control and rebound groups. In both animal models, the number and the size of visible collaterals compared with controls were increased when the femoral artery was ligated during the monocyte rebound (Fig. 2). A distinct collateral network could be observed within the proximal hindlimb. In the rabbit rebound group, the number of visible collaterals was significantly increased (9.0 ± 0.7 vs. 13.3 ± 1.3 collaterals, P < 0.05). The highest number of visible collaterals (20.3 ± 1.33) was observed in MCP-1-treated positive controls (Fig. 2, A-D). Angiograms in mice were performed 21 days after ligation of the femoral artery. After this time, several collaterals could also be observed in untreated mice. However, larger and more visible collateral vessels were found in the rebound group (Fig. 2, D and E). Angiograms from untreated and monocyte-injected mice delivered comparable results (data not shown).

View larger version (71K): [in this window] [in a new window]   Fig. 2.   Collateral artery growth after acute ligation of the femoral artery in rabbit (7 days after ligation) and mouse (21 days after ligation) hindlimbs. A: angiogram of a rabbit hindlimb without further treatment (control). A few visible collaterals spanning the occlusion side can be detected. B: angiogram of a rabbit of the monocyte rebound group. Ligation of the femoral artery was performed during the reactive monocytosis. A markedly increased number of visible corkscrew-like collaterals can be observed. C: rabbit angiogram after 7 days of local monocyte chemotactic protein-1 (MCP-1) administration and femoral artery occlusion. All angiograms were performed 7 days after femoral artery occlusion. D: quantification of visible collaterals in proximal rabbit hindlimbs; * P < 0.001. ** P < 0.0001. E: angiogram of mouse hindlimbs 21 days after right femoral artery ligation without further treatment (control). F: angiogram of mouse hindlimbs 21 days after right femoral occlusion during rebound monocytosis. The left femoral artery is unoccluded.

Quantification of collateral blood flow in 5-FU-treated and control animals. Recovery of blood flow after ligation of the femoral artery in rabbits was assessed by quantification of the collateral conductance 1 wk after femoral artery ligation. Whereas in control groups the maximal collateral conductance was calculated as 99.8 ± 7.09 ml · min¹ · 100 mmHg¹, in the monocyte rebound rabbit group maximal collateral conductance was significantly increased to 153 ± 3.53 ml · min¹ · 100 mmHg¹ (Fig. 3A). The highest level was calculated for the MCP-1-treated control group (210.7 ± 1.89 ml · min¹ · 100 mmHg¹).

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Functional assessment of collateral artery growth. A: collateral conductance in the rabbit hindlimb 7 days after femoral artery occlusion. Collateral conductances were calculated from blood flow in the iliac arteries, and peripheral, and systemic pressures. * P < 0.001. B: blood flow ratio (R/L) in hindlimbs from mice without treatment (control), with rebound monocytosis [5-FU(reb)], or monocyte depletion [5-FU(dep)] assessed postoperatively and 3, 7, 14, and 21 days after femoral artery ligation. C: measurement of hemoglobin oxygen saturation in the mouse foot assessed postoperatively and at days 3, 7, 14, and 21. Open bars, monocyte rebound group; solid bars, untreated animals (control group); gray bars, monocyte depletion group. PBS, phosphate-buffered saline. * P = 0.01, compared with control.

Immunohistological analysis of monocyte accumulation in the rabbit hindlimb. To correlate the increased number of circulating monocytes in the rabbit rebound group with the accumulation of monocytes/macrophages along the growing collaterals we stained tissue sections of rabbits thighs with the macrophage-specific RAM11 antibody (37). Spleen tissue was used to confirm the specificity of this antibody to macrophages (Fig. 4, A and B). The higher magnification for spleen sections compared with sections from the thigh was used to demonstrate the typical nucleus morphology of monocytic cells. Only sections of growing collateral arteries according to a previously described phenotype (1, 31) were assessed. As shown in Fig. 4, C-F, macrophages were stained green in a diffuse pattern and were found in the monocyte rebound group (Fig. 4, C and F) as well as in the control group (Fig. 4, B and E). The accumulation of macrophages around growing collateral vessels often appeared in a cluster formation and was more prominent in the rebound group linking the enhanced arteriogenesis detected in the rebound group to monocytes/macrophages invading the perivascular space of arteriogenic vessels.

View larger version (61K): [in this window] [in a new window]   Fig. 4.   Monocytes/macrophages in rabbit spleen (A, B) and along the growing collateral arteries 3 days after occlusion (C-F). Tissue samples were stained with the rabbit antibody macrophage 11 (RAM11) antibody (monocytes/macrophages; green), tetramethylrhodamine isothiocyanate (TRITC) phalloidin (F-actin, red) and 4',6-diamidino-2-phenylindole (DAPI)/dimeric thiazole blue from Molecular Probes (TOTO-3; nuclear staining, blue). Compared with tissue from control rabbits (C and E), more macrophages (arrowheads) were accumulated in the perivascular space of the growing collateral arteries of the monocyte rebound group. Spleen sections were used to confirm the specificity of the RAM11 antibody. Bars indicate magnification.

Effect of CD11b+ (Mac-1⁺) monocytes on collateral blood flow in monocyte depleted mice. To confirm the hypothesis that monocytes are functionally involved in collateral artery growth, we included an additional experiment, in which the monocyte depletion was experimentally compensated. For this purpose, CD11b⁺ (Mac-1⁺) monocytes, which had been sorted from donor mice blood using flow cytometrical cell sorting (purity as controlled by flow cytometry: >96%, data not shown), were intravenously injected in different concentrations during the hindlimb operation. Compared with monocyte-depleted animals without monocyte rescue, R/L ratio assessed by laser Doppler imaging in the monocyte-injected animals was increased in mice that received monocytes at the seventh postoperative day and continued to increase throughout the observation period similar compared with animals that had not been monocyte depleted (Fig. 5). The injection of both 2 × 10⁵ and 5 × 10⁵ monocytes was sufficient to induce an enhancement in collateral growth; however, the obtained effects appeared to be more pronounced with the higher monocyte concentration.

View larger version (27K): [in this window] [in a new window]   Fig. 5.   Flow measurements (R/L ratio) in hindlimbs from monocyte-depleted and monocyte-rescued mice. Monocyte depletion by 5-FU treatment was compensated by injection of either 2 × 105 or 5 × 105 monocytes (mono transplanted, stippled bars) and data compared with monocyte-depleted mice without injected monocytes [5-FU(dep), open bars]. Untreated animals (gray bars) served as controls.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study, we present for the first time functional data on the role of monocytes in the growth of collateral arteries after an acute ligation of the femoral artery. Using two animal models, we demonstrate the stimulation of arteriogenesis by increased concentrations of blood monocytes. In contrast, collateral artery growth as well as blood flow restoration was inhibited in mice with reduced monocyte counts. This negative effect could be reversed when monocyte depletion was compensated by the injection of monocytes.

The current findings from experiments in two animal models of acute femoral artery ligation suggest a correlation of the blood monocyte concentration and the enhancement of arteriogenesis. The antimetabolite 5-FU was previously demonstrated to affect bone marrow populations in animal experiments, for instance, it influences megakaryopoiesis and granulocytopoiesis (23, 29). In contrast, the lymphocyte concentrations are almost unaltered, which was confirmed by our data (not shown). Whereas there are well-established published protocols (16, 36, 38) for 5-FU treatment in mice, in rabbits, 5-FU was previously used in several concentrations and different modes of administration. For the aims of this study, we ascertained that an intravenous 5-FU injection of 20 mg/kg was tolerated by the rabbits and adequate to induce a rebound monocytosis between day 3 and day 4 after administration. In our study, the monocyte rebound was synchronized with the ligation of the femoral artery leading to an enhancement of the collateral conductance index after 7 days compared with control animals. In postmortem angiograms an increased number and diameter of visible collaterals was observed. The corresponding data from the mouse model confirmed these observations. Our functional data from this study are in line with the immunohistological evidences on the role of monocytes obtained in several earlier studies. For instance, monocytes adhere at the collateral vascular endothelium as early as 12 h after femoral artery ligation in correlation with the endothelial upregulation of the adhesion molecule ICAM-1 (31). The adhesion of monocytes at the shear stress-activated endothelium and the subsequent migration into the perivascular space were suggested to be the basis for subsequent cell proliferation, remodeling processes, and growth of the collaterals as the monocytes were shown to mature within the collateral vessel wall into tissue macrophages, concomitantly releasing growth factors, inflammatory cytokines and metalloproteinases (1, 20, 39). Using the confocal scanning laser microscopy technique, we show that circulating monocyte levels and the enhancement of collateral artery growth correlate with the extent of macrophage accumulation in the rabbit thigh 3 days after occlusion of the femoral artery. Typically, these cells appeared in clusters, suggesting that there may be areas with a stronger activation. This activation may result from enhanced production of vascular adhesion molecules and chemokine production including MCP-1 (8) in response to locally increased shear stress.

Previously, Ito et al. (15) demonstrated that the application of exogenous MCP-1 to the collateral system can enhance the inflammatory process with the attraction of monocytes induced by endogenous MCP-1. Of note, in our study with ligated rabbits, blood perfusion recovery after local MCP-1 infusion was still superior to the values observed with the rebound effect. On the basis of the essential role of MCP-1 elicited integrin upregulation and activation on monocytes (12), it is possible that the shear stress-induced endogenous production of MCP-1 in animals is rather low compared with levels reached by a local MCP-1 administration.

This study is the first experimental demonstration that increasing blood monocyte concentrations lead to increased collateral vessel formation and increase in blood flow in our model of femoral artery occlusion. Increased blood monocyte concentrations were achieved by the so called monocyte rebound effect after 5-FU treatment in both rabbits and mice but also by intravenous infusion of either CD11b⁺ or CD14⁺ (data not shown) positive blood monocytes. In all these cases significant increase in collateral artery formation and blood flow in the impaired regions was demonstrated. These findings are in line with data from a clinical study recently published by Seiler et al. (32) presenting a potential of cerebrospinal fluid (granulocyte monocyte-colony stimulating factor) to improve collateral flow in patients with coronary artery disease via stimulation of collateral artery growth accompanied by a 2.7-fold increase of peripheral blood monocytes after granulocyte monocyte-colony stimulating factor treatment. Conversely, when in our mouse experiments monocyte concentrations were significantly reduced between day 4 and day 10 after 5-FU injection a decrease in collateral artery formation and blood flow was observed. However, collateral growth was not completely impaired in this group. This may be due to our observation that monocytes were not completely removed from the blood during the treatment with 5-FU (~90% removal, see Fig. 1C) and the remaining cells may be sufficient to initiate the remodeling and growth processes. Furthermore, it is worth mentioning that in capillary growth in the rabbit, the thigh was not affected during the observation period (data not shown), which is in line with previous observations (14).

Our demonstration that decreased arteriogenesis in 5-FU-treated animals can be rescued by infusion with purified monocytes adds further evidence for the hypothesis that monocytes are essential mediators of collateral growth. Furthermore, it suggests that the enhancement of collateral growth observed in the rebound group is not secondary to factors produced by mobilized bone marrow cells. Of note, we observed after the injection of a rather small number of isolated monocytes even higher perfusion recovery values than with normal monocyte counts displaying control mice. This may be due to the artificial activation of monocytes caused by the isolation procedure. Activated monocytes have been shown to express increased levels of integrins (for review on leukocyte integrins, see Ref. 11). Therefore, the initial adhesion of the monocytes with the collateral endothelium may have been enhanced or accelerated leading to the superior blood flow reconstitution values.

As a further tool to support information on the stimulation of arteriogenesis we performed measurements of the hemoglobin oxygen saturation distal to the collateral system in the mouse feet. Although we found significant differences between the animal groups 3 days after ligation, indicating the different extent of collateral growth, oxygen saturation in all groups was roughly restored after 7 days. This may be because the animals switch to an "economy mode," thereby avoiding excessive moving of the affected leg. The reduced oxygen consumption by the lower limb tissue under these resting conditions, together with the increased paw blood flow, may therefore be responsible for the differences in progression of the oxygen saturation and the relative blood perfusion recovery curve.

It is likely that the 5-FU treatment in addition to the monocyte rebound also led to the release of further cell types, i.e., stem cells. Recently, evidence was presented for the hypothesis that endothelial progenitors from the bone marrow can induce blood vessel growth after birth (2, 4). The cells that differentiate into endothelial cells in vitro were identified by cellular markers like CD34 (3, 21, 33). In principle, the release of cells from the bone marrow as a rebound reaction after 5-FU treatment can also include these hematopoetic progenitor cells, which may also contribute to the growth of collateral arteries. However, our rescue data with CD11b (Mac-1) purified blood monocytes favor the concept that monocytes but not endothelial progenitor cells play a major role in collateral growth-dependent blood flow restoration. This hypothesis is also supported by a recent report (28) demonstrating that injection of CD34-positive cells improves blood flow restoration in a similar model of hindlimb ischemia only in diabetic but not in normal mice. In addition, the time course of mobilization of these putative precursors from bone marrow (34) and their differentiation into endothelial cells (17, 35) is much longer than our experimental setting, where we found a significant increase in arteriogenesis as soon as 7 days after ligation. Furthermore, CD34⁺ cells represent a heterogeneous population and only a subset of them expressing VEGFR-2 and AC133 antigen (22) and in addition CD34/CD14⁺ monocytes can express typical endothelial markers like von Willebrand factor, VE-cadherin, and endothelial cell nitric oxide synthase when they are under angiogenic stimulation (30). Finally, according to the published data, stem cell rates in peripheral blood after pharmacological mobilization treatment are usually <1%, a number that may be inconsiderable compared with monocyte numbers.

In conclusion, this study provides functional evidence for a key role of monocytes in arteriogenesis. Furthermore, these data suggest that therapeutic intervention leading to increased monocyte counts such as treatment with M-CSF or GM-CSF is a promising strategy for the stimulation of growth of collateral blood vessels in ischemic tissues.