Isolation of rabbit pulmonary microvascular endothelial cells and characterization of their angiotensin converting enzyme activity

Pulmonary Pharmacology (1990) 3 35-40 Longman Group UK Ltd

C 1990

0952-0600/90/0003-0035/510 .00

PULMONARY PHARMACOLOGY

Isolation of Rabbit Pulmonary Microvascular Endothelial Cells and Characterization of their Angiotensin Converting Enzyme Activity W . W . Carley*, L . Tanoue*, M . Merker*, C . N . Gillis$ Departments of Anesthesiology*, Cell Biology' and Pharmacology$, Yale University School of Medicine, New Haven, CT 06510 USA . SUMMARY. An in vitro model using cultured rabbit pulmonary endothelial cells of microvascular origin was developed to define the luminal surface membrane characteristics of microvascular endothelium . Endothelial cells were isolated from peripheral lung segments and sorted after preferential uptake of a fluorescent derivative, diiodoindocarbo cyanine acetylated-LDL . Cells were further characterized by demonstrating angiotensin converting enzyme (ACE) on their surface by means of indirect immunofluorescence . ACE activity and its pharmacologic modification were then studied as functional assays of cell activity . Hydrolysis of Benz-phe-ala-pro (BPAP), a synthetic substrate for ACE was saturable over a concentration range of 1 to 100 gm . Thus, BPAP hydrolysis in cultured microvascular endothelial cells behaves overall in a manner similar to that seen in large resistance vessels except that a portion of the hydrolysis is not inhibited by captopril, an ACE-specific inhibitor, indicating the presence of another protease capable of BPAP hydrolysis . Accordingly, this system can be used to compare ACE and other protease kinetics in microvessel cells with those of large vessel endothelium or perfused lungs .

INTRODUCTION

METHODS

The endothelium regulates transvascular exchange of water, ions, molecules and cells . In some cases the endothelium catalyzes reactions that are luminal and can effect the entire vasculature .' The microvasculature of the lung, in particular, is renowned for its abundance of angiotensin converting enzyme (ACE) whose activity not only helps regulate vascular tone 2 but also serves as an indicator of lung injury . 3 Whole lung perfusion studies in situ have measured the changes in ACE induced by drugs, vasoactive ligands and hyperoxia .3-s To identify the mechanisms of these changes in the lung, in vitro models of lung endothelium need to be well defined . The establishment of endothelial cell cultures from large resistance vessels has provided a more biochemical approach to the study of molecular changes at the luminal surface of the endothelium ; however, microvascular endothelial surface area far exceeds that of large vessels in the lung . It is the microvascular endothelium that should now, therefore, be defined reliably in vitro . Therefore, the isolation of microvascular endothelium from rabbit lung periphery and its morphological characteristics and changes in its surface ACE activity upon `injury' by agents which are known to affect ACE activity in whole lung preparations are reported here .

Cell isolation and culturing New Zealand white rabbits were anesthetized with allobarbitol 400 mg/kg IV) and urethane (100 mg/kg IV) solution . The pulmonary artery outflow tract and the left atrium were cannulated and lungs were perfused with Krebs-dextran solution 6 for 20 min at a flow rate of 50 ml/min. The lungs were removed from the thorax and peripheral segments less than 5 mm from the lung edge were dissected and minced in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal calf serum (FCS, heat inactivated), penicillin (100 U/ml FCS/DMEM) and streptomycin (100 ug/ml FCS/DMEM) . The resulting tissue suspension was centrifuged at 340 x g for 5 min, resuspended in 0 .2% collagenase and 0 .1% bovine serum albumin in PBS, and incubated for 1 h at 37°C . The suspension was filtered through a sterile 100 micron nylon mesh and recentrifuged . The cell pellet was resuspended in `Growth Medium', namely, FCS/DMEM containing heparin (20 ,ug/ml) and retinal-derived growth factor (RDGF ; 5 pl/ml)' and plated on gelatin-coated (1 .5% in PBS) plastic tissue culture flasks. Cells obtained in this manner were grown to confluence . To isolate cells of endothelial origin the primary cultures were incubated with Dil-acetylated LDL, trypsinized and washed, and were subsequently sorted with a fluorescence-activated cell sorter (FACS) . 11 Endothelial cells obtained by this method were grown to confluence and then a second sorting was performed (Fig . 1) . These cells were grown to confluence and used

*Address correspondence to : W . W . Carley, Department of Anesthesiology, Yale University School of Medicine, New Haven, CT 06510 USA .

35



36 Pulmonary Pharmacology

2

E Z Z d U

A

Forward scatter

B

Log fluorescence intensity

Fig. 1- Sorting of rabbit lung endothelial cells by DiI-Acetylated-LDL uptake . Cell lines were obtained as described in Materials and Methods, incubated with DiI-Ac-LDL (5 µg/ml, 37°C, 4 h in DMEM), washed twice with DMEM, trypsinized and resuspended in DMEM with 1 % FCS for sorting (10 6 cells/ml) . (a) Relative cell size is depicted as `forward scatter' of light versus cell number (Both scales are arbitrary units) . (B) Relative cell fluorescence of DiI-Acetylated-LDL (a four log scale is shown) versus cell number (Both scales are arbitrary units) . MDCK : Madin-Darby canine kidney epithelial cells ; R. Lunglsort 1 : rabbit lung endotheial cells incorporating Dil-acetylated-LDL used for the first sort ; R . Lung/sort 2 : rabbit lung cells incorporating Dil-Acetylated-LDL used for the second sort ; BAEC : Bovine aortic endothelial cells .

for all subsequent metabolic and characterization studies . Cell cultures were routinely grown in Growth Medium that was changed every 4 days . Cultures were not used past the twentieth passage . Unless indicated otherwise, cells were detached from the monolayer with trypsin (0 .02%) in PBS prior to passaging . Fluorescence microscopy

Endothelial cells were grown to confluence in an 8-well chamber slide (Miles Scientific) . For Dil-AcetylatedLDL staining or sorting, the cells (passage 2 after the second sort, Figure 1) were incubated with 200 µl of DiI-Acetylated-LDL (5 yg/ml) in Growth Medium for 4 h at 37°C . For detecting ACE the cells were washed twice in 300,41 DMEM per well and then incubated with a 1 : 100 dilution of goat anti-rabbit ACE serum for 20 min at 4°C . Cells were then washed with DMEM and incubated for 20 min at 4°C with a 1 : 200 dilution of rhodamine-conjugated rabbit IgG raised against goat IgG . As a control, non-immune rabbit serum was used at a 1 :100 dilution . Thereafter, the control cell wells were treated identically to those exposed to antiACE antiserum . No detectable fluorescence was seen on control cells . A Zeiss fluorescence microscope, set for rhodamine excitation (546 µm) emission (570 µm), was used to obtain micrographs . Transmission electron microscopy

Rabbit lung endothelial cells at passage 4 (after the second sort) were plated on gelatin-coated 35 min diameter dishes and grown to confluence in Growth Medium . The cells were then processed through the following steps (using 1 ml per step) : 1) glutaraldehyde (2% vol/vol) in PBS, l x 60 min ; 2) cacodylate (0 .2 M, ph 7 .5), 2 x 5 min ; 3) veronal acetate (VA) buffer (ph 7.6), 1 x 5 min ; 4) Os0 4 (1 % wt/vol) in VA buffer, 1 x 60 min ; 5) graded ethanol dehydration . Each dish was then embedded and sectioned for transmission electron microscopy . Measurement of ACE activity

Rabbit pulmonary microvessel endothelial cells were

cultured to pre-confluence . The cells were detached from the monolayer with 2 mM EGTA in PBS for 40 min at 37°C. Cells were centrifuged (340 x g, 5 min) and resuspended in DMEM with Hepes (50 mM, ph 7.2) at a concentration of 4-8 x 10 5 cells/ml and used immediately for assays (0.5 ml per assay tube). Activity of angiotensin converting enzyme was determined by the hydrolysis of the synthetic substrate 3Hbenz-phe-ala-pro . 9 The amount of BPAP hydrolysis was determined by measuring the formation (toluene extraction) of the hydrolytic product 3H-benz-phe after a 120 min incubation at 37°C of the cell suspension with varying concentrations (1-100 µm) of BPAP and 4 uCi (0 .2 nmoles) of 3H-BPAP . For all values reported a boiled sample value (i .e . background of the essay) is subtracted to control for differential metabolite extraction with cells present and for background extraction of BPAP rather than benz-phe . In studies where cells were treated with different drugs, a 10 min incubation at 37°C with DMEM alone or DMEM containing drugs was carried out prior to the addition of substrate . Viability studies

Endothelial cell viability was evaluated by dye exclusion from suspended cells. At the conclusion of the 120 min incubation, cells were incubated with 0 .4% Trypan Blue in DMEM for 5 min . Cells were 90-98% viable by Trypan Blue exclusion, whether or not they had been exposed to drug and regardless of the BPAP concentration used . MATERIALS Bleomycin was the gift of Bristol-Myers Co . (Wallingford, CT) . Unlabeled BPAP was obtained from Vega Biotechnologies, Inc . (Tuscon, Az) . 3H-BPAP (specific activity 20 Ci/mmol) was purchased from Ventrex Laboratories (Portland, ME) . Captopril was obtained from E . R . Squibb and Sons, Inc . (Princeton, NJ) . Collagenase (type II), albumin and bradykinin



ACE Activity of Rabbit Pulmonary Microvascular Endothelial Cells 37

were from Sigma Chemical Co . (St . Louis, MO) . Tissue culture media, fetal calf serum, trypsin, phosphate buffered saline (PBS) and antibiotics were from GIBCO (Grand Island, NY) and cultureware from Corning (Corning, N .Y .) . Di-idoindocarbocyanineAcetylated LDL, Dil-acetylated LDL, was from Biomedical Technologies, Inc . (Stoughton, MA) and heparin was from Fisher Scientific (Fair Lawn, NJ) .

SO

Statistical analysis

Values reported are the means ± SEM . Statistical significance was at the P < 0 .05 level and was determined by two-way analysis of variance with repeated measures on one variable and by comparing individual treatments using the Newman-Keuls test. RESULTS Characterization of isolated endothelial cells The rabbit lung endothelial cells are derived from a peripheral lung segment which contains a high density of continous endothelium . Figure 1 indicates the enrichment capabilities of the FACS application . Prior to sorting for DiI-Acetylated-LDL uptake for the first time (R . Lung/sort 1, Fig . 1) a heterogeneous population of DiI-Acetylated-LDL fluorescing cells is used . Although the fluorescence intensity covers four long units, it is apparent that, while man cells are fluorescence-positive, there are some cells which incorporate little DiI-Acetylated-LDL (i .e . cells with fluorescence intensity of Madin-Barby Canine Kidney epithelial cells) . Therefore, the cells were grown to confluence after the first sort and their fluorescence determined (R . Lung/sort 2, Figure 1) . Further sorts of this same cells population resulted in selected populations with the same fluorescence intensity as R . Lung/sort 2. These sorted rabbit lung endothelial cells incorporate more DiI-Acetylated-LDL than even the positive control, bovine aortic endothelial cells (BAEC) . Cells which have been sorted twice are used for experiments .

Fig . 2-Transmission electron microscopy or rabbit lung microvessel endothelial cells . (A) Micrograph of endothelial cells . Arrows indicate plasmalemmal vesicles open to the surface and apparently within the cytoplasm . Magnification = 34000, Bar=0 .2 µm . (B) Micrograph of endothelial cell junction (arrow) . Magnification = 50000, Bar =0 .2 ,um.

The fluorescence of second sort cells even after ten passages is equivalent to that of the first passage . In situ, the endothelial microvessel cells display plasmalemmal vesicles at the luminal and abluminal membrane as well as large junctional regions . 10 The cells isolated in the present investigation maintain these characteristics but to a smaller extent . The number of plasmalemmal vesicles in the cytoplasm is reduced but localized high densities of vesicles often occur (Figure 2A, arrows) . The cytoplasm of these cells is often filled with highly dilated rough endoplasmic reticulum filled with a proteinaecous material (Fig . 2A) . In addition, if the cells are allowed to

Fig. 3-Endothelial cell fluorescent markers (A) DiI-Acetylated-LDL uptake into lung endothelial cells. Cells are labeled and photographed without fixation . (B) Labeling of lung endothelial cells with antibody directed against angiotensin converting enzyme (Anti-ACE) and fluorophore-labeled secondary antibody without fixation .



38

Pulmonary Pharmacology

bradykinin (at 100,W), and bleomycin (at 10 and 100 pM) significantly decrease BPAP hydrolysis but at 100 aM BPAP, captopril and bradykinin had no effect while bleomycin did inhibit hydrolysis significantly .

0

NU C O E a U) T

2

DISCUSSION

T L

a a m a, 0 Z

40

20

6 0

8 0

100

Concentration BPAP (AM)

Fig. 4-BPAP Hydrolysis kinetics by lung endothelial cells . Upon addition of increasing amount of the ACE substrate BPAP the hydrolysis of BPAP reaches saturation . Substrate concentration (abscissa) ; BPAP hydrolysed during 120 min incubation (ordinate) .

become confluent they often form extensive tight junctions (Fig . 2B, arrow). The cells appear elongated (Fig. 3A) yet form a contact-inhibited monolayer and have a generation time of approximately 24 h at subconfluent densitities . In addition to the FACS data indicating a homogenous population of endothelium incorporating DiI-Acetylated-LDL and subsequent vesicular staining (Fig . 3A), all of the cells also displayed ACE on their surface by indirect immunofluorescence (Fig . 3B) . A typical fluoresence surface staining pattern is seen on these live cells indicating that ACE is surface-associated . These cells have also been characterized in a preliminary report" that attributed the origin of these cells as microvascular when their spectrum of proteins incorporating 3Hgalactose was compared to known capillary, microvessel and large vessel endothelial cells . Angiotensin converting enzyme activity As seen in Figure 4, the hydroloysis of BPAP is saturable when increasing amounts of the substrate are added to the incubation mix . Calculated kinetic constants are Vm , = 85 pmoles/ 10 6 cells/min . and Km=56,uM . Since a number of drugs are known to affect ACE activity and inasmuch as captopril is considered a highly specific inhibitor, the hydrolysis of BPAP was tested after pre-incubating the cells with different drug concentrations . Table 1 indicates that at I and 10 tM BPAP, captopril (at 10 and 100 pM),

The culture of large vessel cells has to date been the major in vitro endothelial cell model from which data have been extrapolated to whole lung . 12-14 The isolation and growth of lung microvessel endothelium provides a model which should more accurately define the response of endothelium associated with the bulk of lung vasculature . Through the use of fluorescencedependent cell sorter and endothelial-specific markers, an endothelial cell population has been isolated from lung periphery . These cells have been deemed microvessel in origin by not only the peripheral location of the tissue used but also by previous labeling of polypeptides common to microvessel endothelium . 11 Specific biochemical approaches have been used in this investigation to isolate microvessel endothelium . The method of Voyta et al . 7 and the advent of endothelium-specific antibodies 15 coupled with the use of a fluorescence-activated cell sorter has allowed the isolation of homogenous populations of endothelium . It is difficult to argue that such populations represent specific microvascular locales (i .e . arterioles, capillaries, postcapillary venules and venules) until more specific probes are available.16 Certainly, repeated cell sorting (Figure 1) indicates that whatever the origin of the microvessel endothelium isolated, it can be enriched for DiI-Acetylated-LDL uptake and will remain a stable population upon repeated passage . An approximation has been made, however, that these cells are largely microvascular in origin when comparing the 3H-galactose-incorporating polypeptides of these cells to those of known capillary or large vessel endothelium.)l In addition, the cells appear more elongated and attenuated (Fig . 3A) much like other microvessel endothelium . 17-19 A somewhat surprising finding is that these cells, unlike many other microvessel endothelium, retain the ability to form extensive junctional complexes (Fig . 2B), indicating they may form a cell monolayer with high electrical resistance and therefore be useful in molecular permeability

Table 1 . Inhibitors of ACE hydrolysis of 3 H-BPAP+ Captopril (pM) BPAP PM 1 10 100

Control

10

.18± .01 1 .45±0 .8 7 .8 ±1 .1

.08± .0.1* 1 .16± .14* 8.2 ±1 .9

100 .06± .01 .73± .15* 5 .9±1 .9

+ Units of activity are nmoles BPAP hydrolyzed in a 120 min incubation . *p< .05.

Bradykinin (pM) 100 .04± .01* .25± .11* 3 .9±1 .5

Bleomycin (pM) 10 .08±0 .1* .66± .12* 5 .3±1 .7

100 .02±0 .1* .38± .11* 2.0±1 .2*

ACE Activity of Rabbit Pulmonary Microvascular Endothelial Cells

studies 2 0 The uptake of DiI-Acetylated-LDL in these cells (Fig . 3A) appears punctate throughout the cytoplasm indicating that a lysosomal vesicle population retains the probe much as that seen in other endothelium . 7 The final indicator that these cells are endothelial in origin is their surface staining by indirect immunofluorescence when incubated with antiserum directed against rabbit ACE (Fig . 3B) . Although all of the sorted cells were stained by this antibody the cells were consistently observed rounding up in response to the antiserum, even when incubated at 4°C . The glycocalyx of the endothelium is known to be `activated' in response to this particular antiserum21 but such an effect on cell adhesion may also indicated adhesive properties for this highly glycosylated metalloproteinase. ACE activity has been measured in these cells (Fig . 4) . ACE activity was measured in detached cells at preconfluent density levels so that cells could be easily detached and concentrated in a small volume . Calculated K M and Vmax values of 56 pM and 85 pmoles/106 cells/min, respectively, were obtained . The only other measurement of ACE activity on pulmonary microvessel endothelial cells (guinea-pig microvessels) 22 reports only a percentage of total substrate (3HBPAP) utilized (1 .2, 2 .2 and 5 .8%) in the ACE assay at various time points (30, 60, and 120 min, respectively) . The magnitude of hydrolysis by these microvessel cells over a similar time course to that used with the microvessel cells reported here (0 .8, 3 .6, and 5 .7% hydrolysis at 30, 60, and 120 min, respectively) agrees quite well. No conclusions, however, can be drawn about Km and Vmax in the guinea-pig endothelium since no kinetic values were reported .22 The Km obtained from large vessel endothelial cells using 3H-BPAP was 7.6,uM (calculated from published data), 11 which is different from that reported here of 56 suM . Other reported values of K m, 8.6 mM, and Vmax, 6.7 nmoles/ 106 cells x min, (Calculated from published data) 13 are difficult to compare since the ACE substrate used was 3 H-hippurylglycylglycine . 12 The Km and Vmax reported here, therefore, are the only reported values known to these authors for lung microvessel endothelium . Finally, the effects of captopril, bradykinin and bleomycin were tested on ACE activities of these cells (Table 1) . Contrary to the reports of three laboratories' 2,13,21 3 H-BPAP hydrolysis is incompletely inhibited (70% at 1 or 10pM substrate) by 100,uM captopril while in the other studies 1-10 pM inhibitor usually results in 90% inhibition in large vessel endothelium. Bradykinin (100 pM) also inhibits hydrolysis by only 80% (at 1 and 10 pM BPAP) . Bleomycin (100,uM) appears to consistently inhibit hydrolysis somewhat more than the other inhibitors . These results suggest the presence of an enzyme, that hydrolyzes 3H-BPAP and is not responsive to these inhibitors. Such an enzyme has been reported by Lanzillo, et al .23 to exist in a latent form on cultured

39

pulmonary artery endothelial cells . Although EDTA inhibits their enzyme, lisinopril and captopril do not . Conceivably this latent enzyme may be activated on the microvessel cells used here since the cells are stimulated to grow by a RDGF/heparin addition to the medium . Such stimulation of growth, in other microvessel endothelium, 18 can dedifferentiate the endothelium and induce new surface proteases, such as plasminogen activators, which reside on endothelial cells as inactive complexes 24,25 The other intriguing possibility is that some modified form of ACE itself hydrolyzes BPAP but does not bind the inhibitors . Since bleomycin, a possible metal chelator '26 appears to inhibit 3H-BPAP hydrolysis more than captopril (Table 1), the captopril-resistant protease may be similar to the captopril-resistant metalloprotease reported by Lanzillo, et al . 23 that is inhibited by another metal chelator, EDTA . It should be mentioned that if captopril-resistant hydrolysis of BPAP is subtracted from the total hydrolysis, the Km becomes 19 pM and the Vmax 23 .1 pmoles/ 106 cells/min, values which much more closely approximate those reported for purified ACE.9 In summary, a reliable technique for isolating lung microvessel endothelium and for characterizing their morphology and ACE activity has been described . The Km and Vmax data for ACE reported here can now be used to compare with other microvessel and large vessel endothelium in culture to explain those obtained in whole lung studies . By measuring such endothelialassociated enzymes and surface antigens in an in vitro system a better molecular description of in situ and in vivo events at the endothelial lumen can be realized . Acknowledgements We would like to acknowledge the expertise of Rocco Carbone and Hans Stukenbrok for assistance with fluorescence cell sorting and electron microscopy, respectively . We thank Dr R. Soffer, Cornell University Medical College, New York, NY for the generous gift of goat anti-rabbit ACE serum . Support This work was supported by US Public Health Service Grants HL-13315 and HL-07410 to C . N . Gillis and by US Public Health Service Grant HL-40863 to C . N . Gillis and W . W. Carley and by a grant from the American Lung Association (Connecticut Chapter) to William W . Carley . References 1 . Soffer R L. Angiotensin-converting enzyme and the regulation of vasoactive peptides . Ann Rev Biochem 1976 ;45 :73-94 . 2. Cohen M L and Kurz K D . Angiotensin converting enzyme inhibition in tissues from spontaneously

40 Pulmonary Pharmacology hypertensive rats after treatment with captopril or MK421 . J Pharmacol Exp Ther 1982 ; 220 : 63-69 . 3 . Dobuler K J, Catravas J D, Gillis C N . Early detection of oxygen-induced injury in conscious rabbits . Am Rev Resp Dis 1982 ; 126 : 534-539 . 4 . Gillis C N, Catravas J D . Altered removal of vasoactive substances in the injured lung : Detection of lung microvascular injury . Ann NY Acad Sci 1982 ; 384 : 458-474 . 5 . Gillis C N, Pitt B R . The pulmonary microcirculation and metabolic functions of the lung . In : Hollinger M, ed . Current Topics in Pulmonary Pharmacology and Toxicology . Philadelphia : Prager, 1986 : 112-142 . 6 . Moalli R, Pitt B R, Gillis C N. Effect of flow and surface area on angiotensin-converting enzyme activity in rabbit lungs . J App Physiol 1987 ; 62 : 2042-2050 . 7 . D'Amore P A, Glaser B M, Brunson S K, Fenselau A H . Angiogenic activity from bovine retina : Partial purification and characterization . Proc Nat! Acad Sci USA 1981 ; 78 : 3068-3072 . 8 . Voyta J C, Via D P, Butterfield C E, Zetter B R . Identification and isolation of endothelial cells based on their increased uptake of acetylated-low density lipoprotein . J Cell Biol 1984 ; 99 : 2034-2040 . 9 . Ryan J W, Chung A, Martin L C, Ryan US . New substrates for the radioassay of angiotensin converting enzyme of endothelial cells in culture . Tissue Cell 1978 ; 10 :555-562 . 10 . Pietra G G, Sampson P, Lanken P N, Hansen-Flasheu J, Fishman A P . Transcapillary movement of cationized ferritin in the isolated perfused rat lung . Lab Investig 1983 ; 49 : 54-61 . 11 . Carley W W, Dolci E D, Palade G E . Differences in polypeptide labeling between cultured endothelia derived from large and small vessels. J Cell Biol 1987 ; 105 : 326a . 12. Catravas J D, Watkins C A . Plasmalemmal metabolic activities in cultured calf pulmonary arterial endothelial cells . Res Commun Chem Path Pharm 1985 ; 50 : 163-179 . 13. Del Vecchio P J, Smith J R. Expression of angiotensin converting enzyme activity in cultured pulmonary artery endothelial cells . J Cell Physiol 1981 ; 108 : 337-345 . 14. Rosen E M, Noveral J P, Mueller S N, Levine E M . Regulation of angiotensin I-converting enzyme activity in serially cultivated bovine endothelial cells . J Cell Physiol 1985 ; 122 : 30-38 .

15 . Gumkoroski, F, Kaminska G, Kaminska M, Morrissey L W, Auberbach R. Heterogeniety of mouse vascular endothelium. In vitro studies of lymphatic, large blood vessel and microvascular endothelial cells . Blood Vessels 1987 ;24 :11-23 . 16 . Gerritsen M E, Carley W W, Milici A J . Microvascular endothelial cells : Isolation, identification and cultivation . In : Maramorosch K, Sato G 1-I, eds . Advances in Cell Culture . San Diego : Academic Press, 1988 :35-67 . 17 . Folkman J, Haudenschild C . Angiogenesis in vitro . Nature 1980 ; 288 : 551-556 . 18 . Milici A J, Furie M B, Carley W W . The formation of fenestrations and channels by capillary endothelium in vitro . Proc Natl Acad Sci USA 1985 ; 82 : 6181-6185 . 19 . Madri J A, Williams S K. Capillary endothelial cell cultures : Phenotypic modulation by matrix components. J Cell Biol 1983 ; 97 : 153-165 . 20 . Shasby D M, Roberts R L . Transendothelial transfer of macromolecules in vitro. Fed Proc 1987 ; 46 : 2506-2510 . 21 . Habliston D L, Whitaker C, Hart M A, Ryan U S, Ryan J W . Isolation and culture of endothelial cells from the lungs of small animals . Am Rev Resp Dis 1979 ; 119 : 853-868 . 22 . Ryan U S. Pulmonary endothelium : A dynamic interface . Clin Investig 1986 ; 9 : 124-132 . 23 . Lanzillo J J, Dasarathy Y, Stevens J, Fanburgh B L . Conversion of angiotensin- I to angiotensin-2 by a latent endothelial cell peptidyl dipeptidase that is not angiotensin-converting enzyme . Biochem Biophys Res Comm 1986 ; 134 : 770-776. 24 . Loskutoff D J, Edgington T S . Synthesis of a fibrinolytic activator and inhibitor by endothelial cells . Proc Natl Acade Sci USA 1977 ; 74 : 3903-3907 . 25 . Levin E G, Loskatoff D J . Comparative studies of the fibrinolytic activity of cultured vascular cells . Thromb Res 1979 ; 15 : 869-878 . 26 . Lazo J S, Lynch T J, McCallister J . Bleomycin inhibition of angiotensin-converting enzyme activity from serum, lungs and cultured pulmonary artery endothelial cells . Am Rev Respir Dis 1986 ; 134 : 73-78.

Date received : 16 June 1989 Date revised : 29 August 1989 Date accepted : 30 August 1989