CD34– Blood-Derived Human Endothelial Cell Progenitors MAGED HARRAZ, CHUNHUA JIAO, HEATHER D. HANLON, REBECCA S. HARTLEY, GINA C. SCHATTEMAN Department of Anatomy and Cell Biology, University of Iowa, Iowa City, Iowa, USA Key Words. Stem Cells · Endothelial cell · Angiogenesis · CD34 · Monocyte · CD14
A BSTRACT A subset of adult peripheral blood leukocytes functions as endothelial cell progenitors called angioblasts. They can incorporate into the vasculature in animal models of neovascularization and accelerate the restoration of blood flow to mouse ischemic limbs. Earlier reports suggested that CD34-expressing (CD34+) but not CD34+ cell-depleted (CD34–) leukocytes can differentiate into endothelial cells (EC) in vitro and in vivo. Recent findings suggest that CD14+ cells, which are typically CD34–, also have angioblast-like properties in vitro. To determine the identity of angioblasts, the potential of CD34+, CD34–, CD34–CD14+, and CD34–CD14– cells to produce EC was compared. We show that a subset of
monocyte (CD34–CD14+)-enriched cells can take on an EC-like phenotype in culture, but that the EC-like cells also express dendritic cell antigens. These findings suggest that monocytes differentiate into macrophages, dendritic cells, or EC depending on environmental cues. The data also demonstrate that angioblasts are more abundant in the blood than previously thought. Finally, we demonstrate that CD34– and CD34–CD14+ cells incorporate into the endothelium of blood vessels in mouse ischemic limbs. However, incorporation of these cells requires co-injection with CD34+ cells, indicating that leukocyte-leukocyte interactions may play a critical role in governing angioblast behavior in vivo. Stem Cells 2001;19:304-312
INTRODUCTION Endothelial cell (EC) progenitors (angioblasts) derived from adult blood play an important role in vascular repair. Blood-derived cells represent ~10% of EC in the neovasculature formed in response to surgical sponge implantation in mice, and a similar fraction of blood-derived cells appears to be present in the neovasculature associated with angiogenesis in mice after induction of hind limb ischemia [1]. Blood-derived cells also contribute to nonpathological neovascularization [2]. EC progenitors have been identified among human leukocytes enriched for CD34-expressing (CD34+) cells, [3-5] and in whole mononuclear cell fractions of peripheral blood and leukapheresis products [6]. It has never been demonstrated directly that CD34+ cells are angioblasts, but several studies suggest that CD34+ cells are enriched for angioblasts and CD34– cells lack them. For example, leukocytes enriched for CD34+ cells, but not those depleted of CD34+ cells (CD34– cells), produced EC in vitro [3, 5]. Also, vital dye labeling of CD34+-enriched cells in coculture with unlabeled CD34– cells suggested that all of the
cells were derived from the CD34+ cell-enriched fraction [3]. Additionally, in vivo, exogenous human leukocytes enriched for CD34+ cells, but rarely CD34– cells incorporated into the hind limb vasculature of mice recovering from hind limb ischemia [3, 7, 8]. Moreover, while exogenous CD34– cells had no effect on the rate of restoration of blood flow to the ischemic limb, CD34+ cells profoundly accelerated it [7, 8]. Nevertheless, three factors led us to consider the idea that angioblasts might be present among CD34– cells. First, over the past several years, it has become clear that subsets of both CD34+ and CD34– cells are capable of long-term hematopoietic repopulation [9-11]. Because embryologically EC progenitors and hematopoietic stem cells are related, we have postulated that the same may be true in adults [12]. Thus, if angioblasts are related to hematopoietic stem cells, they could be either CD34+ or CD34–. Second, because only ~0.1% of leukocytes express CD34 and yet as many as 10% of EC in the mouse neovasculature are blood-derived cells, we questioned whether such a small population could have such a profound effect on neovascularization [8, 13, 14].
Correspondence: Gina C. Schatteman, Ph.D., Anatomy and Cell Biology, BSB 1402, University of Iowa, Iowa City, IA 52242, USA. Telephone: 319-335-7765; Fax: 319-335-7198; email:
[email protected] Received September 8, 2000; accepted for publication April 16, 2001. ©AlphaMed Press 1066-5099/2001/$5.00/0
STEM CELLS 2001;19:304-312
www.StemCells.com
Harraz, Jiao, Hanlon et al. Third, since monocytes traverse the vascular wall during injury, share many antigenic characteristics with EC, and are responsive to vascular endothelial growth factor (VEGF) (and VEGF appears to mobilize angioblasts), we considered the possibility that monocytic cells might also function as EC progenitors [15-18]. In this study we examined the potential of CD34– cells, and two different subsets of CD34– cells that are enriched for monocytes to produce EC. We demonstrate that CD34– and particularly CD34– monocyte-enriched cells readily differentiate into EC-like cells in culture and that they incorporate into the neovasculature in vivo when coinjected with CD34+ cells. Our findings indicate that cells with an angioblastic potential are abundant in the blood, that monocytes probably act as angioblasts, and that leukocyte-leukocyte interactions may play an important role in modulating angioblast behavior. While this work was under way, two studies were published that support the hypothesis that monocytic-like cells can function as angioblasts. In the first, Fernandez-Pujol and colleagues showed that under appropriate in vitro conditions, CD14+ cells (which are typically CD34–) differentiate into EC-like cells exhibiting characteristics of both EC and monocytes [19]. More recently, Moldovan et al. found that when macrophages were induced to infiltrate the heart by overexpression of MCP-1, the invading macrophages appeared to form erythrocyte-containing vascular-like channels reminiscent of the tumor cell-derived vascular channels seen in some tumors [20, 21]. However, cells of the monocytic channels lack some EC antigens and may themselves be subsequently colonized by EC or angioblasts. METHODS Isolation of Leukocyte Subsets Human blood was collected from healthy adult volunteer donors according to protocols approved by the University of Iowa Institutional Review Board as previously described [8]. Briefly, blood was collected into 1/10 volume of 0.13 M sodium citrate as anticoagulant and used within 1 hour. Blood was diluted with 0.5 volumes phosphate-buffered saline with anti-coagulant (PBS/AC) (1.1 mM KH2PO4, 3 mM Na2HPO4, 0.9% NaCl, 0.13 M sodium citrate), pH 7.2, and the leukocytes were fractionated from red blood cells by centrifugation on Histopaque 1077 (Sigma; St. Louis, MO; http://www.sigma-aldrich.com) gradients according to manufacturer’s instructions. Leukocytes were washed once in PBS/AC and twice in PBS/AC with 2% bovine serum albumin (PBS/AC/BSA) then resuspended in PBS/AC/BSA. CD34+ cell enrichments were done using CD34 antibody-coated magnetic beads (Dynal; Lake Placid, NY; http://www.dynal.no)
305
per manufacturer’s instructions. The residual depleted leukocytes served as CD34– cells. To obtain CD34–CD14+ and CD34–CD14– cells, CD14+ cells were selected from CD34– cells using CD14 antibody-coated magnetic beads (Dynal) according to manufacturer’s instructions. The residual CD14– cells were depleted of CD14+ cells a second time. To obtain monocyte-enriched fractions, CD34– cells were depleted of lineage marker-expressing cells using the StemSep magnetic immunodepletion system (Stem Cell; Vancouver, BC) per manufacturer’s instructions. Fluorescence-Activated Cell Sorter (FACS) Analysis To assess the purity of leukocyte subsets, freshly isolated cells were subjected to FACS analysis using a Class 3 mouse anti-CD34 antibody (Silenius) or a rabbit anti-CD14 antibody as previously described [22, 23]. Briefly, cells were resuspended at 1 × 106 cells/ml in PBS/AC with 1% human and 1% goat serum and 4 µg/ml antibody and incubated 30 minutes on ice. Cells were washed twice then pelleted through a fetal calf serum (FCS) cushion. Cells were incubated again as above with a 1:150 dilution of fluorescein isothiocyanate (FITC)-conjugated secondary antibody (BD PharMingen; San Diego, CA; http://www.pharmingen.com). Cells were washed and pelleted through FCS as above and resuspended at 1 × 106 cells/ml in PBS for FACS. Because CD34+ cells represent less than 0.1% of leukocytes before CD34+ depletion, CD34– cells were not analyzed for CD34+ content [13, 14]. Such a small percentage would not be discriminated in our FACS assay. FACS analysis on cultured cells was done as above after cells were detached from the culture plate by incubation with 2 mM EDTA in PBS. For additional analyses, antibodies to tie-2 (Santa Cruz Biotechnology; Santa Cruz, CA; http://www.scbt.com), MUC18 (Chemicon; Temecula, CA; http://www.chemicon.com), endoglin (BD PharMingen), and CD1a and CD45 (Dako; Carpenteria, CA; http://www.dako.dk) were used. A minimum of 1 × 104 events were counted in all analyses. Cell Culture, Counting, and Dye Labeling Cells were plated at 4 × 105 cells per well, except as noted, in 24-well trays precoated with 5 µg/cm2 human fibronectin (BD Biosciences; Franklin Lakes, NJ; http://www.bd.com) according to manufacturer’s instructions. Cells were cultured in medium D (M199 #12340, GIBCO; Gaithersburg, MD), 20% heat-inactivated fetal bovine serum, 2 ml/L bovine brain extract (#CC-4092, Clonetics; San Diego, CA; http://www. clonetics.com), and 2× antibiotic/antimycotic (#15240-0620, GIBCO). The medium was changed on the fourth day after plating and every 3 to 4 days subsequently. Experiments were done in duplicate or triplicate and all experiments with CD34+ cells were performed at least four times, and with CD34– cells
306
at least six times. To determine the number and proportion of EC in cultures, spindle-shaped cells and total cells in 11 representative fields per well (5% of well) were counted using phase contrast microscopy at various times after plating. Spindle shape was used as the criterion for differentiation into EC because in our culture system >95% of cells are spindleshaped by 8 days in culture, and essentially all spindle-shaped cells express Tie-2, and take up acetylated low density lipoprotein (acLDL) [3, 8]. To obtain conditioned medium (CM), CD34– cells were plated at 2 × 106 cells per well as above and cultured for 2 days. The medium was changed, and thereafter the medium was collected and replaced with fresh medium every 3 days for 9 days. For experiments testing the effects of CM, freshly isolated CD34–, CD14–, or monocyte-enriched cells were plated in one-third fresh medium and two-thirds CM. The medium was changed every other day. For CM-DiI or SP-DiO (Molecular Probes; Eugene, OR; http://www.probes.com) labeling, freshly isolated CD34+ or CD34– cells were resuspended at 106 cells/ml and incubated in 2 µg/ml CM-DiI or SP-DiO in media for 5 minutes at 37°C then 15 minutes on ice. For low dye concentration experiments, cells were labeled with 0.2 µg/ml CM-DiI or 0.5 µg/ml Sp-DiO. Cells were pelleted and washed three times in medium (106 cells/ml) to remove unincorporated dye. CMDiI or SP-DiO labeling at these concentrations had no effect on cell viability. Cell Immunolabeling and acLDL Uptake To assess the ability of cells to take up acLDL, 10 µg/ml DiI acLDL (Biomedical Technologies; Stoughton, MA; http://www.btiinc.com) were added to the medium and cells were returned to the incubator for 4 hours. Cells were then washed three times with PBS, and examined by confocal microscopy (due to the DiI tag). The fraction of DiI acLDLlabeled cells was determined from these data. Human coronary artery smooth muscle cells (Cascade Biologicals; Portland, OR; http://www.cascadebio.com) and mouse FTO-2B cells (ATCC; Manassas, VA; http://www.atcc.org) were cultured according to the suppliers’ specifications and were used as negative control cells. Tissue culture cells were rinsed three times in PBS, once in water, and fixed in methanol for 5 minutes. Cells were blocked for 1 hour in 4% calf serum in PBS then incubated for 1 hour at room temperature, then overnight at 4°C in 1 µg/ml anti-Tie-2 (Santa Cruz Biotech), 5 µg/ml vascular endothelial (VE)-cadherin (Chemicon), 10 µg/ml von Willebrand’s factor (vWF) (Dako) or concentration and species-matched nonimmune rabbit or mouse IgG (Sigma). Cells were washed three to six times in PBS, incubated 1-2 hours in biotinylated goat anti-rabbit or mouse IgG antibody (Vector Laboratories;
CD34– Endothelial Cell Progenitors Burlingame, CA; http://www.vectorlabs.com), washed three to six times in PBS, then incubated for 1 hour in 1:100 streptavidin-FITC (Vector) before mounting for confocal microscopy. Cells were examined using a Bio-Rad MRC-1024 (Bio-Rad Laboratories; Hercules, CA; http://www.biorad.com) confocal microscope equipped with a krypton-argon laser or by conventional fluorescence microscopy. Reverse Transcriptase-Polymerase Chain Reaction (RT-PCR) To test for the presence of eNOS mRNA, freshly isolated CD34– cells were plated on fibronectin in medium D as above. After 0 days, 12 days, or 16 days in culture, cells were rinsed twice in PBS and mRNA was extracted using Trizol Reagent (GIBCO) per manufacturer’s instructions. Reverse transcription was carried out for 1 hour at 42°C on 1 µg of RNA in 1× First Strand Buffer (GIBCO) using Moloney murine leukemia virus (M-MLV) RT (GIBCO). For PCR amplification, 2 µg of the produced DNA were amplified for 35 cycles in First Strand Buffer using AmpliTaq DNA polymerase (Perkin Elmer; Foster City, CA; http://www.instruments.perkinelmer.com/index.asp) using the following cycles: 94°C for 3 minutes, 58°C for 1 minute, 72°C for 3 minutes for the first cycle and 94°C for 1 minute, 58°C for 1 minute, 72°C for 3 minutes for 35 cycles. Human primers were GACATTTTCGGGCTCACGCTG (forward) and TTGGGTAGGCAGTTTAGTAGTTCTC (reverse) for EC nitric oxide synthase (eNOS), CACCGTTTGCCCACCCTTCG (forward) and GCCCACTGG-GAGCCGACACT (reverse) for vWF, and GATGCAGAGG- CTCATGATGC (forward) and CTTGCGACTCACGCTTG-ACT (reverse) for VE-cadherin. Hind Limb Ischemia, Cell Injection, and Blood Vessel Labeling All procedures were performed on nude (HFh11nu) mice (Jackson Laboratories; Bar Harbor, ME; http://www.jax.org) according to University of Iowa Animal Care and Use Committee guidelines. Diabetes was induced using streptozotocin as described [8, 24] and surgery was performed 2-3 weeks after the last streptozotocin injection. Nude mice were used to minimize possible host-versus-graft immune responses to the transplanted human cells. Surgery to induce hind limb ischemia was completed as has been described [8, 25]. Three to 6 hours after surgery, the ischemic limb was injected i.m. with 1 × 106 freshly isolated CM-DiI-labeled CD34– or monocyte-enriched cells and 5 × 105 unlabeled CD34+ cells in 0.9% NaCl. Mice were killed 5 days later, and the thigh hind limb muscles were removed, fixed in methanol, paraffin embedded, and sectioned at 8 µm. Sections were incubated with Bandeira simplicifolia lectin B4 to identify murine EC as previously described [26].
Harraz, Jiao, Hanlon et al.
307
Figure 1. Morphology of leukocytes cultured in fresh and CM. Phase contrast images of cultures. Spindle-shaped cells 12 days after plating in (A) CD34+, (B) CD34–, and (C) CD34–CD14+ cell cultures plated at 4 × 105 cells/well in medium D. (D) Cordlike structures in CD34– cell culture. Spindle-shaped cells in CD34–CD14+ cell cultures plated at 2 × 105 cells per well in (E) fresh medium D or (F) CD34– cell CM D for 8 days. (G) CD14– cells plated at 2 × 105 cells per well in CD34– cell CM D for 8 days.
RESULTS Morphology of Cultured Cells Human peripheral blood leukocytes were fractionated using antibody-coated magnetic beads. Cells were enriched for CD34+ cells, and the remaining depleted cells served as the CD34– population. The average purity of CD34+ cells as assessed by FACS was 22% ± 4%. Purity of CD34– cells was not assessed because the starting population is >99.9% pure. To enrich for monocytes, CD14+ cells were selected from CD34– cells (>99% pure per manufacturer) and the depleted cells served as CD14– cells (>99% pure by FACS). A second monocyte-enriched subpopulation (74% ± 5% CD14+ cells by FACS) was obtained by depleting CD34– cells of leukocytes expressing nonmonocyte lineage markers. Freshly isolated cells were plated at either high or low density on fibronectin in medium D, a medium in which human umbilical vein EC (HUVEC) thrive. As previously reported, a subset of freshly isolated CD34+-enriched cells rapidly attached and began to assume a spindle-shaped morphology by 4-6 days in culture [3, 8]. Cells morphologically similar to CD34+ cells were present in cultures of CD34–, CD14+, or monocyte-enriched cells (Fig. 1A-C). The proportion of spindle-shaped cells was higher in CD14+ cultures (82% ± 2%) than in CD34– cultures (63% ± 3%). In CD14– cell cultures no spindle-shaped cells were observed, and viable cells were rarely present after 12 days in culture (data not shown.) Cord-like structures, commonly formed by EC, were observed in low density cultures that produced spindle-shaped cells (Fig. 1D). Since no spindle-shaped cells were produced by CD14– cells cultured in fresh medium D, we tested if factors produced by other leukocytes were required for CD14– cell differentiation. CD14–, CD14+, and CD34– cells were plated in CM from cultures of CD34– cells. While CM profoundly augmented production of spindle-shaped cells in CD34– and CD14+ cell cultures, again, no spindle-shaped cells were found in CD14– cell cultures (Fig. 1E-G).
Antigenic and Physiologic Properties of Cultured Cells CD34+ cell-derived EC have been well characterized [3, 8]. To determine if CD34–, CD14+, or monocyte-enriched cells cultured in medium D exhibit similar properties, cells were fixed after 12 days or 16 days in high density culture and immunolabeled with antibodies to tie-2. As with CD34+derived EC, all spindle-shaped cells in the three cultures labeled with anti-tie-2 antibody (Fig. 2A), whereas incubation with nonspecific IgG yielded no labeled cells (Fig. 2B). In contrast, freshly isolated monocyte-enriched cells do not express tie-2 by FACS analysis (Fig. 2C). When cells were tested for the expression of flk-1, by immunostaining after 10 days in culture all spindle-shaped cells expressed low levels of the antigen (data not shown.) EC take up acLDL. When DiI-labeled acLDL was added to the medium of 12-day cultures of CD34– cells, and the cells were examined 4 hours later, spindle-shaped CD34– cells fluoresced red, indicating that they had taken up the acLDL. (DiI is a red fluorescent dye.) (Fig. 3A). In contrast neither rat smooth muscle cells (data not shown) nor FTO-2B hepatocytes (Fig. 3B) were fluorescent. eNOS is a specific marker of EC, so cultured cells were assayed for eNOS mRNA. CD34+, CD34–, and monocyteenriched cells were plated in medium D, harvested after 1213 days, and their RNA was extracted. RT-PCR analysis using eNOS-specific primers revealed a band of the predicted size in reactions from CD34+, CD34–, and monocyte-enriched cells and in HUVEC (positive control), but not in human smooth muscle cells or freshly isolated CD34– cells (Fig. 3C). The intensity of the bands from CD34+, and monocyteenriched cultures relative to HUVEC mirrored the proportion of spindle-shaped cells in the cultures. DNA sequencing of the excised gel band verified that it represented eNOS. The
308
CD34– Endothelial Cell Progenitors
Figure 2. Tie-2 expression by fresh and cultured monocytes. (A and B) Confocal images of immunostained cultures of CD34– cells 10 days after plating in medium D. Visualization with FITC anti-goat IgG after incubation with (A) anti-Tie-2 or (B) non-immune IgG. (C) Histograms of FACS data from freshly isolated monocyte-enriched fractions immunolabeled with rabbit IgG (left) or anti-Tie-2 (right). Left edge of line indicates 95% gate position.
same populations were examined for production of VEcadherin (CD144) and vWF mRNA after 12-13 days in culture. Whereas vWF mRNA was detected in HUVEC, CD34–, and monocyte-enriched cell cultures, VE-cadherin mRNA was not detected. However, when HUVEC, CD34–, and monocyte-enriched cell cultures were plated at high density and assayed by immunostaining at 24 days in culture, all HUVEC and all spindle-shaped but not round cells in CD34– and monocyte-enriched cell cultures expressed VE-cadherin (Fig. 3D-E). FACS was used to assess expression of other antigens. The EC antigen MUC18 [27] was detected on 14% ± 5% of CD34– cells [27]. Endoglin, which is expressed on activated EC and dendritic cells (which are derived from monocytes), was found on 59% ± 5% of CD34– and 78% ± 9% of monocyte-enriched cells in 16-day cultures [28-30]. The dendritic cell antigen CD1a was detected on 83% ± 5% and 85% ± 9% of 16-day CD34– and monocyte-enriched cells, respectively. Also, the pan-leukocyte antigen CD45 was localized to 94% ± 5% of 12-day CD34– cells. Hence, the phenotype of the cultured cells was intermediate between EC and dendritic cells. Abundance of CD34– and CD34+ EC Progenitors in the Blood To ascertain whether CD34+ cells are enriched for angioblasts, CD34+ and CD34– cells were plated in
Figure 3. Phenotype of cultured leukocytes. Confocal images 12 days after plating of (A) CD34– cells or (B) control FTO-2B cells incubated with DiIlabeled ac-LDL. (C) Ethidium bromide stained agarose gel of eNOS RT-PCR products from mRNA of cells cultured 12-13 days in medium D. Lane (1) Molecular weight standard, (2) freshly isolated CD34– cells, (3) CD34– cells after 12 days in culture, (4) CD34–CD14+ after 12 days in medium D, (5) positive control HUVEC, and (6) negative control human vascular smooth muscle cells. (D and E) Monocyte-enriched cells after 21 days in culture at high density incubated with anti-VE-cadherin and visualized with an FITC-conjugated anti-IgG. (D) Fluorescent microscopic image showing labeling of spindleshaped cells. (E) Phase contrast image of culture in (D) showing two large unlabeled rounded cells (arrows).
monoculture at equal density on fibronectin in medium D. CD34+ and CD34– cells produced similar numbers of tie-2+ spindle-shaped cells regardless of the density of plating (Fig. 4A). Additionally, when CD34+ and CD34– cells were cocultured the number of EC produced was independent of CD34+ cell: CD34– cell ratio (Fig. 4A). That is, the number of EC produced was a function of the total number of cells plated, regardless of the proportion of CD34+ and CD34– cells in the cultures (Fig. 4A). The possibility that the total density of cells in the cultures might alter the ability of CD34+ and CD34– cells to produce EC was also considered. Thus, cells were plated at various densities ranging from 1 × 105 to 8 × 105 cells per well. There was a nearly linear relationship between the number of cells plated and the number of EC produced in monoculture or in coculture
Harraz, Jiao, Hanlon et al.
309
wash conditions, CM-DiI was readily transferred among cells either through cell-cell contacts or via residual dye in the medium (data not shown). This was also true in a 1:1 mixed culture of DiI-labeled and DiO-labeled CD34– cells. To reduce dye transfer, cells then were labeled with 0.2 µg/ml CM-DiI or 0.5 µg/ml Sp-DiO and given two additional washes. In cocultures of equal numbers of CM-DiIlabeled CD34+ and Sp-DiO-labeled CD34– cells, 46%-59% of cells were CM-DiI-labeled whereas all (>99%) of cells were Sp-DiO-labeled (Fig. 5). These data do not support the hypothesis that CD34+-derived EC are the only or predominant phenotype in cocultures of CD34+ and CD34– cells, but rather indicate that the more abundant CD34– cells are the predominant class of EC progenitors in blood. Further, whereas at low concentrations transfer of CM-DiI between cells was not apparent, Sp-DiO was readily transferred to all cells in these conditions.
Figure 4. Effect of coculture of CD34+ and CD34– cells on cell number. (A) 3 × 105 total CD34+ and CD34– cells were plated on fibronectin in medium D at various ratios for 9-11 days. The number of EC (spindle-shaped cells) in the cultures was determined. Data are expressed as the number of cells in the cocultures relative to 300,000 CD34+ cells in monoculture at the same time. (B) Total CD34+ and CD34– cells were plated on fibronectin in medium D at various ratios for 9-11 days at the indicated densities and the number of EC determined. The theoretical line represents the expected number of cells if plating efficiency is not affected by plating density. Error bars indicate standard errors.
regardless of the ratio (which varied from 5:1 to 1:5) of CD34+ to CD34– cells (Fig. 4B). At densities of less than 1 × 105 cells per well, cultures typically were not viable. In an earlier study, when DiI-labeled CD34+ cells were cocultured with unlabeled CD34– cells, essentially all spindle-shaped cells were DiI-labeled [3], suggesting that only CD34+ cells produced EC-like cells. Yet, our current in vitro data suggest that both CD34+ and CD34– cells produce EC-like cells. We, thus, tested the possibility that, DiI might be transferred among cells. CD34+ cells were labeled with CM-DiI and CD34– cells with SP-DiO (a related green dye) or vice versa. Cells were plated and examined 10 days later. In cocultures of equal numbers of CD34+ and CD34– cells, essentially all cells were colabeled with DiI and DiO indicating that at previously used DiI concentrations and
Vascular Incorporation of CD34– Angioblasts Requires CD34+ Cells Earlier studies found that whereas exogenous CD34+ enriched cells incorporate into the neovasculature of mouse ischemic limbs, CD34– cells rarely do. Further, CD34+ but not CD34– cells accelerate the rate of restoration of blood to an ischemic limb. Yet, our in vitro findings suggest that CD34– cells are angioblasts. Since CD34+ cells are typically ~15% CD34+ and 85% CD34–, we tested the hypothesis that CD34+ cells are required for the incorporation of CD34– cells into the vasculature. CD34– cells or monocyte-enriched cells were labeled with CM-DiI at low dye loadings (0.2 µg/ml), at which concentration the dye is not transferred to unlabeled
Figure 5. Dye transfer in cultured leukocytes. Confocal images of 1 × 105 0.2 µg/ml DiI-labeled (red) CD34+ and 1 × 105 0.5 µg/ml DiO labeled (green) CD34– cells cocultured in medium D for 10 days. (A) DiI-labeled cells. (B) DiOlabeled cells. Note that all cells are labeled with DiO whereas only some are DiI-labeled.
310
CD34– Endothelial Cell Progenitors
cells after at least 10 days in culture. 5 × 105 CM-DiI-labeled CD34– or CD14+ and 5 × 105 unlabeled CD34+ cells were coinjected into the muscle of streptozotocin-induced diabetic nude mice 4 hours after induction of hind limb ischemia. Limbs were harvested 3-5 days later. The muscle was sectioned and examined for the presence of CM-DiI-labeled cells in capillaries that were delineated using the lectin Bandiera simplicifolia lectin B4 which binds to murine but not human EC [31]. Elongated CM-DiI-labeled cells in sections from both CD34– and monocyte-enriched cell injected mice were observed interdigitated between lectin-labeled cells in a manner consistent with their being capillary EC (Fig. 6). Consistent with previous results, CM-DiI-labeled cells never or rarely localized to the vasculature in mice injected with CD34– cells alone [3, 7, 8]. DISCUSSION Recently we showed that blood-derived cells represent approximately 10% of endothelial cells in the neovasculature in a sponge model of angiogenesis in the mouse [1]. If angioblasts are a truly rare cell population, this finding seems hard to reconcile with data from earlier studies, including our own, that suggest that CD34+ cells are enriched for adult angioblasts. Previously, we and others were unable to generate EC in CD34– cell cultures [3, 4, 12], whereas here CD34– cells readily generated EC. However, the culture conditions here are different than those reported previously, and we have found that morphology and long-term viability are strongly dependent on culture conditions. Indeed in medium D, EC appear essentially nonproliferative whereas highly proliferative EC have been reported in other leukocyte culture systems [5, 6] and we find that at high-density blood-derived EC are proliferative [Harraz and Schatteman, unpublished observations]. Earlier vital dye-labeling experiments also suggested that only CD34+ cells could produce EC or at least that CD34+ cells somehow suppressed EC production by CD34– cells [3], but it now appears that dye transfer among the cells accounts for the apparent predominance of CD34+-derived cells in those cultures. It was further suggested that coculture of CD34+ with CD34– cells enhanced production of CD34+ cellderived EC. Our data offer a simpler interpretation. The number of EC produced is simply a function of the number of cells plated and is unrelated to whether they are CD34+ or CD34–. If CD34+ cells were repressing production of EC by CD34– cells or CD34– cells were enhancing production of EC by CD34+ cells, one would expect that varying the ratio of CD34+ to CD34– cells would affect the production of EC. No such effect was observed. Our data do not directly answer the question of whether at least some CD34+ cells are angioblasts. Because our CD34+ cell cultures are typically only 15.9% ± 3.3% pure,
Figure 6. Incorporation of CD34– cells into ischemic limb vasculature. Artificially colored confocal images of two 8 µm sections of mouse ischemic limb muscle 5 days after injection with unlabeled CD34+ cells and DiI-labeled (red) CD34– cells. Sections were labeled with biotinylated Bandeira simplicifolia lectin B4 and FITCstreptavidin to visualize mouse EC. Elongated red human cells can be seen incorporated into the green mouse endothelium. Sections appear yellow where green and red cells overlap.
it is possible that no blood-derived EC progenitors express CD34. In support of this, Hernandez et al. [6] were unable to produce EC in CD34+-enriched cells in conditions that supported EC growth from less homogeneous leukapheresis samples. Also, while cultures of CD34– cells always produced EC, occasionally CD34+ cell cultures did not, perhaps because they were too pure. However, our CM data indicate that CD34– cells produce factors that enhance angioblast differentiation, so it is equally likely that factors produced by CD34– cells are lacking in more purified CD34+ cell-enriched cultures. Whatever the case with respect to CD34+ cells, since CD34– cells are many-fold more abundant than CD34+ cells and equal numbers of CD34+ and CD34– cells produce similar numbers of EC, at a minimum CD34– cells are the predominant form of EC progenitors in the blood. Further, because CD34–CD14– fail to produce EC even when cultured in CD34– cell CM, the CD34–CD14+ subset of leukocytes is probably enriched for angioblasts. The phenotype of the EC produced by the cultured leukocytes is noteworthy. Based on immunostaining, all spindleshaped cells in the cultures express tie-2, flk-1, and, after a long period in culture, VE-cadherin. Additionally the level of expression of eNOS mRNA roughly correlates with the fraction of spindle-shaped cells in the culture. That is, the level of
Harraz, Jiao, Hanlon et al. eNOS mRNA expression by CD14+ cultures (82% pure) is slightly less than that of HUVEC (100% pure) and more than CD34– cells (63% pure). Similarly, the proportion of cells expressing endoglin mirrors the proportion of spindle-shaped cells in the culture. Some, but not all cells in the cultures express vWF and the EC-specific antigen MUC [27]. These findings suggest strongly that the cultured cells are EC. However, vWF, tie-2, and flk-1 are all expressed on subsets of leukocytes, and endoglin is expressed by dendritic cells. Further, most of the cultured cells express the dendritic cell antigen CD1a and the pan leukocyte antigen CD45, neither of which is thought to be expressed by EC. Thus, it appears that the cultured cells have a mixed EC-dendritic cell phenotype. The differentiation of monocytes into macrophages, dendritic cells, or EC is presumably dependent on environmental cues, and in our culture system, some factors that drive monocytes into the EC pathway and away from differentiation into either macrophages or dendritic cells are probably lacking. These could include soluble factors, extracellular matrix interactions, an appropriate three-dimensional lattice, or specific cell-cell interactions. This idea is supported by our preliminary studies that show that when CD34–CD14+ cells are plated at high density in the appropriate medium, capillary like structures are formed. Also, while Moldovan et al. report a similar phenotype for CD14+ cell-derived EC, it is not identical to that reported here, but they use different culture conditions [20]. In vivo, exogenous human leukocytes enriched for CD34+ cells, but rarely CD34– cells incorporate into the
311
neovasculature, and exogenous CD34– cells have no effect on the rate of restoration of blood flow to the ischemic limb whereas CD34+ cells profoundly augment it [7, 8]. This has been taken to indicate that CD34+ but not CD34– cells contain angioblasts. Our in vivo data provide a surprising new interpretation of these findings. They suggest that CD34– cells act as angioblasts in vivo, and that CD34+ cells provide stimuli that induce CD34– cell incorporation into the vasculature either directly or by generally stimulating neovascularization. This idea is supported by the finding that hematopoietic stem cells are required for embryonic angiogenesis [32]. It might be argued that the intercalated cells are simply monocytes traversing the endothelium. However, the fact that the cells are elongated along the capillary longitudinal axis argues against this. Also, monocyte trafficking would be expected to occur regardless of whether CD34+ cells are co-injected. ACKNOWLEDGMENT The authors wish to thank Dr. Robert Tomanek for useful discussions, Justin Fishbaugh for help collecting and analyzing FACS data, and our blood donors. Confocal microscopy was performed at the Central Microscopy Research Facility, FACS analysis at the Flow Cytometry Facility, and DNA sequencing at the DNA Facility of the University of Iowa. Supported by a joint NIH and Juvenile Diabetes Foundation International grant #DK5596 and by an American Heart Association grant #0051372Z.
R EFERENCES 1 Crosby JR, Kaminski WE, Schatteman GC et al. Endothelial cells of hematopoietic origin make a significant contribution to adult blood vessel formation. Circ Res 2000;87:728-730.
8 Schatteman GC, Hanlon HD, Jiao C et al. Blood-derived angioblasts accelerate blood-flow restoration in diabetic mice. J Clin Invest 2000;106:571-578.
2 Asahara T, Masuda H, Takahashi T et al. Bone marrow origin of endothelial progenitor cells responsible for postnatal vasculogenesis in physiological and pathological neovascularization. Circ Res 1999;85:221-228.
9 Goodell MA, Brose K, Paradis G et al. Isolation and functional properties of murine hematopoietic stem cells that are replicating in vivo. J Exp Med 1996;183:1797-1806.
3 Asahara T, Murohara T, Sullivan A et al. Isolation of putative progenitor endothelial cells for angiogenesis. Science 1997;275:964-967. 4 Shi Q, Rafii S, Wu MH et al. Evidence for circulating bone marrow-derived endothelial cells. Blood 1998;92:362-367. 5 Boyer M, Townsend LE, Vogel LM et al. Isolation of endothelial cells and their progenitor cells from human peripheral blood. J Vasc Surg 2000;31:181-189. 6 Hernandez DA, Townsend LE, Uzieblo MR et al. Human endothelial cell cultures from progenitor cells obtained by leukapheresis. Am Surg 2000;66:355-358; discussion 359. 7 Kalka C, Masuda H, Takahashi T et al. Transplantation of ex vivo expanded endothelial progenitor cells for therapeutic neovascularization. Proc Natl Acad Sci USA 2000;97:3422-3427.
10 Osawa M, Hanada K, Hamada H et al. Long-term lymphohematopoietic reconstitution by a single CD34– low/negative hematopoietic stem cell. Science 1996;273:242-245. 11 Nakamura Y, Ando K, Chargui J et al. Ex vivo generation of CD34(+) cells from CD34(–) hematopoietic cells. Blood 1999;94:4053-4059. 12 Eichmann A, Corbel C, Nataf V et al. Ligand-dependent development of the endothelial and hemopoietic lineages from embryonic mesodermal cells expressing vascular endothelial growth factor receptor 2. Proc Natl Acad Sci USA 1997;94:5141-5146. 13 Wagemaker G, Neelis KJ, Wognum AW. Surface markers and growth factor receptors of immature hemopoietic stem cell subsets. STEM CELLS 1995;13(suppl 1):165-171. 14 Demirer T, Bensinger WI. Optimization of peripheral blood stem cell collection. Curr Opin Hematol 1995;2:219-226.
312
CD34– Endothelial Cell Progenitors
15 Todd RF III, Nadler LM, Schlossman SF. Antigens on human monocytes identified by monoclonal antibodies. J Immunol 1981;126:1435-1442.
24 Kunjathoor VV, Wilson DL, LeBoeuf RC. Increased atherosclerosis in streptozotocin-induced diabetic mice. J Clin Invest 1996;97:1767-1773.
16 Shen H, Clauss M, Ryan J et al. Characterization of vascular permeability factor/vascular endothelial growth factor receptors on mononuclear phagocytes. Blood 1993;81:2767-2773.
25 Couffinhal T, Silver M, Zheng LP et al. Mouse model of angiogenesis. Am J Pathol 1998;152:1667-1679.
17 Waltenberger J, Lange J, Kranz A. Vascular endothelial growth factor-A-induced chemotaxis of monocytes is attenuated in patients with diabetes mellitus: a potential predictor for the individual capacity to develop collaterals. Circulation 2000;102:185-190. 18 Asahara T, Takahashi T, Masuda H et al. VEGF contributes to postnatal neovascularization by mobilizing bone marrow-derived endothelial progenitor cells. EMBO J 1999;18:3964-3972. 19 Fernandez-Pujol B, Lucibello FC, Gehling UM et al. Endothelial-like cells derived from human CD14 positive monocytes. Differentiation 2000;65:287-300. 20 Moldovan NI, Goldschmidt-Clermont PJ, Parker-Thornburg J et al. Contribution of monocytes/macrophages to compensatory neovascularization: the drilling of metalloelastase-positive tunnels in ischemic myocardium. Circ Res 2000;87:378-384. 21 Maniotis AJ, Folberg R, Hess A et al. Vascular channel formation by human melanoma cells in vivo and in vitro: vasculogenic mimicry. Am J Pathol 1999;155:739-752.
26 Schatteman GC, Motley ST, Effmann EL et al. Platelet-derived growth factor receptor alpha subunit deleted Patch mouse exhibits severe cardiovascular dysmorphogenesis. Teratology 1995;51:351-366. 27 St. Croix B, Rago C, Velculescu V et al. Genes expressed in human tumor endothelium. Science 2000;289:1197-1202. 28 Gougos A, Letarte M. Primary structure of endoglin, an RGDcontaining glycoprotein of human endothelial cells. J Biol Chem 1990;265:8361-8364. 29 Lopez-Casillas F, Cheifetz S, Doody J et al. Structure and expression of the membrane proteoglycan betaglycan, a component of the TGF-beta receptor system. Cell 1991;67:785-795. 30 O’Connell PJ, McKenzie A, Fisicaro N et al. Endoglin: a 180kD endothelial cell and macrophage restricted differentiation molecule. Clin Exp Immunol 1992;90:154-159.
22 Field EH, Noelle RJ, Rouse T et al. Evidence for excessive Th2 CD4+ subset activity in vivo. J Immunol 1993;151:48-59.
31 Coffin JD, Harrison J, Schwartz S et al. Angioblast differentiation and morphogenesis of the vascular endothelium in the mouse embryo. Dev Biol 1991;148:51-62.
23 Field EH, Rouse TM, Gao Q et al. Association between enhanced Th2/Th1 cytokine profile and donor T-cell chimerism following total lymphoid irradiation. Hum Immunol 1997;52:144-154.
32 Takakura N, Watanabe T, Suenobu S et al. Role for hematopoietic stem cells in promoting angiogenesis. Cell 2000;102:199-209.
