|Home | About | Journals | Submit | Contact Us | Français|
We have identified a novel hierarchy of human endothelial colony forming cells (ECFC), which are functionally defined by their proliferative and clonogenic potential and in vivo vessel forming ability. Utilizing previously established clonogenic assays for defining different subpopulations of human ECFC, we now show that a hierarchy of ECFC, identical to the human system, can be isolated from the peripheral blood of rhesus monkeys (Macaca mulatta), and that the frequency of the circulating cells varies with age. Endothelial cells derived from rhesus monkey ECFC share a cell surface phenotype similar to human cord blood ECFC, rapidly form capillary like-structures in vitro, and rhesus monkey endothelial lined vessels in vivo upon implantation in immunodeficient mice in an age-dependent manner. Of interest, while ECFC from the oldest monkeys formed capillary like-structures in vitro, the cells failed to form inosculating vessels when implanted in vivo, and displayed a deficiency in cytoplasmic vacuolation in vitro; a critical first step in vaculogenesis. Thus, these studies establish the rhesus monkey as an important preclinical model for evaluating the role and functions of circulating ECFC in models of primate vascular homeostasis and aging.
The paradigms for vascular formation and repair can be generally divided into the well recognized and extensively studied processes of vasculogenesis, angiogenesis, and arteriogenesis (1). However, postnatal vasculogenesis is a rather novel paradigm proposed by Asahara et al. (2) in which circulating progenitor cells for the endothelial lineage (endothelial progenitor cells, EPCs) are recruited to sites of vascular injury, ischemia, or inflammation and these precursors participate in the regeneration, repair, or vascular remodeling process. Numerous methods for identifying EPCs have been proposed, however, no unique or specific marker to prospectively isolate this cell from circulating blood have been reported. At present, EPCs can be isolated using flow cytometric and in vitro culture methods (1, 3, 4) and many investigators have correlated the circulating concentration of EPCs with the presence and/or severity of numerous cardiovascular disease states (5–7).
We have identified a novel hierarchy of circulating EPCs in human umbilical cord blood and adult peripheral blood using an in vitro colony forming assay (8, 9). Endothelial colony forming cells (ECFC) that circulate in the bloodstream, demonstrate robust clonal regenerative properties, display a wide variety of cell surface molecules commonly observed on human arterial and venous endothelial cells (ECs), and demonstrate self-renewal capacity and lineage restriction to only contribute to the endothelial lineage (8–10). The ECFC progeny form in vitro capillary-like structures and spontaneously form a capillary plexus in type 1 collagen/fibronectin gels upon implantation into immunodeficient mice (10). These transplanted human capillaries inosculate with nearby endogenous murine vessels to become part of the systemic circulation of mouse blood cells (11, 12). Thus, ECFC display all the properties one should expect in a circulating human EPC.
The outgrowth of ECs from the peripheral blood of other mammalian species has been reported, including the rhesus monkey (13). In mice, circulating ECFC are extremely rare and peripheral blood from more than 5 animals is required to assure the growth of at least a single colony (14). Circulating ECFC are also rare in porcine blood (1.5 colonies/10 mL), but the number and proliferative potential increases following an acute myocardial infarction (15). In some instances, ECFC have been identified in cultures of endothelial cells isolated from tissues or blood vessels (16–18). However, in most cases, the clonal proliferative potential of the ECFC has not been rigorously tested. This is a particularly interesting point, since significant differences in the proliferative potential are displayed by the ECFC derived from human umbilical cord and adult peripheral blood suggesting an age related change in ECFC function (8). Studies have suggested that the ECFC from cord blood may display greater in vivo vessel forming ability compared to cells isolated from adult peripheral blood (11).
We hypothesized that the rhesus monkey would be an excellent model to examine the changes in circulating concentrations and functions of circulating ECFC since this nonhuman primate possesses a reasonably long lifespan (approximately a 1:4 age ratio compared to human subjects) and has been used extensively to model age-related processes that occur in human subjects. Indeed, we report that the circulating concentration of ECFC changes with age, that the proliferative potential of individual ECFC progeny declines with age, and that the vessel forming ability of ECFC progeny also declines with age in rhesus monkeys. Given the similar proliferative kinetics, circulating frequency, cell surface phenotype, and in vivo vessel forming ability of young rhesus ECFC to human umbilical cord blood ECFC, we propose that the rhesus monkey provides an invaluable model system to examine the role of ECFC cell therapy to treat human cardiovascular and related disease states.
Blood samples (5–40 ml) were collected from 40 healthy rhesus monkeys from birth to approximately 24 years of age. The Institutional Animal Care and Use Committee (IACUC) at the University of California, Davis approved all protocols for blood sample collection.
Rhesus monkey low density mononuclear cells (MNC) were obtained as previously described with minor modifications (19). Blood was diluted 1:1 with Hanks Balanced Salt Solution (HBSS) (Invitrogen, Grand Island, NY) and overlayed onto an equivalent volume of Histopaque 1077 (ICN, Costa Mesa, CA). Cells were centrifuged for 30 minutes at room temperature at 740 × g. MNCs were isolated and washed three times with Endothelial Cell Growth Medium-2 (EGM-2) medium (Lonza, Walkersville, MD) supplemented with 10% fetal bovine serum (Hyclone, Logan, UT), 2% penicillin/streptomyocin (Invitrogen) and 0.25 μg/ml of amphotericin B (Invitrogen) (complete EGM-2 medium).
MNC were seeded at 3 – 5 × 107 cells per well in 4 ml of complete EGM-2 medium in six well tissue culture plates pre-coated with type I rat tail collagen (BD Biosciences, Bedford MA) at 37°C, 5% CO2 in a humidified incubator. After 24 hours of culture, the non-adherent cells and debris were aspirated, the adherent cells were washed one time with complete EGM-2 medium, and then complete EGM-2 medium was added to each well. Medium was changed daily for the first seven days and then every other day until the first passage.
Colonies of ECs appeared between 7–20 days of culture and were identified as well circumscribed adherent colonies of cobblestone appearing cells. The EC colonies derived from ECFC were enumerated by visual inspection with an inverted microscope (Olympus, Lake Success, NY) under 40x magnification. The ECs derived from the colonies were released from the original tissue culture plates utilizing trypsin EDTA (Invitrogen), resuspended in complete EGM-2 media, and plated onto 75 cm2 tissue culture flasks coated with type I rat tail collagen for further passage.
To assess the ability of ECs to incorporate Ac-LDL, attached cells were incubated with 10 μg/ml of 488-Ac-LDL (Invitrogen) in the medium for 4 hours at 37°C. Cells were washed three times, stained with 1.5 μg/mL of 4′, 6-diamidino-2-phenylindole dihydrochloride (DAPI) (Sigma), and examined by inspection through an inverted fluorescence microscope (Zeiss, Thornwood, NY) at 100x magnification.
Matrigel assays were performed as previously described (8). Early passage (2–3) ECFC-derived ECs were seeded onto 96 well tissue culture plates coated with 30 μl of Matrigel (BD Biosciences) at a cell density of 5,000–20,000 cells per well. Cells were observed every two hours by visual microscopy with an inverted microscope at 40x magnification for capillary-like structure formation. The total length of capillary like structures was quantitated using Image J software as previously described (8).
Early passage (1–2) ECFC-derived ECs (5 × 105) were incubated at 4°C for 30 min with varying concentrations of the primary or isotype control antibody as outlined below in 100 μl of PBS and 2% fetal bovine serum (FBS), washed three times, and analyzed by fluorescence activated cell sorting (FACS®) (Becton Dickinson, San Diego, CA). Primary murine monoclonal antibodies were used against human CD31 conjugated to FITC (BD Pharmingen; clone WM59), anti-human CD34 conjugated to phycoerythrin (PE) (BD Pharmigen; clone 563), anti-human VEGFR2 conjugated to FITC (BD Pharmingen; clone 89106), anti-human CD14 conjugated to PE (BD Pharmingen; clone MrE2), anti-human CD45 conjugated to FITC (BD Pharmingen; clone HI30), anti-human CD146 conjugated to PE (BD Pharmingen; clone P1H12), anti-human AC-133 conjugated to PE (BD Pharmigen; clone JM3), anti-human CD105 conjugated to PE (BD Pharmingen; clone 266), and anti-human CD144 (BD Pharmingen; clone 55-7H1) conjugated to Alexa Fluor 647 (Molecular Probes, Eugene, OR). Isotype control antibodies of the IgG1 and IgG2A subclass were used. All antibodies have been previously validated to cross-react with appropriate cell subsets in the rhesus monkey as indicated in the rhesus resource (http://nhpreagents.bidmc.harvard.edu/nhp/ReagentList.aspx). For detection of the ability to bind to Ulex europaeus agglutinin-I (UEA-I), rhesus ECs were stained with FITC conjugated UEA-1 (Vector Laboratories, Burlingame, CA).
For early passage (2–4) ECFC-derived ECs, the FACS Vantage Sorter (BD) was used to place one cell per well into 96 well plates pre-coated with type I rat tail collagen in 200 μl of complete EGM-2 medium. Individual wells were examined under a fluorescence microscope at 50x magnification to ensure that only one cell had been placed into each well. Cells were cultured at 37°C, 5% CO2 in a humidified incubator. Media were changed every five days. After 14 days of culture, the cells were fixed with 4% paraformaldehyde (Sigma, St. Louis, MO) in PBS for 30 min at room temperature, then washed twice, stained with 1.5 μg/ml DAPI, and examined for the growth of ECs. Those wells containing two or more cells were identified as positive for proliferation under a fluorescent microscope at 10x magnification. Wells containing fewer than 50 cells were counted by visual inspection with a fluorescent microscope at 10x magnification. For those wells with more than 50 cells, colonies were imaged and cell number quantified using an Image J 1.36v program (Wayne Rasband, NIH).
Early passage (3–5) ECFC derived ECs (2×106 cells/ml) were suspended in a 1.5 mg/ml collagen-fibronectin matrix as previously described (10). Volumes (1 ml) were pipette into wells of 12 well plates, allowed to polymerize at 37°C for 30 min, and covered with complete EGM-2 medium for overnight incubation at 37°C, 5% CO2. After 18 hrs of ex vivo culture, cellularized matrices were bisected and implanted into the flanks of 6–8 week-old NOD/SCID mice as previously described (10). After 14 days mice were euthanized and the grafts were harvested, fixed in formalin free zinc fixative (BD Biosciences), paraffin embedded, bisected, and sectioned (6 μm) for analysis by immunohistochemistry (N=6).
Sections were stained as previously described (11). Briefly paraffin embedded tissue sections were deparaffinized and immersed in retrieval solution (Dako, Carpenteria, CA) for 20 min at 90–99°C. Slides were incubated at room temperature for 30 min with anti-human CD31 (clone JC70/A, Abcam), followed by 10 min incubation with LASB2 link-biotin and streptavidin-HRP (Dako), then developed with DAB (Vector, Burlingame, CA) solution for 5 min.
ECFCs were suspended in collagen solutions at a cell density of 2 million cells/ml prior to polymerization to ensure a uniform distribution throughout the type I collagen matrix. The matrix was polymerized for 30 minutes at 37°C and then 120 μL of warm EGM-2 (Lonza, Basel, Switzerland) media supplemented with 50 ng/ml 12-O-tetradecanoyl-phorbol-13-acetate (TPA) (Sigma Aldrich) was added. In vitro cultures were maintained for 48 hour, fixed with 4% paraformaldehyde, and then stained with 0.1% toluidine blue O in 30% methanol. Bright field microscopy was used to visualize the ECFC derived vascular structures. Nine fields of view at 40X were analyzed for both the 1–4 year Rhesus and > 18 year Rhesus ECFCs using a standard image analysis system Metamorph (Molecular Devices, Sunnyvale, CA, California). Total vacuole area was determined by summation of all vacuole areas scored.
Results are generally depicted as the mean ± the standard error of the mean (SEM). Data were analyzed with ANOVA and significant differences were set at the P<0.05 level using GraphPad InStat software. To test if there was a relationship between the age of monkey from which the ECFC were derived and proliferative potential, simple linear regression was conducted using SAS statistical software (SAS Institute, Inc., Cary, NC). A p-value of less than 0.05 was considered statistically significant.
ECFC have been isolated as EC colonies and expanded ex vivo from human adult peripheral and umbilical cord blood MNC (8). Further, we determined that a hierarchy of ECFC exists that can be discriminated by the clonogenic and proliferative potential of individual cells. To determine whether ECFC could also be isolated from the rhesus monkey, we harvested MNCs from peripheral blood and daily examined the culture wells for colony formation. ECFC-derived EC colonies emerged from plated MNCs 7–20 days after initiation of the cultures and the overall average frequency was 24±4 colonies per 108 MNCs (N=28). Among all the subjects tested, 12 animals in the aged group (≥ 18 years) failed to give rise to any colony growth in vitro, whereas colonies were present in every other sample from the blood of the 28 evaluable subjects. However, differences in the total number of colonies obtained were observed in the age groups examined that included young animals birth-<1 yr; median 3 and range 2–25 with a frequency of 1.50 colonies/mL of blood (n = 8) and 1-<4 yrs; median 8 and range 1–31 with a frequency of 3.10 colonies/mL of blood (n = 8), adult animals 4-<18 median 10 and range 2–90 with a frequency of 0.52 colonies/mL of blood (n = 8), and aged monkeys ≥18 yrs; median 2 and range 2–4 with a frequency of 0.06 colonies/mL of blood (n = 4). The male to female ratio of 2:1 was present in all age groups. Thus, circulating ECFC concentration increased with age, reaching a peak in young animals with a significant decline between the 4-<18 yrs and ≥18 yrs (p<0.05). All ECFC-derived EC colonies were able to expand and form a monolayer of EC with a cobblestone morphology (Fig. 1A). In general, ECFC emerged in culture earlier (7–14 days) in blood samples from younger animals (<4 years) and later (12–20 days) in blood samples from the aged animals (data not shown).
Before testing the proliferative and clonogenic potential of ECFC-derived ECs, we verified that the cell progeny derived from rhesus peripheral blood was not contaminated with other cell types. Immunophenotyping revealed that rhesus ECFC-derived cells expressed a very similar pattern of cell surface antigens as human ECFC-derived ECs (8, 10) including CD31, CD105, CD144, CD146, and Ulex europeus lectin binding (Fig. 1B). The ECs also weakly expressed CD34 and VEGFR2, but failed to express the hematopoietic antigens CD14, CD45, or CD133 (Fig. 1B). We have also observed that mesenchymal cells contaminating the ECFC colonies could be readily identified by lack of CD31, CD141, and CD144 expression, lower CD105 expression, and greater CD90 expression (data not shown). When the overall percentage of rhesus ECFC-derived ECs were analyzed for cell surface expression in the different age groups, we saw no significant differences in the cell surface antigen expression in monkeys from birth to adulthood (data not shown).
Moreover, rhesus monkey ECFC-derived ECs uniformly ingested 488-Ac-LDL and formed capillary-like structures in Matrigel, which are characteristics of ECs (Fig. 1C–D). In particular, when we quantitatively analyzed the ability of ex vivo tube formation of these ECs we found there was no statistical difference among the age groups (data not shown). Thus, these studies confirmed that the cell progeny derived from rhesus circulating ECFC were endothelial in origin and not contaminated with hematopoietic or mesenchymal cells.
We have developed a single cell colony-forming assay to quantitate the clonogenic potential of individual human circulating ECFC derived ECs (8). A complete hierarchy of human peripheral blood, cord blood, or vessel wall-derived ECFC has been identified which is composed of high proliferative potential ECFC (HPP-ECFC), low proliferative potential ECFC (LPP-ECFC), endothelial cluster cells, and non-dividing mature ECs (8, 9). Additionally, human ECFC-derived ECs from umbilical cord blood displayed greater proliferative potential than those from adult peripheral blood suggesting an age related change in ECFC function. By using the same experimental method we examined whether there was a hierarchy of proliferative potential in ECFC present in rhesus peripheral blood and whether the distribution of proliferative potential could be altered with age.
It was apparent that plating single circulating ECFC-derived ECs resulted in a diverse display of differences in proliferative ability. Most single plated EC did not divide but significantly more single cells divided in the younger animals (birth-1 yr and 1-<4 yrs) than in the older animals (4-<18 yrs) (Fig 2). Of the single cells that executed a division, heterogeneity in the proliferative potential was observed that varied with the age of subjects (Fig. 2). The progeny of some single plated ECs were able to give rise to more than 2,000 progeny and were capable of being replated to form secondary and tertiary colonies of the same size, thus demonstrating the properties of derivation from a HPP-ECFC (data not shown) as previously described (8). Therefore, a complete hierarchy of ECFC is present in the peripheral blood in rhesus monkeys as in human subjects. Noticeably, when comparing birth-1 year-old animals to 4-<18 year-old animals, there were significantly more plated single cells in 4-<18 year-old animals that divided but gave rise to small sized colonies containing less than 50 progeny (birth-1 yr; 42±7% vs. 4-<18 yrs; 84±13%). In contrast, the percentage of single cells giving rise to colonies of more than 10,000 cells (HPP-ECFC) dramatically decreased in the adult and aged animals (Figure 2B). Thus, the distribution of ECFC was skewed toward HPP-ECFC in the newborn and young animals while weighted towards minimally proliferative endothelial cluster forming cells or LPP-ECFC in the adult and more mature population. Simple linear regression analysis was conducted to determine if there was a relationship between age of the monkey from which the ECFC were derived and proliferative potential. The results demonstrated a statistically significant decrease in percent cells dividing, and percent of colonies containing greater than 10,000 cells with age (p < 0.05) and an increase in the percent of colonies containing both 2–50 and 51–200 cells (p < 0.05) with increasing age. This regression analysis suggests that the proliferative potential of ECFCs are exhausted as the animal progresses through its lifespan.
We (10) and others (11, 12, 20) have reported that human ECFC possess the potential to form de novo blood vessels when transplanted in a collagen-fibronectin matrix subcutaneously into immunodeficient mice. Further, cord blood derived ECFC have been shown to have a greater potential to form de novo blood vessels in vivo than ECFC derived from adult peripheral blood (11, 12) suggesting a decrease in ECFC function with age, although a detailed analysis throughout the human lifespan has never been conducted. By using the same methods as employed for the human studies, we investigated whether there was an age dependent difference in monkey derived ECFC vessel forming potential over the lifespan. While co-implantation of mesenchymal stromal cells or adipose stromal cells is known to enhance the in vivo vessel forming ability of human cord blood and adult peripheral blood ECFC (11, 21), we chose not to include these supportive cells in the scaffolds in the present studies to more directly and rigorously assess the vasculogenic potential of the rhesus ECFC in vivo.
ECFC derived from monkeys of varying ages were suspended in a collagen-fibronectin matrix and transplanted into immunodeficient mice. At 14 days mice were euthanized, the grafts were harvested and analyzed for chimeric blood vessel formation. Grafts were stained for anti-human CD31 antibodies (cross reacts with monkey but not mouse) to differentiate between host and donor vasculature. The grafts were examined for functional monkey-derived capillaries determined by the expression of CD31 and the presence of mouse red blood cells (RBCs) within the vessel. In all experiments ECFC derived from the younger animals (birth-1 yr and 1-<4 yrs) and adult (4-<18 yrs) were able to form functional capillaries (Fig. 3). However, ECFC derived from aged animals (≥18 yrs) failed to form any functional chimeric vessels, suggesting a loss of ECFC vessel formation ability with aging (not shown).
We have assessed in vivo vasculogenesis in the present study by implanting the rhesus ECFC in collagen scaffolds and implanting the gels into host immunodeficient mice in the absence of any co-implanted stromal supportive cells. To begin to address the mechanisms that led to the failure of the aged rhesus ECFC to form vessels in vivo, we have tested the ability of young and old rhesus ECFC to undergo cytoplasmic vacuolation in vitro, since it is this process that is known to represent the fundamental step in endothelial cell formation of lumenized structures in a 3 dimensional gel in vitro and in vivo (22, 23). Vacuole density and total vascular area were significantly increased in matrices seeded with rhesus ECFC derived from 1–4 year old monkeys compared to ECFCs derived from monkeys older than 18 years of age. There was also a trend toward an increase in average vacuole area in matrices seeded with ECFCs from young (1–4 year old) rhesus monkeys compared to aged animals. No difference in the level of apoptosis of the implanted ECFC within the gels among the study groups was noted (data not shown). While this in vitro assay only examines the first critical step in vasculogenesis, it has been determined that blocking this step in vasculogenesis is sufficient to prevent in vivo vessel formation in several vertebrate model systems (22, 23).
While EPCs have previously been isolated from rhesus monkey peripheral blood and vessels (13), their phenotype and function have not been completely identified. In the present study we demonstrated a complete hierarchy of ECFC exists in the peripheral blood of the rhesus monkey as previously demonstrated in human subjects (8, 9). Further, monkey derived ECFC display the potential to form functional vessels upon in vivo implantation in NOD/SCID mice.
Human ECFC frequency and blood vessel forming potential has been shown to decrease in adult peripheral blood when compared to umbilical cord blood derived ECFCs (11, 20), suggesting an effect of aging on ECFC frequency and function. To determine if a similar effect is seen in the rhesus monkey we evaluated the frequency, proliferative potential, and vessel forming capabilities in circulating ECFC over the rhesus monkey’s life span. The current study demonstrated that the circulating concentration of ECFC changes with age, and that the proliferative potential and the vessel forming ability of ECFC progeny declines with age. This is the first study to report such changes over the lifespan of an organism.
Rhesus ECFC display an EPC phenotype that is stable across all age groups. Similar to human derived ECFC the monkey ECFC can express endothelial specific markers CD31, CD105, C144, CD146, and VEGR-2 and can take up the plant lectin UEA-1. Importantly, rhesus monkey-derived ECFCs do not express hematopoietic cell markers such as CD45 and CD14. Further, rhesus ECFC displayed the ability to ingest Ac-LDL in culture and form capillary structures ex vivo on Matrigel. This is consistent with previous reports characterizing rhesus EPCs (13), but extends the prior observations to both younger and older animals.
While the ECFC phenotype was consistent across the life span of the monkeys, the frequency of circulating ECFC changed with age. The average frequency of all samples tested was 24±4 colonies per 100 million MNC which is consistent with prior published data (13). However, the frequency of ECFC colonies was greatest in young animals, remained somewhat stable in adult monkeys, and dramatically declined in the elderly animals. Whether similar changes occur in elderly human subjects has not been addressed.
Human ECFC derived from umbilical cord blood demonstrate a higher proliferative potential compared to adult peripheral blood derived ECFC. Furthermore, a single cell clonogenic assay demonstrated the distribution of ECFC in cord blood is skewed toward HPP-ECFC, and this was accompanied by an increase in telomerase activity in cord blood derived ECFC (8). The lack of adequate ECFC proliferation in colonies derived from the aged group of rhesus monkeys prohibited determination of telomerase activity in this study; however, identification of a significant defect in the ability of the ECFC derived from the aged monkeys to undergo cytoplasmic vacuolation highlights a potential functional target for more investigation to understand this age related defect. While Rhesus ECFC derived from both monkeys 1–4 years of age and greater than 18 years of age displayed the ability to form vacuoles in vitro, the ECFC derived from animals 1–4 years of age did so to a significantly greater degree. The increase in both vacuole density and total vacuole area suggests that the 1–4 year old rhesus monkey ECFC are more readily able to respond to the extracellular matrix cues necessary for endothelial cells to undergo vasculogenesis than are the ECFC derived from older animals. As vacuole formation is proposed to be an initiating step in vessel formation (22, 23), this reduced ability to undergo vacuolization in vitro is a potential reason for the inability of ECFC derived from older animals to form functional vessels in vivo.
The current study demonstrates that the change in proliferative capacity of rhesus monkey ECFC across the life span is consistent with results previously demonstrated in humans, that the aged monkeys display a defect in the initial steps of vasculogenesis and are unable to form vessels in vivo, and suggests that this may be an appealing model system to further study the effects of aging and other stressors such as cardiovascular diseases on the biology of human circulating cells that are fated to assist in vessel repair and regeneration.
Grant Support: These studies were supported by the NIH Center of Excellence in Translational Human Stem Cell Research (#P50 HL085036), the Primate Center base operating grant (#P51 RR00169), and the Riley Children’s Foundation.