Abstract
Putative endothelial cell (EC) progenitors or angioblasts were isolated from human peripheral blood by magnetic bead selection on the basis of cell surface antigen expression. In vitro, these cells differentiated into ECs. In animal models of ischemia, heterologous, homologous, and autologous EC progenitors incorporated into sites of active angiogenesis. These findings suggest that EC progenitors may be useful for augmenting collateral vessel growth to ischemic tissues (therapeutic angiogenesis) and for delivering anti- or pro-angiogenic agents, respectively, to sites of pathologic or utilitarian angiogenesis.
Postnatal neovascularization is thought to result exclusively from the proliferation, migration, and remodeling of fully differentiated ECs derived from preexisting blood vessels (1). This adult paradigm, referred to as angiogenesis, contrasts with vasculogenesis, the term applied to the formation of embryonic blood vessels from EC progenitors, or angioblasts (2).
Vasculogenesis begins as a cluster formation, or blood island, comprising angioblasts at the periphery and hematopoietic stem cells (HSCs) at the center (3). In addition to this spatial association, angioblasts and HSCs share certain antigenic determinants, including Flk-1, Tie-2, and CD34. Conceivably, then, these progenitor cells may derive from a common precursor (3, 4).
The demonstration that HSCs from peripheral blood can provide sustained hematopoietic recovery is inferential evidence for circulating stem cells (5). Here, we have investigated the hypothesis that peripheral blood contains cells that can differentiate into ECs (6). We exploited two antigens that are shared by angioblasts and HSCs to isolate putative angioblasts from the leukocyte fraction of peripheral blood. CD34 is expressed by all HSCs but is lost by hematopoietic cells as they differentiate (7). It is also expressed by many including most activated ECs in the adult (8). Flk-1, a receptor for vascular endothelial growth factor (VEGF) (9), is also expressed by both early HSCs and ECs but ceases to be expressed during hematopoietic differentiation (10, 11).
CD34-positive mononuclear blood cells (MBCD34+) were isolated from human peripheral blood by means of magnetic beads coated with antibody to CD34 (Dynal, Lake Success) (12). Fluorescence-activated cell sorting (FACS) analysis (13) indicated that 15.7 ± 3.3% of selected cells compared with <0.1% of the remaining cells expressed CD34. CD34-depleted cells (MBCD34−) were used as controls. An antibody to Flk-1 was used for magnetic bead selection of Flk-1-positive mononuclear blood cells (MBFlk1+); among MBFlk1+ cells, 20.0 ± 3.3% were Flk-1 positive.
The MBCD34+ and MBCD34− cells were plated separately (14) on tissue culture plastic, collagen type I, or fibronectin. When plated on tissue culture plastic or collagen at a density of 1 × 103 cells/mm2, a limited number of MBCD34+ attached, became spindle shaped, and proliferated for 4 weeks. A subset of MBCD34+ plated on fibronectin promptly attached and became spindle shaped within 3 days (Fig. 1A); the number of attaching cells (ATCD34+) in culture increased with time (probability P < 0.05, by analysis of variance) (Fig. 1B). Attached cells were observed only sporadically among MBCD34− cultures, including cells followed for up to 4 weeks on fibronectin-coated plates.
Attachment, cluster formation, and capillary network development by progenitor ECs in vitro. (A) Spindle-shaped attaching cells (ATCD34+) 7 days after plating MBCD34+ (50 cells/mm2) on fibronectin in standard medium (14). (B) Number of ATCD34+ cells 12 hours and 3 days after culture of MBCD34+ on plastic alone (CD34+/non), collagen coating (CD34+/Col), or fibronectin (CD34+/Fn), and MBCD34− on fibronectin (CD34−/Fn). Network formation (C) and cord-like structures (D) were observed 48 hours after plating coculture of MBCD34+, labeled with DiI, with unlabeled MBCD34− cells (ratio of 1:100) on fibronectin. At 12 hours after coculture, MBCD34+-derived cells had formed multiple clusters (E and F). After 5 days, uptake of acLDL-DiI was detected in ATCD34+ cells at the periphery but not the center of the cluster (G and H).
To confirm that the spindle-shaped cells were derived from CD34-positive cells, we labeled MBCD34+ cells with the fluorescent dye DiI and coplated them with unlabeled MBCD34− cells on fibronectin at an overall density of 5 × 103 cells/mm2; the ratio of the two cell types was identical to that of the original mononuclear cell population (1% MBCD34+, 99% MBCD34−). After 7 days, DiI-labeled cells derived from the MBCD34+ culture, which initially accounted for only 1% of the blood cells, accounted for 60.3 ± 4.7% of total attaching cells as analyzed by FACS. Coincubation with MBCD34− cells increased the proliferation rate to more than 10 times that of MBCD34+ plated alone. Cocultures of MBCD34+ and MBCD34− cells also showed enhanced MBCD34+ differentiation, including the formation of cellular networks and tube-like structures on fibronectin-coated plates (Fig. 1, C and D). These structures consisted principally of DiI-labeled MBCD34+-derived cells (Fig. 1D). Furthermore, within 12 hours of coculture, multiple clusters had formed (Fig. 1E) that contained mostly MBCD34+-derived cells (Fig. 1F). These clusters comprised round cells centrally and sprouts of spindle-shaped cells at the periphery. The appearance and organization of these clusters resembled that of blood island-like cell clusters observed in dissociated quail epiblast culture, which gave rise to ECs and vascular structures in vitro (3). ATCD34+ cells at the cluster periphery took up DiI-labeled acetylated low density lipoprotein (acLDL), whereas the round cells did not (Fig. 1, G and H); the latter detached from the cluster several days later. The MBFlk1+ cells behaved similarly.
To evaluate whether MBCD34+ cells progressed to an EC-like phenotype, we assayed them for the expression of leukocyte and EC markers. Freshly isolated MBCD34+ cells, ATCD34+ cells cultured on fibronectin for 7 days, and human umbilical vein endothelial cells (HUVECs) were incubated with fluorescent-labeled antibodies and analyzed by FACS (Fig. 2). Leukocyte common antigen CD45 was identified on 94.1% of freshly isolated cells but disappeared after 7 days of culture (Fig. 2). In freshly isolated MBCD34+ cells, 15.7 ± 3.3% were CD34+, 27.6 ± 4.3% were Flk-1+, and 10.8 ± 0.9% were CD34+,Flk-1+. Expression of CD34, CD31, Flk-1, Tie-2, and E selectin—all markers of the EC lineage (11, 15)—was greater in ATCD34+ cells after 7 days of culture than in freshly isolated MBCD34+ cells.
FACS analysis of freshly isolated MBCD34+ and ATCD34+ cells after 7 days in culture, and HUVECs. Cells were labeled with fluorescent antibodies to CD45 (DAKO, Carpinteria); CD34, CD31 (Biodesign); Flk-1, Tie-2 (Santa Cruz); and E selectin (DAKO). Similar results were obtained in three or more experiments. The shaded area of each box denotes negative antigen gate, and the white area denotes positive gate. Numbers are the mean ± SEM percentage of cells for all experiments determined by comparison with corresponding negative control labeling.
Additional analyses (16) of ATCD34+ cells after 7 days of culture showed limited (6.0 ± 2.4% cells) expression of CD68, a marker of the monocyte-macrophage lineage; positive immunostaining for factor VIII, ulex europaeus agglutinin-1 (UEA-1), CD31, endothelial constitutive nitric oxide synthase (ecNOS), and E selectin; and more than 80% uptake of DiI-labeled acLDL.
To confirm an EC-like phenotype of ATCD34+ cells, we documented expression of ecNOS, Flk-1/KDR (Flk-1 is also known as VEGFR-2 in mouse, and KDR is the human homolog of VEGFR-2), and CD31 mRNA at 7, 14, and 21 days by reverse transcription-polymerase chain reaction (RT-PCR) (Fig. 3A). Evidence for ecNOS and Flk-1/KDR in ATCD34+ cells was also demonstrated in a functional assay. Nitric oxide was produced in the cells in response to the EC-dependent agonist acetylcholine (Ach) and the EC-specific mitogen VEGF (Fig. 3B); the latter response also confirms that the cells express a functional Flk-1 receptor (17).
Progenitor ECs express ecNOS, Flk-1/KDR, and CD31 mRNA and release NO. (A) Complementary DNA (from 106 cells) was amplified by PCR (40 cycles) with paired primers (23) (B) NO release from ATCD34+ and ATCD34− cells cultured in six-well plates was measured as described (24). NO production was measured in a well with incremental doses of VEGF and Ach. HUVECs and bovine aortic ECs were used as positive controls, and human coronary smooth muscle cells (HCSMCs) as negative control. The values are means ± SEM of 10 measurements for each group.
To determine if MBCD34+ cells contribute to angiogenesis in vivo, we used mouse and rabbit models of hindlimb ischemia. For administration of human MBCD34+ cells, C57BL/6J × 129/SV background athymic nude mice were used to avoid potential graft-versus-host complications. Two days after creating unilateral hindlimb ischemia by excising one femoral artery, we injected mice with 5 × 105 DiI-labeled human MBCD34+ or MBCD34- cells into the tail vein. Histologic examination 1 to 6 weeks later revealed numerous (Fig. 4A) including proliferative (Fig. 4, C and D) DiI-labeled cells in the neovascularized ischemic hindlimb. Nearly all labeled cells appeared integrated into capillary vessel walls. In MBCD34+-injected mice, 13.4 ± 5.7% of all CD31-positive capillaries contained DiI- labeled cells, compared with 1.6 ± 0.8% in MBCD34--injected mice (18). By 6 weeks, DiI-labeled cells were clearly arranged into capillaries among preserved muscle structures (Fig. 4, I and J).
Heterologous (panels A to L), homologous (M), or autologous (panels N and O) EC progenitors incorporate into sites of angiogenesis in vivo. (A and B) DiI-labeled MBCD34+ (red, arrows) between skeletal myocytes (M), including necrotic (N) myocytes 1 week after injection; most are colabeled with CD31 (green, arrows). Note a preexisting artery (A), identified as CD31-positive, but DiI-negative. (C and D) Evidence of proliferative activity among several DiI-labeled MBCD34+-derived cells (red, arrows), indicated by coimmunostaining for antibody to Ki67 (Vector Lab, Burlingame, California) (green). Proliferative activity is also seen among DiI-negative, Ki67-positive capillary ECs (arrowheads); both cell types contribute to neovasculature. (E) DiI (red) and CD31 (green) in capillary ECs (arrows in E and F) between skeletal myocytes, photographed through a double filter 1 week after DiI-labeled MBCD34+ injection. (F) A single green filter shows CD31 (green) expression in DiI-labeled capillary ECs integrated into the capillary with native (DiI-negative, CD31-positive) ECs (arrowheads in E and F). (G) Immunostaining 1 week after MBCD34+ injection showing capillaries comprising DiI-labeled MBCD34+-derived cells expressing Tie-2 receptor (green). Several MBCD34+-derived cells (arrows) Tie-2 positive and integrated with some Tie-2-positive host capillary cells (arrowheads) identified by the absence of red fluorescence. (H) Phase-contrast photomicrograph of the same section shown in (G) indicates the corresponding DiI-labeled (arrows) and -unlabeled (arrowheads) capillary ECs. (I and J) Six weeks after administration, MBCD34+-derived cells (red, arrows) colabel for CD31 in capillaries between preserved skeletal myocytes (M). (K and L) One week after injection of MBCD34−, isolated MBCD34--derived cells (red, arrows) are observed between myocytes but do not express CD31. (M) Immunostaining of β-Gal in a tissue section harvested from ischemic muscle of C57BL/6J,129/SV mice 4 weeks after the administration of MBFlk-1+ isolated from transgenic mice constitutively expressing β-Gal. (Flk-1 cell isolation was used for selection of EC progenitors because of the lack of a suitable antibody to mouse CD34.) Cells overexpressing β-Gal (arrows) were incorporated into capillaries and small arteries; these cells were identified as ECs by anti-CD31 and BS-1 lectin (16). (N and O) Section of muscle harvested from rabbit ischemic hindlimb 4 weeks after administration of autologous MBCD34+ cells. Red fluorescence in (N) indicates localization of MBCD34+-derived cells in capillaries seen (arrows) in the phase-contrast photomicrograph in (O). Each scale bar is 50 μm.
No labeled cells were observed in the uninjured limbs of either MBCD34+- or MBCD34−-injected mice. DiI-labeled cells consistently colocalized with cells immunostained for CD31 (Fig. 4, B, F, and J), Tie-2 (Fig. 4G), and UEA-1 lectin (16). In contrast, in hindlimb sections from mice injected with MBCD34−, DiI-labeled cells were typically found in stroma near capillaries, but they did not form part of the vessel wall nor did they colocalize with cells that stained with antibodies to either UEA-1 or CD31 (Fig. 4, K and L).
In a second set of mouse experiments, 1 × 104 MBFlk1+ cells were isolated from whole blood of 10 transgenic mice constitutively overexpressing β-galactosidase (β-Gal) (all mice were Flk-1+/+). MBFlk1+ or MBFlk1− cells were injected into nontransgenic mice of the same genetic background that had hindlimb ischemia of 2 days duration. Immunostaining of ischemic tissue, harvested 4 weeks after injection, for β-Gal demonstrated incorporation of cells expressing β-Gal in capillaries and small arteries (Fig. 4M); these cells were identified as ECs by staining with antibody to CD31 (anti-CD31) and BS-1 lectin.
In vivo incorporation of autologous MBCD34+ cells into foci of neovascularization was also tested in a rabbit model of unilateral hindlimb ischemia. MBCD34+ cells were isolated from 20 ml of blood obtained by direct venipuncture of normal New Zealand White rabbits immediately before surgical induction of unilateral hindlimb ischemia (19). Immediately after surgery, freshly isolated autologous DiI-labeled MBCD34+ were reinjected into the ear vein of the same rabbit. Histologic examination of the ischemic limbs 4 weeks later revealed that DiI-labeled cells were localized exclusively to neovascular zones of the ischemic limb (Fig. 4, N and O) and were incorporated into 9.7 ± 4.5% of the capillaries that consistently expressed CD31 and reacted with BS-1 lectin.
In summary, our findings suggest that cells isolated with anti-CD34 or anti-Flk-1 can differentiate into ECs in vitro. The in vivo results suggest that circulating MBCD34+ or MBFlk1+ cells may contribute to neoangiogenesis in adult species, consistent with vasculogenesis, a paradigm otherwise restricted to embryogenesis (2, 3). A potentially limiting factor in strategies designed to promote neovascularization of ischemic tissues (20) is the resident population of ECs that is competent to respond to administered angiogenic cytokines (21). This issue may be successfully addressed with autologous EC transplants. The fact that progenitor ECs home to foci of angiogenesis suggests potential utility as autologous vectors for gene therapy. For anti-neoplastic therapies, MBCD34+ cells could be transfected with or coupled to antitumor drugs or angiogenesis inhibitors. For treatment of regional ischemia, angiogenesis could be amplified by transfection of MBCD34+ cells to achieve constitutive expression of angiogenic cytokines or provisional matrix proteins or both (22).
