Isolation of murine lung endothelial cells

Abstract

Several protocols for the isolation of endothelial cells (ECs) from murine lung have been described in the literature. We, however, encountered a number of problems while using these procedures that prevented us from consistently or reliably obtaining pure populations of ECs from the lungs of mice. By incorporating specific elements from previously published protocols, as well as adding some novel features, we developed a new strategy for isolating ECs from murine lung. In this approach, a suspension of lung cells is initially prepared from the lungs of 7- to 14-day-old mouse pups using procedures that prevent intravascular clotting and leukocyte activation, minimize mechanical trauma to the lung tissue, and limit exposure to the digesting enzymes. The resulting cell suspension is cultured for 2–3 days, trypsinized to produce a suspension of single cells, and then subjected to fluorescence-activated cell sorting using an anti-ICAM-2 antibody. The sorted cells are then plated and split 1:2 at each passage to maintain a high density of the cells. Using this approach, we have been able to isolate pure populations of ECs that were sustainable for extended periods in culture without the emergence of fibroblast overgrowth or the development of senescence. We believe the success of this approach will provide opportunities to take advantage of the large and growing number of knockout and transgenic mouse lines to investigate the endothelial-specific roles of targeted molecules in the pulmonary vasculature.

the isolation of various cell types from the large and growing number of knockout and transgenic mouse lines (9, 21) has provided opportunities to investigate the cell-specific roles of targeted molecules. This has certainly been the case for various molecules expressed by the endothelium of the vasculature (14, 30), including endothelial cells (ECs) of the pulmonary circulation (22, 24, 27). With respect to ECs, a number of strategies for isolating murine ECs have been described in the literature. These have included techniques for the isolation of mouse ECs from a variety of tissues (1, 3, 5, 10, 13, 15, 20, 26 29, 31) including the lung (5, 6, 12, 16, 18, 19, 22, 24, 27).

In our studies to assess the role of endothelial cell adhesion in angiogenesis and lung development, a number of these approaches (see Table 1) were employed in an effort to isolate ECs from the lung of mice deficient in the expression of several cell surface molecules of interest to us, including CD44 and PECAM-1 (2, 4). Despite their reported effectiveness, we, however, found that these strategies failed to consistently produce significant numbers of ECs or resulted in impure populations of ECs that were variably contaminated by fibroblasts. In this report, we describe an approach that enabled us to consistently isolate pure populations of murine lung ECs, highlighting some of the important features we believe were responsible for its effectiveness, which may also be helpful to others isolating murine ECs from other organs.

Table 1.

Published protocols for the isolation of ECs from murine lung

Source (Reference Citation)	Age of Mouse	Technique Used to Disperse Lung	Purification Method	Surface Marker(s) Used for Magnetic Sorting
Demeule et al. (5)	Adult*	Collagenase A digestion of diced tissue	Single magnetic bead sort	PECAM-1
Dong et al. (6)	Adult (18–20 g)	Collagenase A digestion of minced tissue	Double magnetic bead sort	PECAM-1 (for both sorts)
Khan et al. (12)	5–7 days	Collagenase A digestion of minced tissue	Single magnetic bead sort	ICAM-2
Kuhlencordt et al. (16)	3–4 mo	Collagenase A digestion of minced tissue	Double magnetic bead sort	ICAM-2 (for both sorts)
Lim et al. (18)	8–10 wk	Collagenase I digestion of minced tissue	Double magnetic bead sort	PECAM-1 (1st sort) ICAM-2 (2nd sort)
Marelli-Berg et al. (19)	6–8 wk	Collagenase (type not specified) digestion of diced tissue	Single magnetic bead sort	PECAM-1, endoglin, and isolectin B4

Summarized are some of the major features of published procedures for the isolation of endothelial cells (ECs) from mouse lungs.

^*Protocol originally described for rats.

MATERIALS AND METHODS

Reagents and chemicals.

All reagents and chemicals were obtained from Sigma (St. Louis, MO) unless otherwise specified. FITC-labeled, isolectin B4 [Griffonia (Bandeiraea) simplicifolia lectin I] was obtained from Vector Laboratories (Burlingame, CA). DiI-Ac-LDL (acetylated low-density lipoprotein, labeled with 1,1′-dioctadecyl-3,3,3′,3′-tetramethyl-indocarbocyanine perchlorate) was obtained from Biomedical Technologies (Stoughton, MA). DAPI (4′-6-diamidino-2-phenylindole) was obtained from Molecular Probes (Eugene, OR). TNFα was purchased from eBioscience (San Diego, CA).

Antibodies.

The following antibodies against murine surface receptors were employed: rat anti-PECAM-1 antibody, clone 390 (32); anti-ICAM-2 antibody, clone 3C4, unlabeled and FITC labeled, from Southern Biotech (Birmingham, AL); anti-VE-cadherin and anti-eNOS antibodies from BD Biosciences (San Jose, CA); KM81, anti-CD44 antibody from American Type Culture Collection (ATCC; Rockville, MD); anti-S100A4 from Abcam (Cambridge, MA); anti-α-smooth muscle actin (Sigma) and anti-GAPDH antibody from Millipore (Temecula, CA). FITC-conjugated donkey, anti-rat IgG was obtained from Jackson ImmunoResearch Laboratories (West Grove, PA).

Cell lines.

H5V murine EC (8), B16 murine melanoma cells (from ATCC), and 3T3 fibroblasts (from ATCC) were cultured in DMEM containing 10% FBS, penicillin/streptomycin, and 2 mM l-glutamine (DMEM complete). Lung ECs were isolated as described below from wild-type and PECAM-1- and CD44-null mice. Isolated cells were cultured in M199 medium containing 15% FBS, 50 μg/ml endothelial growth factor (BD Bioscience), 100 μg/ml heparin, and 1 mM glutamine (M199 complete). When assessing for the expression of mouse PECAM-1, cultured cells were detached using an enzyme-free cell dissociation solution from Chemicon (Temecula, CA).

Animals.

The Institutional Animal Care and Utilization Committees at both the Wistar Institute and the University of Pennsylvania School of Medicine approved all animal care procedures. Wild-type mice, on a C57BL/6 background, were obtained from Taconic (Germantown, NY). CD44-null mice (28), on a C57BL/6 background, were the kind gift of Dr. Tak Mak (Amgen, Toronto, Canada). PECAM-1-null mice (7), on a C57/Bl6 background, were the kind gift of Steven Albelda (Univ. of Pennsylvania, Philadelphia, PA).

Isolation and culture of ECs from murine lung.

For each isolation, four to six mouse neonatal pups (7–14 days old) were used for our procedures. Each mouse initially received a 25-μl intramuscular injection of heparin (1,000 USP U/ml). Ten minutes later, the mouse was anesthetized (ketamine/xylazine, 140/14 mg/kg), followed by exposure of the thoracic cavity. Five milliliters of cold DMEM was then injected via the right ventricle to flush the lung of blood cells. One milliliter of collagenase A (1.0 mg/ml for 7- to 10-day-old pups or 1.5 mg/ml for 11- to 14-day-old pups) was then quickly instilled through the trachea into the lungs, and the trachea was then tied off. The lungs (without the heart) were subsequently removed and then incubated with 5 ml of collagenase A in a 50-ml tube for 30 min in a 37°C water bath. Every 5–8 min during this incubation, the tube was gently agitated for a few seconds. After the 30-min incubation, 25 ml of 1× PBS was added to the tube. The tube was then vigorously shaken for 30 s to dissolve the lung, and the resulting tissue/cell suspension was filtered through a 70-μm strainer. The filter was washed with 5 ml of 1× PBS. The filtered cell suspension (∼35 ml) was centrifuged for 4 min at 900 rpm. After removal of the supernatant, the cell pellet was washed once with complete DMEM and then resuspended in 10 ml of complete DMEM and plated into a gelatin-coated T-75 tissue culture flask. The following day, the medium was changed to M199 complete, and the cells were cultured for an additional 1–2 days. At this time, the cells from each lung were removed by trypsin and pooled together into one suspension for the sorting. The lung cell suspension was then subjected to FACS sorting using anti-ICAM-2 antibody as described below. The sorted cells were resuspended in M199 complete, plated at a concentration of 300,000 cells/ml into gelatin-coated T-25 flasks, and subsequently split 1:2 at each passage.

Magnetic bead sorting to isolate murine ECs.

Lungs were prepared and digested, and the resulting lung cells were then plated and cultured as described above. The cultured cells were then detached by trypsin, and the ECs were isolated according to the manufacturer's (Dynal Biotec, Lake Success, NY) instructions, using Dynabeads coupled to anti-ICAM-2 antibody and the Dynal MRC-L Magnetic Particle Concentrator.

Immunofluorescence staining and confocal imaging.

Cells cultured on fibronectin-coated chamber slides were washed in PBS, fixed with 3% paraformaldehyde for 10 min, and then permeabilized with ice-cold 0.5% Nonidet P-40 for 1 min. After being washed, cells were subjected to immunofluorescence staining with the appropriate rat anti-mouse antibody followed by FITC-conjugated goat anti-rat IgG secondary antibody to identify the cell surface expression for PECAM-1, VE-cadherin, and ICAM-2. For DAPI staining, after removal of the secondary antibody and washing of the slide, the DAPI solution was added for 5 min in the dark and then washed with distilled water. For DiI-Ac-LDL labeling, plated cells were incubated for 5 h at 37°C with DiI-Ac-LDL diluted to 10 μg/ml in complete growth medium. The cells were then washed with probe-free medium and fixed with 3% paraformaldehyde. The slides were viewed by confocal microscopy with a Zeiss phase-epifluorescent microscope using a ×60 fluorescence lens.

Fluorescence-activated cell sorting analysis.

Cells were treated with various anti-murine antibodies for 1 h at 4°C. The primary antibody solution was then removed, the cells were washed with PBS, and a 1:200 dilution of FITC-labeled goat anti-rat secondary antibody (Jackson) was added for 40 min at 4°C. After washing in PBS, FACS was performed using an Ortho Cytofluorograph 50H cell sorter equipped with a 2150 data handling system (Ortho Instruments, Westwood, MA). For the endothelial cell isolations, the cells were resuspended to a concentration of 20–40 × 10⁶ cells/ml in 1× PBS with 4% serum for the sorting. The sorted cells were collected into M199 complete medium with 50% serum and then resuspended in M199 complete with 15% serum for plating.

Western blotting.

Total endothelial cell lysates were loaded in equal protein amounts (10 μg) determined by BCA (Pierce, Rockford, IL). Proteins were resolved by SDS-PAGE (Novex, Invitrogen), followed by transfer onto a nitrocellulose membrane using the iBlot Dry Blotting System (Invitrogen), which employs semidry electrotransfer. Membranes were washed in 1× TTBS for 2–3 min, blocked with 5% Blotting Grade Blocker solution (Bio-Rad Laboratories, Hercules, CA), and incubated with the primary antibody in 2% BSA for 1 h at room temperature. Unbound antibodies were washed off with TTBS before membranes were incubated with HRP-labeled species-specific secondary antibodies for 1 h at room temperature. After again washing the membranes with PBS, bound antibody signals were detected by ECL substrate and documented on X-ray film. The chemiluminescent signals were quantified by densitometry (ImageQuant; Amersham, Piscataway, NJ) and normalized to the housekeeping protein, GAPDH.

In vitro tube formation assay.

In vitro tube formation was studied using previously described procedures (32). Fifty microliters of Matrigel (Collaborative Biomedical Products, Bedford, MA) were added to each well of a 96-well plate and allowed to form a gel at 37°C for 30 min. Cells (20,000) in 200 μl of complete medium were subsequently added to each well and incubated for 6 h at 37°C in 5% CO₂. The wells were washed, and the gel and its cells were fixed with 3% paraformaldehyde. Total tube length per well was determined by computer-assisted image analysis using the Image-Pro Plus program.

RESULTS AND DISCUSSION

Isolation of murine lung ECs using previously published protocols.

Several techniques have been have been described in the literature (Table 1) for the isolation of ECs from murine lung (5, 6, 12, 16, 18, 19). Features common to all of these protocols are the use of adult mice [except for Khan et al. (12), who used 5- to 7-day-old mice]; digestion of the lung using collagenase treatment (×45–60 min) of minced or diced lung; and the use of magnetic bead sorting to directly isolate the ECs from collagenase-digested lung cell suspensions [except for Kuhlencordt et al. (16), who sorted lung cells that had been cultured for 2–4 days]. In our hands, however, we were unable to consistently or reliably obtain pure populations of ECs from murine lung using any of these approaches. One or more of the following problems were encountered with each of the protocols we employed: 1) few if any viable cells were present after the digestion of the lung; 2) significant fibroblast contamination despite sorting with endothelial-specific markers such as PECAM-1 or ICAM-2; 3) endothelial cell populations that were reasonably pure initially (>85%) but quickly became overgrown by fibroblasts with serial passaging of the cells; 4) ECs that had low proliferative capacities and/or quickly became senescent; and 5) inconsistency in the outcomes.

The reasons for our lack of success in all instances are not clear, but several things can be cited. First, the various digestion procedures (mincing or dicing followed by collagenase digestion and filtering) result not in a suspension of single cells but in clumps of cells composed of a mixed population of cells (Fig. 1). Thus in sorting these suspensions of cells, one is also likely to separate ECs associated with other cells, including fibroblasts. In addition, digestion of the lung in this way appears to result in physical and physiological trauma that reduces the immediate number of viable cells (data not shown). Second, during the magnetic bead sorting, there is nonspecific binding of the beads to non-endothelial cells (Fig. 2). Third, isolectin B4, which has been described as EC specific (17) and has been used for sorting (19), turns out to bind to not only murine ECs but to other cell types, including murine fibroblasts and tumor cells, as assessed by FACS analysis (Fig. 3A). Fourth, as others have done, we initially relied on the uptake of DiI-Ac-LDL as a marker of endothelial identity (25). We, however, came to appreciate that murine fibroblasts, like murine ECs, also have the ability to take up DiI-Ac-LDL, although less avidly than the ECs (Fig. 3B). This resulted in an underestimation of the level of fibroblast contamination. Last, when we did see some success, it was with ECs isolated from neonatal mouse pups, suggesting that the proliferative capacity of ECs isolated from adult mice is limited.

Fig. 1.

Cell morphology in suspensions of collagenase-digested lung. Cell suspensions of collagenase-digested lungs were plated, fixed, and stained immunofluorescently for ICAM-2, an endothelial cell marker, and with DAPI to identify cell nuclei. A: observed are clusters of endothelial cells (ECs) of various sizes. *Cell cluster that is analyzed in B and C. On closer inspection, it is clear that the clusters are composed of not only ICAM-2-postive (i.e., ECs) but also ICAM-2-negative cells (B) as well. This is illustrated by the dark “holes” (arrows) in areas that contain cells as shown by the DAPI staining (C).

Fig. 2.

Murine ECs isolated by magnetic bead sorting. Shown are ECs isolated by magnetic bead sorting with ICAM-2 antibody that were immunofluorescently stained for VE-cadherin to identify them and with DAPI to identify their nuclei. The magnetic beads appear as orange spheres in the image, clustered primarily on the ECs. Also present, however, are cells (within the box) that do not express VE-cadherin, the nuclei of which are detected by DAPI staining. Magnetic beads are also associated with some of these VE-cadherin-negative cells (arrows).

Fig. 3.

The specificity of isolectin B4 binding and DiI-Ac-LdL uptake as markers of ECs. A: the binding of FITC-labeled isolectin B4 to various murine cells was assessed by FACS analysis. Rat IgG, blue tracing; isolectin B4, red tracing. Isolectin B4 bound not only to ECs but to the murine 3T3 fibroblast and B16 melanoma lines as well. B: the uptake of DiI-Ac-LdL by murine ECs was determined by immunofluorescence microscopy. 3T3 murine fibroblasts demonstrated the capacity to take up DiI-Ac-LdL, although less than that of the ECs.

Development of an alternative method for the isolation of ECs from murine lung.

Given the difficulties we encountered using previously published protocols, we developed an alternative strategy for isolating ECs from murine lung. In doing this, our goals were to develop an approach that 1) enabled a more complete digestion of the lung and dispersal of the cells; 2) produced more viable cells after sorting; 3) minimized fibroblast contamination; and 4) resulted in ECs that could be serially cultured to high passage numbers. Using elements from some of the previously published protocols, as well as introducing some novel features, we were able to develop a new strategy that accomplished these goals. This approach is detailed in materials and methods and is summarized in Fig. 4A.

Fig. 4.

Purification of ECs from mouse lung. A: the major steps of the protocol used in this paper for the isolation of ECs from murine lung. B: shown are cells from digested lung that had been plated and cultured for 24 h and then stained immunofluorescently with an anti-PECAM-1 antibody. The staining demonstrates the presence of clusters and aggregations of ECs. Cell nuclei are detected by DAPI staining in blue. C: FACS, using ICAM-2 antibody, of cells from digested lung that had been cultured for 2–3 days. Background, black tracing; ICAM-2, red tracing. A population of ICAM-2-expressing (endothelial) cells was present (circle) and was selected for sorting.

Briefly, in our protocol, the ECs are isolated from the lungs of 7- to 14-day-old mouse pups. Ten minutes after an intramuscular injection of heparin, the chest is opened, and cold media is perfused into the lung via the right ventricle to flush blood from the vasculature of the lung. Collagenase A is next instilled through the trachea into the lungs. The lungs are then removed and incubated in collagenase A for 30 min in a 50-ml tube. At the end of the incubation, the tube is vigorously shaken for 30 s. With these procedures, the lung completely disperses without the persistence of large visible pieces of tissue. After being filtered and washed, the lung cells are plated and cultured. When viewed, these cultures reveal the presence of endothelial cell clusters and aggregates (Fig. 4B). After 48–72 h in culture, the lung cells are removed using trypsin to produce a suspension of single cells and then subjected to FACS using anti-ICAM-2 antibody (Fig. 4C). Typically, 1–5% of the suspension cells are ECs. The sorted cells are then plated and split 1:2 at each passage to maintain a high density of the cells.

Consistent isolation of ECs from murine lung.

Using the approach described above, we were able in 30 of 34 attempts to isolate pure populations of ECs that demonstrated the cobblestone appearance typical of cultured ECs (Fig. 5A). (The 4 unsuccessful attempts were associated with the first 7 isolations during the initial application of our procedures; subsequent isolations have been successful.) The endothelial identity of the cells obtained from wild-type mice was confirmed by the expression of ICAM-2, PECAM-1, VE-cadherin, and eNOS, as demonstrated by FACS analysis, immunofluorescence staining, and/or Western blotting (Fig. 5, B–E). These cells did not express the fibroblast maker FSP-1/S100A4 (Fig. 5E) or stain for smooth muscle actin (data not shown). This technique also proved successful in isolating ECs from PECAM-1- and CD44-null mice (Fig. 5). ECs isolated by these procedures formed tubular/cordlike networks on Matrigel, and stimulation with TNFα for 24 h upregulated their expression of ICAM-1 (Fig. 6). We found that for most isolations, we were able to culture the cells to at least passage 8–10 without the development of senescence. We, however, did observe that the proliferative capacity of the late passage cells (> passage 10) and the expression of PECAM-1 and VEGFR-2 appeared to wane somewhat in the higher passages (data not shown). Consequently, cells were only used during passages 4–8 for our in vitro functional studies. Cultures were routinely assessed at passages 4 and 6 by immunofluorescence staining for PECAM-1 and ICAM-2 to confirm the continued purity of the ECs. In the two instances where there was the late emergence of significant numbers (>10%) of contaminating cells, repeat sorting by FACS was performed to restore the purity of the ECs. In our experience, we found that the isolated ECs did not proliferate well after freezing and thawing. We would also note that because whole lungs were used to isolate the ECs, there could be ECs from the bronchial circulation. However, given the small size of the bronchial circulation relative to the pulmonary circulation, there is no reason to believe that bronchial ECs will be anything but a very small percentage of the ECs isolated.

Fig. 5.

Confirmation of the endothelial identity of cells isolated from murine lung. A: confluent monolayers of early passage ECs from wild-type mice demonstrated the cobblestone appearance typical of cultured ECs. As shown by FACS analysis (B), immunofluorescence staining (C), and Western blotting (D), these cells expressed ICAM-2, PECAM-1, and VE-cadherin. These cells also expressed eNOS, but not FSP-1/S100A4, which was detected in 3T3 fibroblasts (E). ECs were also successfully isolated from PECAM-1- and CD44-null (KO) mice. For the FACS analysis, background, black tracing; rat IgG, blue tracing; PECAM-1, green tracing; and ICAM-2, red tracing.

Fig. 6.

Functional analyses of isolated murine lung ECs. A: shown are tubular/cordlike structures of ECs from murine lung that formed on Matrigel after 6 h. B: stimulation of these cells with TNFα for 24 h upregulated the expression of ICAM-1.

Factors contributing to the successful isolation of murine ECs.

There are several aspects of this approach that we believe have contributed to its overall effectiveness as a strategy for the isolation of murine ECs from the lung. The first is the use of lungs from 7- to 14-day-old mouse pups rather than lungs from older or adult animals. Although we found that we could isolate ECs from adult mice with our procedures, their growth in culture was typically not sustained for more than one or two passages, and they were more susceptible to fibroblast overgrowth. This is not surprising given the evidence that cellular senescence and the potential for cell division of human primary cultures depend on donor age (23). The tsA58 large T antigen (tsA58T Ag) is a mutated SV40T Ag that results in temperature-dependent, cell type-independent cell proliferation (11). A transgenic mouse line has been developed that expresses tsA58T Ag in an endothelial cell-specific manner using Cre/lox P recombination, and ECs isolated from these animals proliferate continuously at 33°C without undergoing senescence (31). These mice may therefore facilitate the culturing of wild-type ECs from older mice without the problems of senescence. It should be noted, however, that while these transgenic mice grow normally, they die suddenly within 6–12 wk after birth, thus limiting somewhat the usefulness of these animals as a reliable source of ECs from older mice. The tsA58T Ag has also been used to immortalize freshly isolated murine ECs and may provide another strategy for overcoming cellular senescence (27), although the potential for molecular and functional changes in the transformed cells will always be a concern.

A second important feature of our strategy is the way in which the lungs are prepared and digested. The previous procedures all involved the potential for intravascular clotting and leukocyte activation, mechanical trauma to the lung tissue, and exposures to the collagenase of up to 1 h. With our protocol, however, clotting was prevented, and the lungs were flushed of blood cells before any significant manipulation of the lungs. In addition, the intratracheal instillation of collagenase A allowed us to limit the incubation time in the collagenase to 30 min and enabled the lungs to disintegrate more completely without much direct physical manipulation of the lung. These interventions, we believe, act to limit injury to the ECs, thereby increasing the likelihood of their subsequent survival.

Another critical element that contributes to the success of our approach is that the digested cells are plated, cultured for 2–3 days, and then trypsinized immediately before they are sorted by FACS. These procedures provide time for viable cells to recover, allow dead or dying cells to detach (and thus be excluded from the sorting), and results in a suspension of single cells for the sorting. Our approach differs from several earlier published procedures in which the suspension of digested lung cells is immediately subjected to magnetic bead sorting. As illustrated in Fig. 1, the various digestion procedures employed by these other groups result in clumps of cells composed of a mixed population of cells. Thus, in sorting the cells in this way, one is also likely to separate ECs associated with other cells, including fibroblasts.

The last significant factor, we believe, is the use of FACS, instead of magnetic bead selection, for the sorting and isolation of the ECs. We found that although PECAM-1 or ICAM-2 antibody-coupled magnetic beads bound preferentially to ECs, there was still some nonspecific binding that could also lead to the isolation of other cell types (Fig. 2). In our own head-to-head comparisons of FACS vs. magnetic bead sortings of the same suspension of trypsinized cells, the results of magnetic bead sorting were always inferior to that of FACS (data not shown). In most instances, less than 50% of the cells recovered by magnetic bead sorting were ECs. Whether the use of other products for magnetic bead sorting (e.g., MACS products, Miltenyi Biotec; see Refs. 1, 5, 19, 31) would have improved the outcomes is unclear. The efficiency might be improved with serial or repeated magnetic bead sorting. We did in fact do this but did not find this was effective in efficiently improving the purity of the ECs. Thus, although sorting by magnetic beads may be a cheaper and more convenient approach for cell sorting, it is our conclusion that these advantages are outweighed by the apparent relative inefficiency of this approach compared with FACS. In this regard, we would note that we sorted by targeting ICAM-2 rather than PECAM-1 due to the sensitivity of mouse PECAM-1 to cleavage by trypsin (19), which we employed to detach the cultured lung cells.

Conclusions.

Our experience may be instructive to others developing strategies for the isolation of murine ECs from other organs. In these efforts, procedures should be employed that minimize the mechanical trauma to the organ and ultimately disperse the tissue into single cells. It is also important to keep in mind such factors as the inefficiencies of magnetic bead sorting, some of the lack of specificity of isolectin B4 binding and acetylated-LDL uptake as markers of endothelial identity, and the trypsin sensitivity of mouse PECAM-1. In addition, the purity of the isolated ECs should be monitored, and caution should be exercised in using cells from higher passages.

The isolation of various cell types from the large and growing number of knockout and transgenic mouse lines has provided opportunities to investigate the cell-specific roles of targeted molecules. This has certainly been the case for molecules expressed by the endothelium of the vasculature (14, 22, 24, 27, 30). Where adult mice must be used and pure populations of ECs or extended culturing are not necessary, the earlier protocols may be adequate. However, we believe our approach is very suitable for those settings in which mouse age is irrelevant and relatively large numbers of pulmonary vascular ECs, uncontaminated by other cells, are required. Finally, when the source of the ECs is not a concern, our procedures also represent a suitable means of obtaining murine ECs in general.

GRANTS

This work was supported by grants from the Department of Defense (PR043482) and National Heart, Lung, and Blood Institute (HL-079090).