Endothelial progenitor cells (EPCs) have been isolated postnatally from bone marrow, blood, and both the intima and adventitia of conduit vessels. However, it is unknown whether EPCs can be isolated from the lung microcirculation. Thus we sought to determine whether the microvasculature possesses EPCs capable of de novo vasculogenesis. Rat pulmonary artery (PAEC) and microvascular (PMVEC) endothelial cells were isolated and selected by using a single-cell clonogenic assay. Whereas the majority of PAECs (∼60%) were fully differentiated, the majority of PMVECs (∼75%) divided, with ∼50% of the single cells giving rise to large colonies (>2,000 cells/colony). These highly proliferative cells exhibited the capacity to reconstitute the entire proliferative hierarchy of PMVECs, unveiling the existence of resident microvascular endothelial progenitor cells (RMEPCs). RMEPCs expressed endothelial cell markers (CD31, CD144, endothelial nitric oxide synthase, and von Willenbrand factor) and progenitor cell antigens (CD34 and CD309) but did not express the leukocyte marker CD45. Consistent with their origin, RMEPCs interacted with Griffonia simplicifolia and displayed restrictive barrier properties. In vitro and in vivo Matrigel assays revealed that RMEPCs possess vasculogenic capacity, forming ultrastructurally normal de novo vessels. Thus the pulmonary microcirculation is enriched with EPCs that display vasculogenic competence while maintaining functional endothelial microvascular specificity.
- endothelial progenitor cells
- pulmonary circulation
endothelium forms a contiguous interface between blood and tissue and is responsible for the regulated delivery of hormones, vasoactive autacoids, anti- and proproliferative signals, inflammatory mediators, metabolic products, and circulating cells to appropriate targets (1, 2, 12). Although adult (postnatal) endothelium does not actively proliferate in the intact mature blood vessel, severely injured or senesced endothelial cells must be continually replenished to maintain vascular and tissue/organ homeostasis (1, 2, 14). Schwartz and Benditt (42) revealed the presence of fast-growing endothelial “niches” within the aortic intima that were enriched with replication-competent cells needed to repopulate injured or senesced cells. Their results, together with recent findings in the intact circulation, suggest that 1–3 yr approximates the endothelial cell turnover in conduit blood vessels in vivo (1, 12).
However, conduit vessel endothelial cell behavior is not indicative of microvascular endothelial cell behavior. Endothelial cell heterogeneity exists along the arterial-capillary-venule axis of the same organ (1, 2). The literature is bereft of accurate estimates for the lifespan of microvascular endothelial cells in mature capillaries. Yet normal capillary growth can achieve extraordinarily high rates, as occurs with capillaries in the internal theca that cyclically proliferate to supply blood to the avascular granulosa during the periovulatory period (40). Similarly, the estimated rate of microvascular endothelial cell proliferation in the endometrium during the luteal phase of the menstrual cycle can be comparable with that observed in aggressive tumors (15). Exercise triggers rapid capillary growth into metabolically active muscle (20), and overfeeding-induced adipocyte accumulation triggers rapid capillary growth into fat (46). These findings support the notion that microvascular endothelial cells possess an intrinsically high proliferation rate and angiogenic/vasculogenic potential that is essential to their physiological function.
Microvascular endothelial cells retain their proproliferative phenotype in vitro. Comparison of the pulmonary artery (PAEC) and pulmonary microvascular (PMVEC) endothelial cell growth revealed that PMVEC populations grow approximately two times faster than do PAEC populations (27). Moreover, PMVECs possess increased cyclin D1 expression, cyclin-dependent kinase (CDK4) and CDK2 activity (promitogenic proteins), and a constitutive inactivation of retinoblastoma (antimitogenic protein) (44). However, it is not evident whether all PMVECs exhibit enhanced proliferative potential or whether, as suggested by the in vivo findings of Schwartz and Benditt (42), PMVECs are enriched with replication-competent cells that account for such high proliferative rates. Recently, Ingram (23) and colleagues demonstrated that populations of human umbilical vein and aortic endothelial cells possess a modest number of cells that underlie monolayer growth potential and display endothelial colony-forming cell (ECFC) activity. In their studies, ∼50% of cells seeded at a single cell density in a clonogenic assay were replication competent, with only ∼12.5% exhibiting high proliferative behavior. Upon reseeding at a single cell density, fast-growing cells reconstituted the entire hierarchy of ECFC growth potentials, giving rise to ECFC with high proliferative potential (self-renew). On the basis of their high proliferative index and capacity to self-renew, fast-growing ECFC were considered to be endothelial progenitors (21–23). Importantly, the isolated progenitor cells displayed an intrinsic vasculogenic capacity, as evidenced by their ability to form de novo vessels in vivo, an important functional endothelial attribute. These findings are compatible with those of Schwartz and Benditt (42) in endothelium of the intact vessel wall and indicate that endothelial cell populations derived from conduit vessels are enriched with progenitor niches. Because pulmonary microvascular endothelium possesses a larger surface area than any other vascular bed and exhibits an enhanced growth potential, we sought to determine whether PMVEC populations are enriched in progenitor cells and furthermore whether these fast-growing resident microvascular endothelial progenitor cells (RMEPCs) possess an intrinsically rapid vasculogenic capacity.
All studies involving the use of animals were approved by our Institutional Animal Care and Use Committee and conformed to the National Institutes of Health guidelines for the Care and Use of Laboratory Animals. CD40 rats (Charles River Laboratories) 10–15 wk old were used in the present study.
Isolation and Culture of Rat Pulmonary Endothelial Cells
PAECs and PMVECs were isolated and cultured as previously described (27). Briefly, anesthetized (sodium pentobarbital, 50 mg/kg ip) CD40 rats were subjected to thoracotomy, and the heart and lung were excised en bloc. For PAECs, the main pulmonary artery was dissected from the root to a second vessel generation and under sterile conditions placed in a 60-mm dish containing ice-cold DMEM (GIBCO). The vessels were inverted, the intima was scraped, and the collected cells were strained with a 20 μm filter (BD Biosciences). Harvested cells were transferred to a T75 flask and were supplemented with DMEM enriched with 20% FBS (Hyclone) and 100 U/ml penicillin-100 μg/ml streptomycin (GIBCO) and were incubated at 37°C with 5% CO2-21% O2. For PMVECs, the outer edges of the lung were sliced, and the subjacent tissue was carefully dissected and placed in a 60-mm dish containing cold DMEM (4°C). Tissue was digested with type II collagenase (Worthington), rinsed with DMEM, transferred to a T75 flask, and incubated as described for PAECs. PAECs and PMVECs were characterized by flow cytometry after cell expansion (passages 3–7) by using a previously described multipanel of cell surface and cytosolic markers (27). Cell-culture medium was replaced with DMEM enriched with 10% FBS and 1% penicillin/streptomycin (standard culture media), and passages 7–12 were used for consecutive experiments.
PAECs and PMVECs cells were seeded at 2 × 104 cells/cm2 in separate six-well plates (∼19,200 cells/well) containing DMEM and 1% penicillin/streptomycin. Following 48-h incubation at 37°C with 5% CO2-21% O2, the medium was replaced by standard culture media (serum-stimulated cell growth protocol) or serum-free DMEM (serum-free cell growth protocol); this time point was considered day 0. Growth studies were performed in triplicate, where three separate measurements were obtained per phenotype per day. Growth was determined by cell counts after day 0, at 24-h intervals, by using a Beckman Coulter counter. Results denote the average of triplicate measurements repeated three separate times. Statistical analysis compares the absolute numbers between groups at corresponding time points. Population doublings were calculated during the logarithmic growth phase by using Prism 4.03 GraphPad software.
Single-cell clonogenic assay.
The single-cell clonogenic assay was performed in expanding PAECs and PMVECs (60–80% confluent), as described elsewhere (22). Briefly, cells were trypsinized, strained using a 40-μm filter, and transferred to flow cytometry tubes containing standard culture medium at 1 × 106 cells/tube. Cells were sorted by a FACSVantage sorter (BD Biosciences) at a rate of 100 events/s. Individual-sorted cells were placed in collagen I (BD Biosciences)-coated wells of a 96-well plate containing 200 μl of standard culture medium. Each phenotype was seeded in triplicate. Cells were cultured at 37°C with 5% CO2-21% O2, and standard culture medium was replaced every 4 days. After 14 days, cells were fixed with 4% paraformaldehyde (Sigma) and were stained with Hoechst-33258 (Invitrogen). Each well was examined by fluorescent microcopy to detect the presence of fluorescently stained cells. Quantification was performed by visual inspection at ×40 magnification (<100 cells/well) or by imaging and counting them with ImageJ version 1.36 software (Wayne Rasband, National Institutes of Health; >100 cells/well). Results denote the average of triplicate measurements repeated three separate times. RMEPCs were obtained after expanding colonies that grow to >2,000 cells under standard culture media.
RMEPCs were suspended at 1 × 106 cells per tube in 1.7-ml microtubes containing 1 ml of cold PBS (calcium and magnesium free, pH 7.4, 4°C, GIBCO) supplemented with 2 mmol/l EDTA (Sigma). Cells for antibody staining were incubated with 0.05% Tween 20 (in PBS; Sigma) for 15 min at room temperature. All cells were rinsed with cold PBS and were incubated with the corresponding fluorescent-conjugated primary antibody (1:10), lectin (1:20), or isotype control antibody (1:10) for 30 min at room temperature in darkness. Cells were rinsed twice in cold PBS and were incubated for 30 min with a secondary antibody when necessary (1:10). At the end of the primary (or secondary) antibody incubation, cells were transferred to flow cytometric tubes and taken for flow analysis by using a FACSVantage SE. We used the following fluorescent-conjugated primary antibodies: CD34-FITC (mouse; Miltenyi), endothelial nitric oxide synthase (eNOS)-FITC (rabbit; Santa Cruz), CD133-phycoerythrin (PE) (mouse; Miltenyi), CD45-FITC (mouse; Invitrogen), mouse IgG2-PE (isotype control; Miltenyi), and rabbit isotype-FITC (Zymed). For detection of eNOS, cells were incubated with the antibody in the presence of 0.15% Triton X-100 (Sigma). For lectin bindings, we used the PE-conjugated Helix pomatia (EY Laboratories) and the FITC-conjugated Griffonia simplicifolia (Sigma). We used the nonfluorescent primary antibodies CD31 (goat; Santa Cruz), vascular endothelial cadherin (VE-cadherin; mouse; Santa Cruz), von Willebrand factor (sheep; Affinity Biologicals), neuronal cadherin (N-cadherin; mouse; Santa Cruz), vascular endothelial growth factor receptor-2 (VEGFR-2; goat; Abcam), and CD105 (rabbit; Santa Cruz). For immunodetection of von Willebrand factor, cells were incubated with the antibody in the presence of 0.15% Triton X-100. We used the following conjugated secondary antibodies: donkey anti-goat-FITC (Santa Cruz), goat anti mouse-FITC (Santa Cruz), donkey anti sheep-FITC (Jackson ImmunoResearch), and donkey anti rabbit Alexa 488 (Molecular Probes). Cells were incubated as a control with secondary antibodies exclusively to test their immunospecificity (negative controls). Confirmation of protein expression or lectin interaction in RMEPCs is observed at the flow charts by a right shift in the curve (dark areas) compared with that from a corresponding control (open areas).
Transelectrical Membrane Resistance
Barrier properties were assessed by using the electrical cell impedance sensor (ECIS) technique (29). Equivalent density of PAECs, PMVECs, and RMEPCs (5 × 104 cells/cm2) were seeded on gold electrode (8W10E, with 10 electrodes, each 250 μm diameter) arrays (∼23,000 cells/array), supplemented with standard culture media (400 μl total volume/array), and incubated at 37°C with 5% CO2-21% O2. Cell culture medium was changed 4 days after seeding. Individual phenotypes were seeded in triplicate, and electrical resistances were measured on a daily basis by using an ECIS model 1600R (Applied BioPhysics) and were reported at day 5, the time at which a plateau in the resistances was observed. Results denote the average of triplicate measurements repeated five separate times.
Expanding PAECs, PMVECs, and RMEPCs (60–80% confluent) were transferred to 1.7-ml microtubes at 1 × 106 cells/tube. Cells were suspended in a hybridization solution containing a FITC-conjugated telomere probe or in a probe-free hybridization solution according to manufacturer directions (telomere assay kit; Dako Cytomation). Each phenotype was hybridized in duplicate. Following overnight hybridization at room temperature, cells were rinsed and incubated with propidium iodide to select cells during G0/G1 to normalize the data to equivalent genome loads. Cells were taken for flow analysis to detect the fluorescence intensity of the samples using a FACSVantage SE. Results denote measurements of the relative telomere length, which indicates the telomere fluorescence per chromosome/genome in the sample with respect to that in control (cells loaded without the fluorescent probe). Results denote the average of duplicate measurements repeated three separate times.
Network Formation: Matrigel in Vitro
PAECs, PMVECs, and RMEPCs were seeded at 2 × 104 cells/cm2 in Matrigel-coated (100 μl/cm2) 48-well plates (∼20,000 cells/well) supplemented with standard culture medium (400 μl total volume/well). Fibroblasts isolated from lung parenchyma were used as control under similar experimental conditions. The plates were incubated at 37°C with 5% CO2-21% O2. Each phenotype was seeded in triplicate, and medium was replaced 4 days after seeding. Wells were analyzed for network formation by using phase-contrast microscopy at 8, 24, and 48 h and 1 wk after seeding. Network formation was determined by counting the number of branches that connect two distant cells observed at ×10. Average from counts of three wells at ×10 were considered (n = 1). Results denote the average of measurements repeated three separate times.
Assessment of de Novo Vessel Formation: Matrigel in Vivo
PAECs, PMVECs, and RMEPCs were suspended in 1.7-ml microtubes containing 250 μl cold (4°C) standard culture medium at a density of 375,000 cells/tube. The cell-containing solution was mixed with 500 μl of unpolymerized Matrigel at 4°C (to maintain the mixture in a fluid phase). The 750-μl mixture solution was injected subcutaneously (left and right lumbar abdominal regions, two plugs per animal) into mildly sedated (ketamine 75 mg/kg ip) CD40 rats using a 23-gauge needle. The injected mixture polymerizes at body temperature and becomes a plug following subcutaneous contact (∼1 min). In control experiments, Matrigel-medium mixture was injected with no cells. At 4 and 10 days after injection, Matrigel plugs were excised from the abdominal wall of anesthetized animals (sodium pentobarbital, 50 mg/kg ip) and immersion fixed in 4% paraformaldehyde. Fixed specimens were dehydrated in ethanol, embedded in paraffin, cut in 5-μm sections, and stained with hematoxylin and eosin for light microscopy analysis. Stained sections were examined to count the total number of tubes containing red blood cells (blood vessels) within the gel. Blood vessels at the periphery of the gel in close proximity to the subjacent tissue or embedded within native tissue anywhere in the gel were not included for quantification. Results denote the average from six different plugs obtained from different animals.
Transmission Electron Microscopy From in Vitro Studies
PAECs, PMVECs, and RMEPCs were seeded at 4 × 104 cells/cm2 onto Matrigel-coated (100 μl/cm2) 0.4 μm polycarbonate membranes (Nunc, ∼20,000 cells/membrane) and supplemented with standard culture medium (200 μl total volume/well). Cells were incubated at 37°C with 5% CO2-21% O2 for 8, 24, and 48 h and 1 wk after seeding. Culture medium was replaced after 4 days. Each phenotype was loaded in triplicate. After each time point, cells were fixed in 3% glutaraldehyde (in cacodylate buffer). Specimens were postfixed with 1% osmium tetroxide, dehydrated with a graded alcohol series, and embedded in PolyBed 812 resin. Semithin sections (1 μm) were cut and stained with toluidine blue and were examined by light microscopy. Thin sections (80 nm) were cut with a diamond knife, stained with uranyl acetate, counterstained with Reynold's lead citrate, and examined by transmission electron microscopy (TEM) using a Philips CM 100 (FEI). All fixative and staining reagents were purchased from Polysciences. Images are representative of random micrographs obtained from three separate experiments.
TEM From in Vivo Studies
Matrigel plugs were excised 4 and 10 days after injection as described under Assessment of de Novo Vessel Formation: Matrigel in Vivo and were immersion fixed in 3% glutaraldehyde (in cacodylate buffer). Specimens were processed for TEM and were examined as described for the TEM in vitro studies. Images are representative of random micrographs obtained from six different plugs.
Green Fluorescent Protein Transfection
PAECs, PMVECs, and RMEPCs were seeded in a six-well plate, supplemented with standard culture medium, and incubated at 37°C with 5% CO2-21% O2. Following 70% confluence, the medium was removed and culture supernatant containing retroviral constructs, encoding green fluorescent protein (GFP) and a puromycin-resistant cassette, was added to the cells in the presence of Polybrene (permeabilizing agent; Sigma). Plates were centrifuged at 750 rpm for 30 min at room temperature to ensure homogenous distribution of the retrovirus within the plate. Standard culture medium was added to the plates, and cells were incubated as described for 48 h or until a confluent monolayer was observed. Successful transfected cells were selected with puromycin (PAEC, 5 μg/ml; PMVEC and RMEPC, 15 μg/ml; Sigma). Selection was considered complete after 3 days of antibiotic incubation, the time at which the medium was removed and cells were supplemented with fresh standard culture medium. Medium was changed every 4 days until cells were confluent and used for in vivo Matrigel injections.
Immunodetection of RMEPCs in Vivo
Excised Matrigel plugs were processed for immunostaining as described under Assessment of de Novo Vessel Formation: Matrigel in Vivo. Sections of 5 μm were deparaffinized in two changes of histological-graded xylenes (Fisher). Tissue was rehydrated in a graded ethanol series followed by antigen retrieval using sodium citrate buffer at 10 mM with 0.05% Tween 20, pH 6.0, at 70°C. Slides were cooled at room temperature and were permeabilized by using Triton X-100 (0.15%; Sigma). After consecutive rinsing in PBS (GIBCO), tissue was blocked with 5% goat serum (Jackson ImmunoResearch) supplemented with 5% bovine serum albumin (Sigma) and 0.5% glycine (Sigma) for 1 h. To detect expression of GFP cells, sections were incubated with rabbit anti-GFP antibody (Abcam, 1:200) or vehicle (negative control) overnight at room temperature, followed by Alexa-Fluor 647-conjugated donkey anti-rabbit antibody (Invitrogen, 1:200) for 60 min. DAPI reagent (Invitrogen) was added to stain nuclei, and the slides were rinsed in PBS, air-dried, and coverslipped with Gold Antifade mounting medium (Invitrogen). Immunohistofluorescence and immunohistochemistry slides were imaged by using a fluorescent microscope with a ×60 objective. Experiments were repeated by examining tissues from three different plugs.
PAECs, PMVECs, and RMEPCs were seeded at 2 × 104 cells/cm2 (∼19,200 cells/well) in multiple six-well plates and were supplemented with DMEM and 1% penicillin/streptomycin (serum-free media). Following 48 h incubation at 37°C with 5% CO2-21% O2, serum-starved cells were supplemented with standard culture medium and allowed to grow. Cells during expansion (60–75% confluent) or confluent (100%) were incubated in Krishan solution [a mixture containing propidium iodide (Sigma) and NP-40 (Sigma), a permeabilizing agent] for 4 h at 4°C. Cell-cycle profiles were determined by analysis of propidium iodide incorporation into the cell genome by flow cytometry using a FACSVantage SE. Results denote the average of three different measurements.
Anchorage-Independent Growth-Soft Agar
PAECs, PMVECs, RMEPCs, and MDA-MB-435 (a breast cancer cell line used as a control) were suspended at a density of 2 × 104 cells/cm2 in 0.35% agar (in standard culture medium; BD Biosciences) and seeded onto a bottom layer of 0.75% solid agar (in dH2O) of a six-well plate (∼19,200 cells/well). Cells were incubated at 37°C with 5% CO2-21% O2. After 10 days, plates were taken for phase-contrast microscopy imaging to examine cell growth. Efficiency of anchorage-independent growth was determined by the capacity of cells to form colonies (>50 cells) within the agar. Experiments were performed in duplicate three separate times.
GraphPad Prism 4.03 software was used for statistical data analysis and figure generation. Quantitative data are represented as means ± SE. Comparison among groups was made using one- or two-way ANOVA, as appropriate; the Tukey post hoc test was used to identify intergroup differences. P < 0.05 was considered statistically significant.
Pulmonary Endothelial Cells Exhibit Phenotypic Heterogeneity in Cell Growth
PAECs and PMVECs possessed a 48-h lag phase (Fig. 1). After 72 h, PMVECs exhibited ∼3.5-fold increase in cell growth for up to 1 wk (P = 0.007); confluence was achieved by 5 days. PMVEC and PAEC population doublings, calculated during the logarithmic growth phase, were at 39 and 58 h, respectively. Although the underlying mechanism(s) of such growth inequities are incompletely resolved, the data suggest that PMVEC possess an intrinsic proproliferative behavior.
Pulmonary Microcirculation is Enriched in Endothelial Progenitor Cells
Yoder and colleagues (21, 23, 52) have recently demonstrated that vessel wall-derived endothelial cells from several systemic conduit vessels contain endothelial progenitor cells (EPCs) based on their high proliferative potential and self-renewal capacity. Therefore, we sought to determine whether PMVECs possess progenitor cells that account for their increased growth potential. We performed the single-cell clonogenic assay, where endothelial cells are seeded at the single-cell level, and their growth was measured over a 2-wk period (22, 23). In this assay, >60% of PAECs remained as single differentiated cells (i.e., cells did not divide), whereas <30% of PMVECs displayed similar behavior (Fig. 2, P < 0.01). Of the dividing cells, there were no significant differences between the two phenotypes in the percent of single cells giving rise to clusters or colonies of <2,000 cells. However, >50% of PMVECs, and only 20% of PAECs, displayed high proliferative potential, because single cells gave rise to colonies of >2,000 cells (>P = 0.03). Moreover, individual analysis of colony size revealed that high-proliferative-potential PMVECs were capable of forming colonies consisting of 100,000 cells, whereas PAECs typically grew to ∼10,000 cells. Importantly, when cells isolated from colonies that proliferate to >2,000 cells were reseeded as single cells in the clonogenic assay, only fast-growing cells were capable of reconstituting the complete proliferative hierarchy of all cells (i.e., a single cell could give rise to cells that remain as single cells, proliferate to between 2 and 2,000 cells, or divide to >2,000 cells). It is significant that within PMVECs there is a population of cells that not only exhibit a high proliferative behavior but that also possess the ability to give rise to endothelial cells exhibiting high proliferative potential, an indication of their capability to self-renew the cell colony. The abundance of these RMEPCs within the PMVEC population provides an explanation for the accelerated growth that is exhibited by PMVECs compared with PAECs.
Characterization of RMEPCs
Expression of surface and cytosolic markers.
We sought to determine whether RMEPCs express endothelial cell, microvascular, and progenitor markers. “Classical” endothelial markers include CD31 (platelet endothelial cell adhesion molecule), VE-cadherin (CD144), eNOS, and von Willebrand factor, proteins that were all expressed in RMEPCs (Fig. 3A; Table 1). The adhesion protein N-cadherin (CDw325) (10) was also expressed in RMEPCs. N-cadherin is found in migrating and proliferating cells, particularly endothelial and bone marrow-derived stem cells from the embryo and adult tissues (10, 30). Lectin reactivity discriminates lung endothelial phenotypes in vivo and in vitro, because Helix pomatia interacts mainly with PAECs and Griffonia simplicifolia interacts exclusively with PMVECs (27). RMEPCs interacted with Griffonia simplicifolia but did not interact with Helix pomatia. Thus RMEPCs express global endothelial cell markers and display a microvascular phenotype.
We next sought to determine whether RMEPCs express markers of circulating EPCs. Because circulating EPCs were originally described by Asahara and colleagues (6, 7), investigators have tried to identify unique surface marker(s) for this phenotype. Concurrent expression of CD34 (mucosilain), VEGFR-2 (CD309), CD105 (endoglin), and CD133 (prominin-1) are conventionally accepted circulating EPC markers, whereas CD34, VEGFR-2, and CD45 (leukocyte common antigen), but not CD105 or CD133, define a population of resident conduit vessel wall EPCs (18, 26, 48, 52, 53). Similar to circulating and vessel wall-derived EPCs, RMEPCs express CD34 and VEGFR-2. RMEPCs were negative for CD133 but positive for CD105 (EPC marker associated with vascular repair) (49) and negative for CD45 (a pan-leukocyte antigen; Table 1). The findings demonstrate that, in contrast to circulating early outgrowth- or resident conduit vessel wall-derived EPCs, RMEPCs express conventional endothelial markers while retaining some markers uniquely found in circulating progenitor cells and lacking expression of leukocyte antigens.
Endothelial barrier properties.
Endothelium forms a semipermeable barrier that restricts fluid and solute flux between blood and the subjacent interstitium (47). Recent findings indicate that PMVECs possess an inherently tighter barrier function compared with PAECs, both in vivo and in vitro (25, 35). We measured transelectrical resistance (TER) to determine whether RMEPCs retain an intrinsically resistant barrier function. Both PMVECs and RMEPCs possessed a significantly higher TER than did PAECs (Fig. 3B, P < 0.001). Interestingly, RMEPCs possessed a higher TER than PMVECs (P < 0.05). These results further support the idea that RMEPCs maintain a microvascular phenotype and indicate that a restrictive barrier function is a quality that does not limit rapid proliferation per se.
Telomeres are portions of genetic material involved in stabilizing the chromosome ends (3, 5). Short telomeres have been associated with replicative senescence in vitro and loss of stem cell proliferative capacity in vivo, whereas long telomeres are prominently found in rapidly growing cells (3, 5). Thus we examined their length in PAECs, PMVECs, and RMEPCs by fluorescently labeling the telomeres. In this assay, the telomere length is established as a percent of the maximal (100%) response. RMEPCs possessed the longest telomere length among all phenotypes studied (Fig. 3C). The findings suggest that the rapid proliferation of RMEPCs is associated with the presence of longer telomeres, a typical feature found in several stem cells in vivo that is coupled with enhanced regenerative properties (3, 5, 33).
PMVECs and RMEPCs exhibit rapid vascular network formation.
Endothelial cells possess the intrinsic ability to form vascular networks on Matrigel in vitro. We therefore examined the capacity for PAECs, PMVECs, and RMEPCs to form networks in Matrigel-coated wells. Both PMVEC and RMEPCs formed elaborate networks as early as 8 and 24 h after seeding, whereas PAECs failed to generate extensive networks at these time points (Fig. 4, A and B, P < 0.001 and P < 0.01, respectively). Network formation reached a plateau in PMVECs and RMEPCs by 48 h (∼90 networks/cm2), whereas it reached a plateau at 1 wk in PAECs (∼60 networks/cm2). Fibroblasts were used as a control, and they did not display organized, mature networks at any time point (data not shown). Ultrastructural analysis of endothelial cell networks revealed that RMEPCs formed tubes with distinguishable lumens (Fig. 4C, left), as did PMVECs and PAECs (data not shown). In general, all phenotypes displayed mature junctional complexes, with sealed cell-cell borders (Fig. 4C, middle). Consistent with their high proliferative behavior, RMEPCs displayed abundant dense fibrillar components (Fig. 4C, right), nucleolar structures that contribute to early transcriptional events and folding and assembly of ribosomal RNA (19, 34). The data indicate that RMEPCs not only grow at a fast rate but also have the ability to rapidly form ultrastructurally normal vascular networks, as do PMVECs.
Pulmonary Microcirculation Possesses Rapid Vasculogenic Capacity
We examined whether RMEPCs possess the ability to rapidly form tubes and blood vessels in vivo, critical attributes that should be displayed by any cell denoted as an EPC. PAECs, PMVECs, and RMEPCs were mixed in Matrigel at equal densities and injected subcutaneously into the rat abdomen. Plugs were harvested at 4 and 10 days after seeding. After 4 days, blood vessels (tubes containing red blood cells) were found in all of the plugs that were seeded with cells but were not observed in cell-deficient plugs (Fig. 5A). Vessels resembled both capillaries, with a single layer of cells delineating a small lumen (<10 μm), and arterioles/venules, with two or three cell layers delineating a larger lumen (<50 μm). In control plugs (cell free), mesenchymal-like cells infiltrated the gel, but these cells failed to form lumens (Fig. 5A). Although there was a high number of vascular lumens (tubes) that formed in RMEPC-seeded plugs (data not shown), there was no difference in the number of blood vessels that formed in PAEC-, PMVEC-, and RMEPC-seeded plugs at this time point (Fig. 5B).
At 10 days after injection, plugs seeded with all cell phenotypes were enriched in numerous, well-organized vessels (Fig. 5C). PMVECs and RMEPCs formed twofold more blood vessels than did PAECs (Fig. 5D; P < 0.001). Control plugs (without cells) also possessed blood vessels (Fig. 5C), although there were threefold fewer compared with PMVEC- and RMEPC-seeded plugs. Vessels that formed de novo were anatomically similar to host vessels that were found near the plug-tissue border.
A diversity of cellular infiltrates was observed within the plugs (Fig. 5, A and C) and raised the question as to whether de novo vessels were derived from the injected cells or whether they formed after the recruitment of host EPCs. To address this issue, GFP-transfected RMEPCs were seeded in plugs and injected, and the location of injected cells was resolved by GFP immunodetection. GFP-positive cells (red) lined the vessel lumen (Fig. 5E), indicating they were responsible for vessel formation. GFP-negative cells (blue nuclei, no visible cytosol) were observed within the gel, suggesting that host cells were recruited into Matrigel but without directly contributing to lumen formation. A similar response was observed in plugs injected with GFP-transfected PAECs and PMVECs (data not shown). GFP staining was negative altogether in plugs that were injected without cells (data not shown).
Ultrastructural analysis of the de novo vessels arising from RMEPCs was examined by TEM at day 4 and day 10 after injection to determine the quality of the vessels that formed. Junctional complexes were observed at the interendothelial borders, suggesting tight cell-cell apposition consistent with a restrictive barrier (Fig. 5F). Similar characteristics were found in PAECs and PMVECs at 4 or 10 days after injection (data not shown), illustrating that all endothelial cell phenotypes generate functional and ultrastructurally normal blood vessels.
RMEPCs Display “Physiological” Growth Behavior
Rapidly growing cells may transform in culture and lose cell-cycle regulatory-suppressive genes, leading to loss of cell-cell contact inhibition and uncontrolled growth. Given the high proliferative behavior, the capacity for self-renewal, and the high vasculogenic capability exhibited by RMEPCs, we sought to determine whether these cells display a malignant phenotype. First, we examined whether RMEPC growth required serum. Unlike transformed cells, RMEPC growth was suppressed in serum-deprived medium (Fig. 6A). Serum restriction also inhibited PAEC and significantly attenuated PMVEC growth, similar to previous reports (27, 45). Second, we measured propidium iodide incorporation to determine whether cell-cell contact reduces the percent of cells in the S phase of the cell cycle. At subconfluence, ∼25–35% of PAECs, PMVECs, and RMEPCs were in S phase (Fig. 6B), whereas at confluence, only 5–15% of cells remained in S phase, indicating that confluence decreases cell growth (P = 0.01). Third, we measured anchorage-independent cell growth in 0.7% agar matrix (43). As a control, a breast cancer cell line (MDA-MB-435) was seeded in parallel. PAECs displayed little proliferative capacity, and only small clusters (2–25 cells) were found scattered through the field (Fig. 6C). PMVEC and RMEPC growth in agar generally reflected that seen with PAECs, although infrequently larger clusters were observed (25–50 cells), whereas breast cancer cells formed numerous, large colonies (>50 cells) throughout the field. Collectively, these findings support the idea that, although RMEPCs constitute a fast-growing cell type with the capacity for self-renewal, they do not display a malignant or transformed phenotype.
Microvascular endothelial cells possess an intrinsic ability to proliferate at high rates, both in vivo and in vitro, an essential attribute for vascular homeostasis and repair (9). In the present study, we sought to determine whether such proproliferative behavior was due to the presence of progenitor cells residing within the PMVEC population and to determine whether they exhibited a high vasculogenic capacity. Our results indicate that ∼50% of the cells comprising a PMVEC population proliferate rapidly and are able to self-renew the colony's hierarchy of growth potentials. These RMEPCs retain their endothelial cell phenotype and, more specifically, their microvascular phenotype, on the basis of both the conservation of surface antigens and functional behaviors, such as barrier resistance. RMEPCs are vasculogenic in in vitro and in vivo Matrigel assays. To date, these endothelial cells isolated from the lung microcirculation represent the most enriched source of EPCs possessing neovasculogenic capacity.
EPCs exist in a variety of sites in vivo, including bone marrow, blood, and the intima and vascular wall of conduit vessels (6, 7, 23, 37–39, 53). Asahara and colleagues (6, 7) described a source of circulating mononuclear cells that arose in culture soon after their isolation and formed cobblestone cell colonies. These cells expressed endothelial surface markers and the leukocyte markers CD14 and CD45 (Table 1). When expanded in culture and injected into the circulation of animals undergoing ischemic injury, these early-outgrowth cells engrafted in the injured vessel wall and restored blood flow. However, early-outgrowth cells are not alone capable of angio- and/or vasculogenesis (52, 54). A second cell population that can be isolated from the circulation arose slowly in culture. These late-outgrowth cells are positive for endothelial surface markers and lack expression of CD14 and CD45 (28, 52). Late-outgrowth cells are capable of undergoing de novo vasculogenesis (52). A third group of putative EPCs that originate in the bone marrow and that can be isolated from the blood are the multipotent adult progenitor cells (MAPCs) (39). Whereas MAPCs express endothelial cell-surface markers and possess a vasculogenic capacity, they appear to do so only after exposure to the proangiogenic vascular endothelial growth factor (39). The lineage commitment of MAPCs' “endothelial” progeny has not been studied. Recently, a vascular wall source of EPCs was identified at the medial-adventitial border of the internal thoracic artery (53). These progenitor cells possess the capability of differentiating into either mature endothelial cells, hematopoietic cells, or immune cells. Nonetheless, despite their underlying plasticity (50), it is not clear whether conduit vascular wall-derived EPCs exhibit high proliferative behavior and the capacity for self-renewal. In our studies, PAEC growth was representative of that observed in endothelial cells isolated from human umbilical veins and aorta (23). However, it was the PMVECs that displayed a remarkable distribution of highly proliferative progenitor cells, possessing a capacity for self-renewal and exhibiting vasculogenic capacity.
It is not surprising that PAECs and PMVECs display such profound functional heterogeneity. Study of endothelial cell function reveals marked structural and functional heterogeneity along individual vascular segments, a unique imprint that is conserved throughout evolution (1, 2, 51). Hagfish, the last common ancestor of vertebrates, possesses an arterial-capillary-venule endothelium that is structurally and functionally diverse (51). In the mammalian pulmonary circulation, microvascular endothelium forms a much tighter barrier than does either arterial or venule endothelium (35). Moreover, 14,15-epoxyeicosatrienoic acid or 4α-phorbol 12,13-didecanoate acid disrupt the capillary endothelial cell barrier but do not increase extra-alveolar permeability, whereas thapsigargin disrupts the arterial and venule endothelial cell barrier but does not increase capillary permeability (4, 11) . Capillary endothelial cells interact with Griffonia simplicifolia and do not interact with Helix pomatia (27). These unique site-specific attributes in vivo are maintained by PMVECs and PAECs in culture, illustrating that endothelial cells retain some imprinted memories in vitro. We have not identified a surface marker suitable to prospectively identify RMEPCs in the intact capillary. Nonetheless, the rich source of progenitor cells within the PMVEC population is consistent with the rapid growth rates and angiovasculogenic potential of microvascular endothelium in vivo (15, 20, 46). Our findings suggest that the microvasculature is an essential endothelial progenitor niche, which regenerates endothelium needed to maintain vascular homeostasis and to contribute to repair.
Despite this evidence for the angiogenic capacity of RMEPCs, angiogenesis within the postnatal pulmonary circulation has been difficult to convincingly resolve and is not a universally accepted phenomenon. Indeed, angiogenesis occurs within the lung's systemic circulation (i.e., bronchial and vaso vasorum) (13, 16), and, in some circumstances, the adjacent systemic circulation (i.e., intercostal) (32) invades the lung, where it anastomoses with the pulmonary circulation to restore blood flow. Although evidence for angiogenesis of the pulmonary circulation has been difficult to systematically resolve, several reports indicate that it does occur. There is striking evidence to support a role for lung microvascular endothelium in postnatal angiogenesis in scleroderma in humans (8), in pulmonary venoocclusive disease in humans (41), in idiopathic interstitial pneumonia in humans (36), following caloric restriction in mice (31), following intermediate levels of chronic hypoxia in rats (17), and in lung tumorigenesis in dogs (24). Knowing now that RMEPCs are intrinsically capable of forming de novo vessels, it will be important to rigorously assess how these cells contribute to normal endothelial cell turnover and the response to injury. Indeed, this work may help to guide our understanding of how and when angiogenesis occurs within the pulmonary microcirculation.
In summary, PMVECs are enriched with a population of progenitor cells, the RMEPCs. These RMEPCs are highly proliferative, and they are capable of renewing the entire hierarchy of endothelial cell growth potentials. RMEPCs maintain their endothelial cell phenotype and, more distinctively, they maintain their microvascular specification while expressing progenitor markers. Although they rapidly proliferate, they do not exhibit characteristics of transformed cells. The vasculogenic capacity of RMEPCs confers an enriched source of angiovasculogenic progenitors to the vast surface area of the microcirculation that is necessary for homeostasis and repair.
This work was supported by National Heart, Lung, and Blood Institute Grants HL-66299 and HL-60024 and by the Riley Children's Foundation.
We thank Dr. Raymond Hester, Dr. Mikhail Alexeyev, Dr. Lalita Shevde, Dr. David Ingram, Linn Ayers, Anna Bedford, and Freda McDonald for their contribution in the development of the current work.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society