British Journal of Huernatology. 1992, 82, 151-160

Mitogenic role for extracellular calmodulin-like activity in normal human umbilical vein endothelial cells R. A. DAWSONA N D S . MAC NEILDepartment of Medicine, University of Shefield, Clinical Sciences Centre, Northern General Hospital, Shefield

Received 29 January 1992; accepted for publication 24 April 1992

Summary, Normal human umbilical vein endothelial cells cultured on gelatin-coated plastic dishes were found to produce a protein in their media which had calmodulin-like immunoreactivity and biological activity. Further identification of the protein was achieved by examining the incorporation of 14C leucine into protein found in the conditioned medium. Cells produced 14Clabelled protein in their medium which specifically bound to an affinity column for calmodulin. This latter material stimulated calmodulin dependent phosphodiesterase activity in vitro and this stimulation was inhibited by the addition of the calmodulin antagonist W7. The presence of calmodulin-like activity and immunoreactivity in the media varied as the cells grew from low to high density, a peak of extracellular calmodulin-like activity

preceding an increase in cell number. Extracellular calmodulin-like activity did not correlate with the presence of lactate dehydrogenase in the medium. The addition of pure pig brain calmodulin affected the rate of cell proliferation: significant proliferation to pure calmodulin was only seen in cells at low density, at higher density calmodulin either had no effect or inhibited proliferation. Inhibition of extracellular calmodulin activity by a calmodulin antagonist immobilized on agarose beads, or by an antibody to calmodulin significantly decreased proliferation in all dividing cultures. Taken together this data suggests that, in vitro, calmodulin, or a very closely related protein, influences endothelial cell proliferation through an autocrine mechanism.

Many of the actions of calcium are known to occur when it is bound to the calcium binding protein calmodulin. In particular, the calcium calmodulin complex is thought to play a vital role in the regulation of cell proliferation (reviews by Means & Dedman, 1980; Tomlinson et al, 1984). The levels of intracellular calmodulin vary throughout the cell cycle. A specific peak has been identified at the GI phase preceding DNA synthesis (Chafouleas et al, 1984) and a genetic manipulation of the mRNA for calmodulin to increase the intracellular calmodulin results in a shortening of the cell cycle time (Rasmussen & Means, 1987). However, investigations into a relationship between intracellular calmodulin levels and rates of cell proliferation have so far remained inconclusive (Chafouleaset al, 1981;Veigletal, 1982,1984). Addition of pure calmodulin to the culture medium has been shown to stimulate cell proliferation in a rat liver cell line, human peripheral blood lymphocytes, mouse melanoma cells and human leukaemic lymphocytes (Boynton et al, 1980; Gorbachevskaya et al, 1983; Mac Neil et al, 1984b: Crocker et al, 1988). The conditioned media of a variety of

cells have been shown to contain undefined growth stimulants (Lincoln et al, 1982; Koos, 1986). Our own work has identified a calmodulin-like substance in a wide range of cell conditioned medium and established a critical role for extracellular calmodulin-like activity in the regulation of leukaemic lymphocyte proliferation (Crocker et al, 1988). A rich potential source of calmodulin in vivo is from the platelet which in vitro contributes relatively high levels of calmodulin to the serum (Mac Neil et al, 1984a, 1988). A possible target for this calmodulin could be the closely situated endothelial cell. The aim of this study was to determine whether extracellular calmodulin plays a role in the regulation of normal human endothelial cell proliferation.

Correspondence:Dr S. Mac Neil, Department of Medicine, University of Sheffield, Clinical Sciences Centre, Northern General Hospital, Sheffield S5 7AU.

151

MATERIALS AND METHODS Materials Medium 199 and glutamine were purchased from Flow Laboratories Ltd, Rickmansworth, Herts., U.K. Human AB serum was obtained from Trent Regional Blood Transfusion Service and was heat inactivated at 56°C for 45 min before use. Antibiotics benzyl penicillin sodium from Glaxo, streptomycin sulphate from Evans and a fungicide, amphotericin,

152

R. A. Dawson and S. Mac Neil

were obtained from the Northern General Hospital pharmacy. Tissue culture plastics were purchased from Falcon, Becton and Dickinson. 2 Bridgewater Lane. Lincoln Park, New Jersey, U S A . Collagenase Type IV. gelatin W7 agarose, Dowex 1 x 8 (200-400 mesh). and lactate dehydrogenase kits were purchased from Sigma Chemical Co.. Poole. Dorset. U.K. Calmodulin-deficient ox heart phosphodiesterase and King-cobra (Ophiophagus hcznncrh venom were purchased from Boehringer Mannheim, Bell Lane, Lewes. East Sussex. U.K. Calmodulin radioimmunoassay kits were purchased from Amersham International plc. Lincoln Place. Aylesbury. Buckinghamshire. I'.K. Cyclic [PY~HIAMP.[ 'Hlthymidine and ['T]leucine were purchased from Amersham. Buckinghamshire, U.K. Sodium bicarbonate was purchased from BDH. Fourways. Atherstone. Warwickshire. U.K. Endothelial cell growth supplement (ECGS). was purchased from Collaborative Research. Oak Park. Bedford. U.K. Sheep IgG was purchased from Nordic Immunological Labs. Ltd. Maidenhead, Berkshire, U.K. Polyclonal anti-calmodulin IgG was purchased from Peninsula Laboratories. Jackson Street, St Helens. Merseyside, U.K. Met hods Cell culture. The isolation and culture of human umbilical vein endothelial cells was a modification of the method described by Jaffe et a1 (1973). This method briefly entails clamping each end of the human umbilical cord (collected in medium within 4 8 h of delivery) and filling the vein with a 032% solution of the enzyme collagenase and incubating for 10 min at 37°C. Cleaved endothelial cells were then collected by washing the vein with Medium 1 9 9 with 10%human AB serum. The cells were cultured on gelatin coated tissue culture plastics in medium 199 containing 103,human AB serum. 22 g/l NaHCO$. 2 msr glutamine. 10' U i l penicillin. 100 mgjl streptomycin and 2 . 5 mg,1 fungizone. Cells were grown to confluence in TZ 5 flasks then transferred to 6 . 12 or 24 well plates for use at first passage. After plating (between 5 x 10%and 1 x loi cells;ml using 1 ml levels for initial plating) cells were allowed to attach (approximately 24 h ) then washed with fresh medium prior to the initiation of experiments. Viability was measured using the trypan blue dye exclusion test. This method routinely produced cultures of endothelial cells as determined by their morphology. ability to produce prostacyclin (by the method of Greaves & Preston, 1 9 8 2 )and positive staining for von LVillebrand's factor (by the method of Jaffe et al, 1973). Measurement of cnlniodulin in cells and in conditioned media. Cell conditioned medium collected from cultures of high viability was centrifuged for 5 rnin at 5 0 0 g to remove unattached cells prior to preparation for calmodulin assay. For measurement of intracellular calmodulin. cells were enzymatically collected (using collagenase) and lysed by suspension in 1 ml of distilled water followed by sonication. Cellular debris was then discarded after centrifugation at 11 6 0 0 g for 5 min leaving the cell contents in the supernatant. Conditioned medium and cell lysates were then prepared for calmodulin assay by heating at 95OC for 5 min followed by immediate cooling on ice. Samples were then

centrifuged at 11 6 0 0 g for 5 min and the heat treated supernatants stored at - 20°C prior to assay of calmodulin. Measurement of calmodulin biological activity was determined using a calmodulin-dependent beef heart cyclic nucleotide phosphodiesterase as described by Mac Neil et a1 (1984b). The interassay coefficient of variation is 17.7% for this assay. Immunoreactive levels of calmodulin were measured using a commercially available radioimmunoassay. Calmodulin values were expressed as ng/ml or pg/106 cells. Cell proliferation. Cell proliferation was determined by measuring [ 3H]thymidine into DNA as previously described (Mac Neil et al, 1984b) and by determining cell number. This was done in two ways, by detaching cells with collagenase and counting them using a haemocytometer and also on occasion by counting the number of adherent cells per mmz of the tissue culture plates using a graticule eye-piece. Bradford protein assay. Measurements of protein were carried out using the Bradford protein assay (Bradford, 1976). Lrictate dehydrogenase activity. Release of the intracellular enzyme lactate dehydrogenase (LDH) was determined as a n index of cell damage. Media were collected, centrifuged at 1000 g for 1 0 min to remove cells and stored at 4°C until assayed within 5 d using lactate dehydrogenase kits. Deliberate progressive lysis of the cells gave increasing levels of LDH in the medium. Preparation of pure calmodulin. Pig brain calmodulin was prepared by the method of Kakiuchi et al (1981). Preparations were over 95% pure as determined by SDS/polyacrylamide-gel electrophoresis (Laemli, 1 970). Extraction ofcalmodulin-like activity. A volume (100 ml) of conditioned medium was heated to 95°C for 5 min then cooled on ice. After centrifugation at 2500 g at 4°C for 6 0 min the protein present in the supernatant was precipitated with the addition of 4 g trichloroacetic acid per 1 0 0 ml sample. The precipitate was mixed for 2 h before centrifugation at 2 5 000 g for 30 min at 4°C and the pellet redissolved at neutral pH. After a final centrifugation of 2 5 000 g at 4°C for 6 0 min, the supernatant was ready for addition to a phenyl sepharose column at 4OC. The column was regenerated and pre-equilibrated using 50 mM Tris/HCl pH 7.4. The sample was then loaded and eluted using 40 mM Tris/HCl with 5 mm CaCI2 at pH 7.4. Protein present in the eluate was measured using the Bradford (1976) method. When there were negligible levels of protein present in the fractions the column was eluted with 40 mm Tris/HCI with 2 mM EGTA at pH 7.4 until again the levels of protein in the fractions were negligible. The EGTA present will cause protein bound to the column to be eluted. Fractions containing the bound protein were pooled and dialysed against 50 mM Tris/HC1 at pH 7 before being freeze dried. The sample was then reconstituted and examined for calmodulin-like activity. Preparation of W7. W7 agarose. agarose and anticalmodulin IgG. A 10ml stock of W7 was prepared in dimethylsulphoxide (DMSO). aliquoted and stored at -20°C. This stock was diluted in serum free medium to give the final concentration required. We have confirmed that the highest concentration of DMSO present (1%)did not inhibit cell proliferation. W7 agarose and control agarose beads were washed twice in a

Calmodulin and Endothelial Proli,feration B

A

r

14 C

l4 13 12 11 10

153

--

13 12

11 10

-

' '

9 -

9 1

8 -

8'

7 -

7 -

6 .

6 '

-

5 '

5'

4 =

2r 1

0

AI 0 1

3 -

2 '

1

10 pI extract

100

1 -

0-

I

1

1

10

100

1000

CaM (ng/incubation)

L

I

0 1

m

I

10

100

W7

(pM)

Fig 1, Calmodulin-like activity in endothelial cell conditioned media taken from subconfluent proliferative cells. Fig 1A shows activation of calmodulin-dependentbeef heart phosphodiesterase (as described in Materials and Methods) by pure calmodulin ( 0 )and by heat-treated cell conditioned media ( 0 ) .Fig 1B shows inhibition of calmodulin-like activity in this cell-conditioned media by the addition of the calmodulin antagonist W7. Half-maximal inhibition of activity was achieved with 30 PM W7. Results are meanskSEM of triplicate determinations.

large excess of serum-free medium, followed by centrifugation at 1000 g for 10 min. This was carried out to remove any free W7 and the 0.02% thimerosal preservative present in the W 7 agarose stock. Control non-coupled agarose beads were prepared and used similarly. The anticalmodulin IgG contained a low percentage of azide. We confirmed that the concentration of azide present in the highest concentration of antibody used (O.OOOl2%) did not inhibit cell proliferation. Statistics. Student's t-test was used to determine statistical significance and values of P