Atherosclerosis, 91 (1991) 77-83 6 1991 Elsevier Scientific Publishers
ATHERO
77 Ireland,
Ltd. All rights reserved
0021-9150/91/$03.50
04717
Human arterial endothelial cell detachment in vitro: its promotion by homocysteine and cysteine N.P.B. Dudman, C. Hicks *, J. Wang and D.E.L. Wilcken Department of Medicine, The Prince Henry Hospital, Unillersity of New South Wales, Little Bay (Sydney), NS W 2036 (Australia) (Received 21 May, 3991) (Revised, received 31 July, 1991) (Accepted 5 August, 1991)
Summary
We have explored earlier evidence that premature atherosclerosis in homocystinuria is triggered by homocysteine-induced loss of vascular endothelium. We used a reproducible sluicing assay to test in vitro detachment of human arterial endothelial cells. Cell detachment was induced by exposure of cultured endothelial cells to the sulphydryl-containing amino acids homocysteine and cysteine, whereas methionine, alanine, valine and isoleucine at comparable concentrations were ineffective. This cellular detachment was greatly diminished by growth of the endothelial cells on fibronectin coated- rather than plain tissue culture dishes. Considerably higher concentrations of homocysteine were required for in vitro effects than are associated with atherogenesis in homocystinuria, and despite the cysteine associated changes, cysteine itself is not known to be related to atherogenesis. These data suggested that in vitro detachment of cultured endothelial cells, induced by sulphydryl-containing amino acids, may have marginal relevance to mechanisms of atherogenesis in homocystinuria.
Key words: Vascular disease; assay; Fibronectin
Endothelial
cell detachment;
Introduction
Initiation of the precocious atherosclerosis found in patients with homocystinuria [l] has
* Current address: Department of Haematology, University of New South Wales, The Prince of Wales Hospital, Randwick (Sydney), New South Wales, 2033, Australia. Correspondence to: Dr. Nicholas P.B. Dudman, Department of Medicine, The Prince Henry Hospital, Little Bay (Sydney), New South Wales, 2036, Australia. Tel. 02 694 5694; Fax: 02 311 3483.
Atherogenesis;
Homocystinuria;
Sluicing
been ascribed to detachment of vascular endotheha1 cells caused by endogenous homocysteine, a sulphydryl-containing amino acid abundant in these patients’ plasma ([2]; baboons in this study received homocysteine thiolactone; L.A. Harker, personal communication). This proposal is supported by two major experimental findings. First, when baboons were infused with homocysteine (as the thiolactone), substantial areas of vascular endothelium became desquamated, followed by atherogenic changes in the underlying vascular walls [2,3]. Secondly, experiments in which cul-
78 tured venous endothelial cells were exposed to 2.5-10 mmol/l o,i_-homocysteine (again as the thiolactone) resulted in a major decrease in the adhesivity of the cells to their substrate of tissue culture plastic [4,5]. Adhesivity was assessed by counting cells which had remained attached following application of a jet of saline aimed at the cells [5]. Features of the in vitro evidence, as discussed below, suggested to us that homocysteine-induced desquamation of cultured endothelial cells, under the conditions described, might not be a valid model for the homocysteine-related atherogenesis found in homocystinuria. These difficulties, together with our view that endothelial desquamation in the baboon experiments could have been caused by compounds other than homocysteine [6], led us to re-examine the evidence for the homocysteine-induced endothelial desquamation hypothesis, as we now report. In the present experiments we have defined more fully the effects of homocysteine and homocysteine thiolactone on cultured human vascular endothelial cells, using a modification of the original saline jet technique for determining cell adhesivity. The specificity of these effects was probed using cysteine, as well as several nonsulphydryl amino acids. We also studied whether endothelial cells cultured on a bed of human fibronectin, rather than on the previously used tissue culture plastic, would respond similarly towards homocysteine exposure. Materials
and methods
Endothelial cells were obtained from human umbilical arteries, cultured, and identified as already reported [7]. o,L-Homocysteine, L- and D,i_-homocysteine thiolactone hydrochloride, D,Lhomocystine, and L-cysteine were provided by the Sigma Chemical Co., (MO, U.S.A.). Human serum fibronectin was prepared and used for coating tissue culture plastic surfaces as described previously [8]. Detachment assays
Cells were plated into 35-mm petri dishes (Lux, type 5221, from Flow Laboratories) at approximately 2 x 10’ per dish in 2.0 ml medium, using 6
petri dishes for each level of additive in the dose-response assays, and 3 dishes at each time point for each treatment in the kinetic progress curve assays. Following the methods of Harker et al. [2,4,5], we plated cells directly onto tissue culture plastic in preparation for these assays, unless fibronectin-coated dishes were used as stated below. After incubation for 24 h, the medium was replaced by freshly prepared medium containing either homocysteine or its thiolactone, or cysteine, or another amino acid. After a further 24 h incubation the dishes were each sluiced twice essentially as described previously [9], using a specially designed nozzle connected to an Oxford model 470A automatic dispenser, which delivered 18 ml isotonic saline over 1.15 set through 21 equidistant 20-gauge hypodermic needles. Then 1 ml of saline was gently layered over the remaining attached cells in the dish until they were trypsinised, resuspended in Isoton solution and counted using a Coulter counter model ZB. The number of cells dislodged was proportional to the volume of saline delivered, and to the speed of delivery [9]. Spontaneous detachment was accounted for by trypsinising and counting cells which had not been treated with sulphydryl-containing amino acids nor sluiced; these counts were used as the 100% controls. Controls were also run to evaluate the effect of homocysteine and cysteine on cellular growth rate, in experiments in which cell detachment was measured following exposure to these compounds. In experiments in which the changes in ease of detachment over a 48 h period were observed, 24 h after cells were plated the medium was replaced with fresh medium containing homocysteine or cysteine. Some petri dishes were sluiced immediately after the addition of this fresh medium, and others were sluiced at timed intervals up to 48 h after the medium change. Twenty-four h after the initial exposure of cells to medium containing homocysteine or cysteine, a further medium change with fresh homocysteine or cysteine was made. The viability of cells which had become detached from their dishes during incubation with homocysteine or cysteine was tested as follows.
79 The cells were resuspended by gentle shaking, removed in medium, and centrifuged, and then either sampled for counting or resuspended in DMEM and incubated in fibronectin-coated petri dishes for 24 h. The dishes were then washed gently with isotonic saline, after which any cells which had remained reattached were trypsinised and counted. For our study of the autooxidation of homocysteine and cysteine in DMEM, solutions of homocysteine were prepared immediately before use by dissolving either or_- or L-homocysteine thiolactone hydrochloride (15.4 mg) in 1.0 ml of N,bubbled NaOH solution (0.5 mol/l>. The solution was allowed to stand at 25 OC under N, for 7.0 min, and then neutralised with 1.0 ml N,-bubbled 0.44 mol/l HCI. L- and o-cysteine were prepared as their hydrochloride solutions in N,-bubbled H,O immediately before use. Medium was prepared by dissolving a 1-litre sachet of DMEM powder (Product 10-331-20 from Flow Laboratories) in H,O, adjusting the pH to 7.4 with sodium carbonate buffer, and adjusting the volume to 1 litre with H,O. Medium (84 ml) was then mixed with 16 ml of either phosphate buffered saline (PBS) or fetal calf serum. The autooxidation assays were started by adding 0.1 ml homocysteine or cysteine to 0.9 ml of medium plus PBS or to medium plus serum, at 37°C and incubating at that temperature. Aliquots (50 ,ul> of these reaction mixtures were quenched after various periods by addition to 9.95 ml 5,5’-dithiobis-(2-nitrobenzoic acid) (0.2 mmol/l) in sodium borate (0.05 mol/l), and the absorbance at 412 nm (c = 141.50) was measured at once. We used Student’s unpaired r-test to assess the significance of the differences between treated and untreated cells. Results
Effects of homocysteine Exposure of cells to various concentrations of o,L-homocysteine for 24 h before sluicing resulted in homocysteine-dependent cell detachment. Fig. 1 shows the results in the form of a dose-response curve; D,L- homocysteine promoted cell detachment especially at concentrations above 1 mmol/l. Microscopic examination
M
.c .u ;
80
n
z
0
II
0
/I
0
d
-2
.3
-.I
Log concentration of D,L-homocysteine (mid/L) Fig.
1. Dose response
for endothelial
exposed to various concentrations
h and then sluiced. Points represent percentage ing.
of cells remaining
Points
mmoi/l tached. r = K/(K
with
values
homocysteine The +
broken
cells which have been
of r&L-homocysteine
in petri
significantly
value
dishes following from
f-test
probability
have their
required
H is log[homocysteine
the
0 at-
curve of the form
r is the proportion
moved, K is log[homocysteine
sluic-
different
line is a theoretical
H),where
for 24
the mean (+ SD, n = 6)
of cells not re-
for 50% removal], and
in the medium].
showed visible detachment of the cells after exposure to 5 or 10 mmol/l homocysteine for 24 h, though ceils exposed to lower concentrations of homocysteine showed no visible changes. This experiment was performed 4 times, using cells from 2 different cords, and in each case o&-homocysteine promoted cell detachment. The mean f SD homocysteine concentration for 50% detachment was 3.8 + 1.7 mmol/l. The proportion of cells removed by homocysteine with sluicing achieved statistical significance (P < 0.05) at homocysteine concentrations of 0.5 mmol/l and above, in Fig. 1. o,r_-Homocysteine at concentrations up to 10 mmol/l had no effect on total cell numbers present in the dishes after 24 h exposure. A 48 h study of the detachment kinetics of cells in DMEM containing 5 mmol/l o,L-homocysteine (Fig. 2) showed that exposure to homocysteine followed by sluicing caused steady and progressive cell loss, rather than rapid or synchronous loss of cells. The effects of D,L-homocysteine and D&-homocysteine thiolactone were compared by incubating cells for 24 h in various concentrations of the 2 compounds before sluicing and counting. The dishes of cells containing a final concentra-
ATHERO
77 Ireland,
Ltd. All rights reserved
0021-9150/91/$03.50
04717
Human arterial endothelial cell detachment in vitro: its promotion by homocysteine and cysteine N.P.B. Dudman, C. Hicks *, J. Wang and D.E.L. Wilcken Department of Medicine, The Prince Henry Hospital, Unillersity of New South Wales, Little Bay (Sydney), NS W 2036 (Australia) (Received 21 May, 3991) (Revised, received 31 July, 1991) (Accepted 5 August, 1991)
Summary
We have explored earlier evidence that premature atherosclerosis in homocystinuria is triggered by homocysteine-induced loss of vascular endothelium. We used a reproducible sluicing assay to test in vitro detachment of human arterial endothelial cells. Cell detachment was induced by exposure of cultured endothelial cells to the sulphydryl-containing amino acids homocysteine and cysteine, whereas methionine, alanine, valine and isoleucine at comparable concentrations were ineffective. This cellular detachment was greatly diminished by growth of the endothelial cells on fibronectin coated- rather than plain tissue culture dishes. Considerably higher concentrations of homocysteine were required for in vitro effects than are associated with atherogenesis in homocystinuria, and despite the cysteine associated changes, cysteine itself is not known to be related to atherogenesis. These data suggested that in vitro detachment of cultured endothelial cells, induced by sulphydryl-containing amino acids, may have marginal relevance to mechanisms of atherogenesis in homocystinuria.
Key words: Vascular disease; assay; Fibronectin
Endothelial
cell detachment;
Introduction
Initiation of the precocious atherosclerosis found in patients with homocystinuria [l] has
* Current address: Department of Haematology, University of New South Wales, The Prince of Wales Hospital, Randwick (Sydney), New South Wales, 2033, Australia. Correspondence to: Dr. Nicholas P.B. Dudman, Department of Medicine, The Prince Henry Hospital, Little Bay (Sydney), New South Wales, 2036, Australia. Tel. 02 694 5694; Fax: 02 311 3483.
Atherogenesis;
Homocystinuria;
Sluicing
been ascribed to detachment of vascular endotheha1 cells caused by endogenous homocysteine, a sulphydryl-containing amino acid abundant in these patients’ plasma ([2]; baboons in this study received homocysteine thiolactone; L.A. Harker, personal communication). This proposal is supported by two major experimental findings. First, when baboons were infused with homocysteine (as the thiolactone), substantial areas of vascular endothelium became desquamated, followed by atherogenic changes in the underlying vascular walls [2,3]. Secondly, experiments in which cul-
78 tured venous endothelial cells were exposed to 2.5-10 mmol/l o,i_-homocysteine (again as the thiolactone) resulted in a major decrease in the adhesivity of the cells to their substrate of tissue culture plastic [4,5]. Adhesivity was assessed by counting cells which had remained attached following application of a jet of saline aimed at the cells [5]. Features of the in vitro evidence, as discussed below, suggested to us that homocysteine-induced desquamation of cultured endothelial cells, under the conditions described, might not be a valid model for the homocysteine-related atherogenesis found in homocystinuria. These difficulties, together with our view that endothelial desquamation in the baboon experiments could have been caused by compounds other than homocysteine [6], led us to re-examine the evidence for the homocysteine-induced endothelial desquamation hypothesis, as we now report. In the present experiments we have defined more fully the effects of homocysteine and homocysteine thiolactone on cultured human vascular endothelial cells, using a modification of the original saline jet technique for determining cell adhesivity. The specificity of these effects was probed using cysteine, as well as several nonsulphydryl amino acids. We also studied whether endothelial cells cultured on a bed of human fibronectin, rather than on the previously used tissue culture plastic, would respond similarly towards homocysteine exposure. Materials
and methods
Endothelial cells were obtained from human umbilical arteries, cultured, and identified as already reported [7]. o,L-Homocysteine, L- and D,i_-homocysteine thiolactone hydrochloride, D,Lhomocystine, and L-cysteine were provided by the Sigma Chemical Co., (MO, U.S.A.). Human serum fibronectin was prepared and used for coating tissue culture plastic surfaces as described previously [8]. Detachment assays
Cells were plated into 35-mm petri dishes (Lux, type 5221, from Flow Laboratories) at approximately 2 x 10’ per dish in 2.0 ml medium, using 6
petri dishes for each level of additive in the dose-response assays, and 3 dishes at each time point for each treatment in the kinetic progress curve assays. Following the methods of Harker et al. [2,4,5], we plated cells directly onto tissue culture plastic in preparation for these assays, unless fibronectin-coated dishes were used as stated below. After incubation for 24 h, the medium was replaced by freshly prepared medium containing either homocysteine or its thiolactone, or cysteine, or another amino acid. After a further 24 h incubation the dishes were each sluiced twice essentially as described previously [9], using a specially designed nozzle connected to an Oxford model 470A automatic dispenser, which delivered 18 ml isotonic saline over 1.15 set through 21 equidistant 20-gauge hypodermic needles. Then 1 ml of saline was gently layered over the remaining attached cells in the dish until they were trypsinised, resuspended in Isoton solution and counted using a Coulter counter model ZB. The number of cells dislodged was proportional to the volume of saline delivered, and to the speed of delivery [9]. Spontaneous detachment was accounted for by trypsinising and counting cells which had not been treated with sulphydryl-containing amino acids nor sluiced; these counts were used as the 100% controls. Controls were also run to evaluate the effect of homocysteine and cysteine on cellular growth rate, in experiments in which cell detachment was measured following exposure to these compounds. In experiments in which the changes in ease of detachment over a 48 h period were observed, 24 h after cells were plated the medium was replaced with fresh medium containing homocysteine or cysteine. Some petri dishes were sluiced immediately after the addition of this fresh medium, and others were sluiced at timed intervals up to 48 h after the medium change. Twenty-four h after the initial exposure of cells to medium containing homocysteine or cysteine, a further medium change with fresh homocysteine or cysteine was made. The viability of cells which had become detached from their dishes during incubation with homocysteine or cysteine was tested as follows.
79 The cells were resuspended by gentle shaking, removed in medium, and centrifuged, and then either sampled for counting or resuspended in DMEM and incubated in fibronectin-coated petri dishes for 24 h. The dishes were then washed gently with isotonic saline, after which any cells which had remained reattached were trypsinised and counted. For our study of the autooxidation of homocysteine and cysteine in DMEM, solutions of homocysteine were prepared immediately before use by dissolving either or_- or L-homocysteine thiolactone hydrochloride (15.4 mg) in 1.0 ml of N,bubbled NaOH solution (0.5 mol/l>. The solution was allowed to stand at 25 OC under N, for 7.0 min, and then neutralised with 1.0 ml N,-bubbled 0.44 mol/l HCI. L- and o-cysteine were prepared as their hydrochloride solutions in N,-bubbled H,O immediately before use. Medium was prepared by dissolving a 1-litre sachet of DMEM powder (Product 10-331-20 from Flow Laboratories) in H,O, adjusting the pH to 7.4 with sodium carbonate buffer, and adjusting the volume to 1 litre with H,O. Medium (84 ml) was then mixed with 16 ml of either phosphate buffered saline (PBS) or fetal calf serum. The autooxidation assays were started by adding 0.1 ml homocysteine or cysteine to 0.9 ml of medium plus PBS or to medium plus serum, at 37°C and incubating at that temperature. Aliquots (50 ,ul> of these reaction mixtures were quenched after various periods by addition to 9.95 ml 5,5’-dithiobis-(2-nitrobenzoic acid) (0.2 mmol/l) in sodium borate (0.05 mol/l), and the absorbance at 412 nm (c = 141.50) was measured at once. We used Student’s unpaired r-test to assess the significance of the differences between treated and untreated cells. Results
Effects of homocysteine Exposure of cells to various concentrations of o,L-homocysteine for 24 h before sluicing resulted in homocysteine-dependent cell detachment. Fig. 1 shows the results in the form of a dose-response curve; D,L- homocysteine promoted cell detachment especially at concentrations above 1 mmol/l. Microscopic examination
M
.c .u ;
80
n
z
0
II
0
/I
0
d
-2
.3
-.I
Log concentration of D,L-homocysteine (mid/L) Fig.
1. Dose response
for endothelial
exposed to various concentrations
h and then sluiced. Points represent percentage ing.
of cells remaining
Points
mmoi/l tached. r = K/(K
with
values
homocysteine The +
broken
cells which have been
of r&L-homocysteine
in petri
significantly
value
dishes following from
f-test
probability
have their
required
H is log[homocysteine
the
0 at-
curve of the form
r is the proportion
moved, K is log[homocysteine
sluic-
different
line is a theoretical
H),where
for 24
the mean (+ SD, n = 6)
of cells not re-
for 50% removal], and
in the medium].
showed visible detachment of the cells after exposure to 5 or 10 mmol/l homocysteine for 24 h, though ceils exposed to lower concentrations of homocysteine showed no visible changes. This experiment was performed 4 times, using cells from 2 different cords, and in each case o&-homocysteine promoted cell detachment. The mean f SD homocysteine concentration for 50% detachment was 3.8 + 1.7 mmol/l. The proportion of cells removed by homocysteine with sluicing achieved statistical significance (P < 0.05) at homocysteine concentrations of 0.5 mmol/l and above, in Fig. 1. o,r_-Homocysteine at concentrations up to 10 mmol/l had no effect on total cell numbers present in the dishes after 24 h exposure. A 48 h study of the detachment kinetics of cells in DMEM containing 5 mmol/l o,L-homocysteine (Fig. 2) showed that exposure to homocysteine followed by sluicing caused steady and progressive cell loss, rather than rapid or synchronous loss of cells. The effects of D,L-homocysteine and D&-homocysteine thiolactone were compared by incubating cells for 24 h in various concentrations of the 2 compounds before sluicing and counting. The dishes of cells containing a final concentra-
