Apoptosis and adaptive responses to oxidative ... - The FASEB Journal

Apoptosis and adaptive responses to oxidative stress in human endothelial cells exposed to cyclosporin A correlate with BCL-2 expression levels BIANCAMARIA LONGONI,*,†,1 ELENA BOSCHI,* GIAN CARLO DEMONTIS,‡ GIAN MICHELE RATTO,§ AND FRANCO MOSCA* Departments of *Oncology, Transplants and Advanced Technologies in Medicine, †Physiology and Biochemistry, ‡Psychiatry and Neurobiology, University of Pisa, School of Medicine; and §Institute for Neurophysiology, CNR, Pisa, Italy Treatment of transplanted patients with cyclosporin A (CSA) may cause adverse effects such as nephrotoxicity and hypertension. As CSA is known to induce oxidative stress in several tissues, it may cause vascular problems by triggering oxidative stress in endothelial cells (EC). However, oxidative stress has been reported for acute exposure to CSA concentrations exceeding its clinical range, whereas immunosuppression requires life-long treatment with therapeutic concentrations. We therefore compared the effects of 21 h pharmacological (200 ␮M) vs. 8 days clinical (0.5–2.5 ␮M) doses of CSA on cultured human EC. Pharmacological doses of CSA cause a decrease in cell density via apoptosis and a down-regulation of the antiapoptotic protein Bcl-2. However, these effects are independent of CSA-induced oxidative stress. In contrast, therapeutic concentrations of CSA cause Bcl-2 up-regulation and modification of EC morphology, both effects blocked by antioxidants. Therefore, a low level of oxidants may act in EC as second messengers that up-regulate Bcl-2, thus promoting survival of impaired EC. Our data suggest that the oxidative stress induced by clinical concentrations of CSA may be involved in the adverse effects of the drug on the vascular system of transplanted patients via an adaptive response involving Bcl-2 up-regulation rather than an apoptotic process.—Longoni, B., Boschi, E., Demontis, G. C., Ratto, G. M., Mosca, F. Apoptosis and adaptive responses to oxidative stress in human endothelial cells exposed to cyclosporin A correlate with BCL-2 expression levels. FASEB J. 15, 731–740 (2001)

ABSTRACT

Key Words: confocal microscopy 䡠 reactive oxygen species 䡠 antioxidants 䡠 dihydrorhodamine 123 䡠 melatonin 䡠 transplantation

Cyclosporin A (CSA) is an efficient immunosuppressive agent of significant clinical importance due to its widespread use in organ transplantation (1). However, severe side effects have been found to be associated with life-long treatment of transplanted patients with CSA, including renal and vascular toxicity (2– 4). The observation that CSA causes oxidative stress in 0892-6638/01/0015-0731 © FASEB

rat hepatocytes (5) and that antioxidants such as vitamin E, melatonin, and other indolic compounds inhibit lipid peroxidation in animal models of cyclosporin A-induced nephrotoxicity (6 –9) suggests a role for oxidative stress in CSA-induced damage. As the endothelium is known to influence vascular tone (10, 11) and is considered to be the main site of impaired vascular regulation by CSA (12–15), endothelial cell (EC) damage by oxidative stress may play a key role in the vascular problems of transplanted patients treated with CSA. Accordingly, we have recently shown that CSA induces a concentration-dependent increase in oxidants and lipid peroxidation in cultured human endothelial cells (16). Considering the well-known relationship between oxidative stress and apoptosis (17, 18), CSA-induced oxidative stress may damage EC via apoptotic cell death. However, oxidative stress-induced apoptosis occurs for acute exposure to CSA concentrations outside its therapeutic range (19, 20), and it is presently unclear whether it may play a causal role in the vascular problems of patients chronically treated with therapeutic doses of CSA. The main aim of the present work was to provide a better understanding of the pathogenesis of CSA-induced vascular toxicity by comparing its acute (high levels for short time) vs. chronic (low levels for long time) effects on oxidative stress and apoptosis in human EC. Our study suggests that acute treatment with CSA is associated with apoptosis and with a decrease of Bcl-2 protein expression in EC. On the other hand, chronic treatment with therapeutic levels of CSA affects EC via mechanisms that involve adaptation to CSA-induced oxidative stress through Bcl-2 up-regulation, which may have a key role in promoting the survival of EC with an impaired function, as suggested by their altered morphology. 1 Correspondence: Department of Oncology, Transplants, and Advanced Technologies in Medicine, University of Pisa, Via Paradisa 2, I-56124 Pisa, Italy. E-mail: biancam@dfb. unipi.it

731

MATERIALS AND METHODS Human umbilical vein endothelial cell (HUVEC) cultures HUVECs were isolated from human umbilical veins as described (21) with minor modifications. After collection, umbilical cords were placed in cold saline solution containing penicillin 25 U/ml and streptomycin 50 ␮g/ml. The umbilical veins were cannulated and perfused with saline solution to wash out the residual blood. Both ends of the cord were clamped and then infused with saline containing 0.25% type IA collagenase. After 10 min incubation at 37°C, the collagenase solution containing detached HUVECs was flushed out of the cord with 10 ml of Medium 199. The cells were then centrifuged at 100 g and resuspended in complete Medium 199 containing 20 mM HEPES buffer, 20% heat-inactivated fetal calf serum, 0.2% growth factor for endothelial cells, 2 mM glutamine, 5 U/ml heparin, 25 U/ml penicillin, and 50 ␮g/ml streptomycin. HUVECs were cultured overnight in 25 cm2 gelatin-coated tissue culture flasks (Corning, Corning, N.Y.) at 37°C in a NAPCO incubator (5% CO2, 95% air). The next day contaminating cells were rinsed out of the flask and 9 ml of complete medium was added. At confluence, HUVECs were trypsinized and plated into three flasks for an additional passage. HUVECs were used for experiments in the third subpassage. HUVECs were identified by their typical cobblestone appearance and positive indirect immunofluorescence for factor VIII. All reagents used for cell dissociation and culture were obtained from Sigma (Milan, Italy). Phase-contrast microscopy For phase-contrast microscopy, cells were cultured on 75 cm2 culture flasks (Greiner) and exposed to experimental conditions. Phase-contrast microscopic examination was performed using an inverted Zeiss Televal 3 (light source 25 W halogen lamp) equipped with a Nikon camera. Cell density was evaluated by the trypan blue exclusion procedure. The cell suspension was diluted 1:2 with the dye (trypan blue 0.4%, Sigma) prior to counting of viable cells with a hemocytometer. Monitoring of reactive oxygen species production For microscopic observation, HUVECs were plated on gelatincoated glass coverslips until they reached subconfluence. The glass coverslip was adapted to a Plexiglas chamber and HUVECs were loaded with 10 ␮M dihydrorhodamine 123 (DHR123) (Molecular Probes, Eugene, Oreg.) for 30 min at 37°C in culture medium. The chamber was positioned for imaging on the stage of an inverted microscope (Nikon Eclipse TE300) equipped with a laser confocal scanning system (Radiance Plus, Bio-Rad, Hercules, Calif.). The stock solution (10 mM) of dihydrorhodamine 123 in DMSO contained less than 1% of contaminating rhodamine 123 (RHD123), as assessed by spectrofluorometry using a PerkinElmer LB50 Fluorescence Spectrometer. Contamination of rhodamine 123 present in traces in the dihydrorhodamine stock was responsible for the low background fluorescence observed at the beginning of the recording session, which did not change in untreated cells over a period of 2 h. Cyclosporin A was added after 15 min of observation (baseline); 2 ␮M SNARF (Molecular Probes) was also included in some experiments to assess cell morphology and viability. Acridine orange (Molecular Probes) was used at 10 ␮g/ml as a membrane permeant and an intercalating agent to label nucleic acids. Fluorescence images were collected every 3 min using the 732

Vol. 15

March 2001

488 nm excitation wavelength from an argon laser and the 515 nm emission filter. To minimize photo-oxidation of the probe, the laser beam was attenuated to 50% of maximal illumination and exposure of cells to light was limited to the image acquisition intervals (⬃2 s every 3 min) via the acquisition software. Analysis of the fluorescent signal was performed with Adobe Photoshop 5.0 (Adobe, San Jose, Calif.) for each cell and for each period of time over a similar region of interest, which included most of the cells. To compensate for the pronounced variability of fluorescence from cell to cell (see Results, Fig. 1A, and Fig. 6A) that was already present in controls, for each cell the effect of CSA on fluorescence was expressed as the relative increase over the basal value at the beginning of recording. In control experiments, without dihydrorhodamine 123, CSA did not affect the low autofluorescence of HUVECs. Lactic dehydrogenase (LDH) assay LDH release was measured as the enzymatic activity present in the cell medium. Media were collected from cell cultures at appropriate times and centrifuged to remove contaminating cells and debris. LDH activity was assessed by standard enzymatic methods (22). Lipid peroxidation assay Lipid peroxidation was evaluated by using a commercial colorimetric kit by Calbiochem (San Diego, Calif.) for malonaldehyde and 4-hydroxyalkenals (MDA⫹4-HNA). MDA ⫹ 4-HNA production was determined at different time points in control and CSA-treated cell cultures in accordance with the manufacturer’s instructions. DNA extraction and gel electrophoresis Endothelial cells (107 cells/sample) were washed with NaCl 0.9%, trypsinized, and harvested by centrifugation (100 g for 10 min). The pellet was washed with NaCl 0.9% and resuspended in 500 ␮l of digestion buffer (100 mM NaCl, 10 mM Tris-HCl pH 8, 25 mM Na2EDTA pH 8, 0.5% SDS, 0.1 mg/ml proteinase K, and 1% 2-mercaptoethanol). After overnight incubation at 50°C, the samples were extracted twice with phenol/chloroform/isoamyl alcohol (25/24/1) and centrifuged at 10,000 g for 10 min. DNA in the aqueous phase was precipitated by adding to the sample 0.5 volume of 7.5 M ammonium acetate and 2 volumes of ethanol, and recovered by centrifuging at 10,000 g for 2 min. The pellet was rinsed with 70% ethanol and air dried, and DNA was resuspended in TE buffer (10 mM Tris-HCl pH 8, 1 mM Na2EDTA pH 8). Residual RNAs were removed with 1 ␮g/ml DNase-free RNase-A by incubating for 1 h at 37°C. DNA was resolved in 1% agarose gel in the presence of 2 ␮g/ml ethidium bromide and visualized with UV transillumination. Photographs were directly taken with a Polaroid M4 camera. Bcl-2 and nucleosome ELISA Human Bcl-2 protein and nucleosomes were quantified in HUVEC extracts by using the Bcl-2 and the nucleosome ELISA kits from Oncogene Research Products (Cambridge, Mass.) in accordance with the manufacturer’s instructions. Readings of colored products were taken by using a spectrophotometer plate reader (Titertek Multiskan MCC) and compared with the standard curves. Values were expressed as units of Bcl-2 or nucleosomes/mg of protein.

The FASEB Journal

LONGONI ET AL.

Protein measurement Protein concentration was determined in accordance with the method of Lowry et al. (23). CSA application CSA was obtained either as a powder from Alexis Corporation (La¨ufelfingen, Switzerland) or as a solution in castor oil (Novartis, Pharma, Basel, Switzerland). CSA powder was dissolved in ethanol at 10 mM and brought to final concentrations in Medium 199. The CSA formulation in castor oil was dissolved in Medium 199. Similar effects were obtained with both formulations of CSA; control experiments with ethanol and castor oil ruled out a possible effect of the vehicle. As CSA binds to plasma proteins in vivo, the free CSA concentration ([CSA]) in serum-containing Medium 199 is expected to be lower than the analytical concentration (Ctot). Determination of [CSA] in Medium 199 was performed with an FPIA kit (Fluorescence Polarisation Immunoassay) from Abbott Laboratories (Abbott Park, Ill.). The measurement of [CSA] over a range of Ctot values indicates that [CSA]/Ctot ⬵ 0.5; in other words, the effective CSA concentration is approximately half Ctot. Concentrations of CSA are reported here as free concentrations ([CSA]). Statistical analysis Measurements are reported as mean ⫾ standard error of the mean (se). Analysis of HUVEC fluorescence shows the presence of cells with large differences in intensity, even in basal conditions. Accordingly, plots of average cellular fluorescence result in a bimodal distribution (data not shown), indicating that HUVECs may not be considered as a homogeneous population normally distributed. Consequently, the statistical significance of the differences in fluorescence observed before and after CSA application was assessed by either nonparametric tests (Mann-Whitney U test) or by 2-way analysis of variance (2-way ANOVA), using tabulated values of probabilities from the F distribution (24).

Figure 1. Generation of ROS in human cultured endothelial cells by CSA. ROS formation was visualized with confocal microscopy imaging of rhodamine 123 fluorescence, as described in Materials and Methods. After collecting baseline images (A), 200 ␮M CSA was added and the image in panel B was acquired 13 min after application. Note the increase in fluorescence for each cell after CSA addition. Average fluorescence in control conditions was 10.3 ⫾ 1.0 counts/pixel (mean⫾se). Note the variability in cellular fluorescence in basal conditions. Data are representative of three independent experiments with similar results.

the drug (P⬍0.005 and P⬍0.001, respectively). Note that despite clear differences in fluorescence intensity between cells already present in control conditions, the average relative increase in fluorescence over basal values (see methods) induced by CSA was similar for all cells, averaging 105.4 ⫾ 11.9% (mean⫾se) after 13 min of exposure to the drug. The increase in cellular oxidants monitored by the oxidation-sensitive probe is associated with an increased release of malonaldehyde and 4-hydroxyalkenals, by-products of lipid peroxidation (25), in the culture medium from 9.6 ⫾ 0.7 pmol ml⫺1 in control conditions to 109.4 ⫾ 21.2 pmol ml⫺1 after 21 h in CSA 200 ␮M (P⬍0.05, n⫽3, paired t test). To correlate CSA-induced oxidative stress with apoptosis, HUVECs were double stained with DHR 123 and

RESULTS Oxidative stress and apoptosis upon acute exposure to CSA The relationship between oxidative stress and apoptosis in HUVECs during acute treatment with pharmacological concentrations of CSA was investigated by laser scanning confocal microscopy. Treatment of HUVECs with 200 ␮M CSA causes an increase in the conversion of the oxidation-sensitive probe dihydro rhodamine 123 to its fluorescent analog rhodamine 123, which suggests an increased production of cellular oxidants in response to CSA. This fluorescence increase is illustrated in Fig. 1B 13 min after application of CSA, when a steady state has been reached, and may be compared to control conditions (Fig. 1A). 2-way ANOVA for the effect of the duration of CSA application vs. control was carried out in 75 cells from three independent experiments and indicated that the increase in fluorescence induced by CSA was statistically significant after both 10 and 15 min of exposure to OXIDATIVE STRESS IN ENDOTHELIAL CELLS

Figure 2. Apoptosis of HUVECs after CSA application. A) Endothelial cells were simultaneously loaded with DHR 123 and SNARF as described in Materials and Methods. Images were pseudocolored (DHR 123 in green and SNARF in red) based on pixel values. Blebbing formation is clearly observable 20 min after 200 ␮M CSA addition, when the DHR 123 signal has already increased, and is indicated by the white arrows. B) Clear signs of nuclear condensation and membrane blebbing in endothelial cells stained with the nuclear indicator acridine orange and exposed to 200 ␮M CSA for 50 min. Images in panels A and B were from different experiments; qualitatively similar results were obtained in two additional experiments. 733

the cytoplasmic indicator SNARF. As illustrated in Fig. 2, when 200 ␮M CSA treatment for 20 min had already caused an increase in rhodamine 123 fluorescence (green), there was the appearance of multiple cytoplasmic (white arrows) membrane blebs, an early sign of apoptosis (Fig. 2A). Additional indications for the occurrence of apoptosis in response to 200 ␮M CSA were assessed by staining the cells with acridine orange, a membrane-permeant and nucleic acid-intercalating probe. After 50 min from the application of 200 ␮M CSA, clear changes in nuclear morphology, including chromatin condensation and lobule formation, were present (Fig. 2B). Note that these early signs of apoptosis were not present simultaneously in every cell. CSA did not cause an acute release of lactic dehydrogenase, suggesting that pharmacological concentrations of CSA trigger an apoptotic rather than a necrotic process in HUVECs. Molecular basis of CSA-induced damage To assess the ability of HUVECs to complete the apoptotic process triggered by 200 ␮M CSA, we performed DNA fragmentation analysis. As illustrated in Fig. 3A, the DNA from control cells appears to be intact (lane 2) whereas treatment with CSA 200 ␮M (lane 3) for 21 h induces DNA degradation, with the ladder-like appearance of the classic nucleosomal-sized DNA fragmentation characteristic of apoptosis (26 –28). Note that simultaneous treatment with 300 ␮M melatonin, a hormone with antioxidant properties (29, 30), fails to protect HUVECs from DNA degradation (lane 4).

To verify the role of oxidants in the proapoptotic effects of CSA, we performed quantitative estimates of DNA fragmentation by measuring the levels of nucleosomes with ELISA. Column bars of Fig. 3B indicate that in HUVECs acutely treated with CSA (200 ␮M for 21 h), nucleosome levels increased from a control value of 0.6 ⫾ 0.1 U/mg protein to 29.3 ⫾ 2.3 U/mg protein. However, nucleosome levels in HUVECs treated simultaneously with 200 ␮M CSA and either 300 ␮M melatonin or 100 ␮M Trolox (a water-soluble analog of vitamin E) were 27.8 ⫾ 1.0 or 27.3 ⫾ 2.0 U/mg protein, significantly higher than control levels (P⬍0.001 for both antioxidants) and not significantly different from CSA alone (1-way ANOVA). The increase in oligonucleosome formation was not significant 2 h after CSA application (data not shown), suggesting that an induction time is required for the activation of enzymes involved in DNA fragmentation. To further characterize the molecular basis of the apoptotic process in consideration of the importance of antiapoptotic proteins, like Bcl-2, we measured the effects of acute CSA treatment on expression of the antiapoptotic protein Bcl-2 in HUVECs. Figure 3C shows that treatment with 200 ␮m CSA for 21 h, which induces ladder-like DNA fragmentation, also induces a decrease in Bcl-2 protein expression from 4.6 ⫾ 0.1 U/mg protein (control) to 1.4 ⫾ 0.1 U/mg protein (treated cells); melatonin did not prevent Bcl-2 from falling to 1.8 ⫾ 0.3 U/mg of protein. These results suggest that pharmacological levels of CSA affect HUVEC survival by altering the balance between pro- and antiapoptotic factors although by

Figure 3. DNA fragmentation and Bcl-2 analysis upon acute CSA treatment. A) DNA laddering visualized by DNA electrophoresis in 1% agarose gel and fluorescent staining by ethidium bromide. Lane 1: 500 bp DNA ladder molecular weight marker; 2: DNA from HUVECs after 21 h incubation in control medium; 3: DNA from HUVECs incubated for 21 h in the presence of CSA 200 ␮M; 4: DNA from HUVECs incubated for 21 h in the presence of CSA 200 ␮M and 300 ␮M melatonin. Note the presence of ladder-like DNA fragmentation in the presence of both CSA and CSA plus melatonin. To compensate for the decreased cell density in CSA, similar amounts of DNA were loaded in lanes 2– 4. B) Column bars indicate nucleosome levels in HUVECs incubated for 21 h in control conditions, in the presence of 200 ␮M CSA and in the presence of 200 ␮M CSA plus 300 ␮M melatonin (MLT) or 200 ␮M CSA plus 100 ␮M Trolox. Note the lack of protection by antioxidants against the DNA degradation induced by CSA. Results are the mean ⫾ se of three different experiments performed in duplicate. Analysis of variance (1-way) indicate that CSA significantly increases (P⬍0.001) nucleosome formation compared with controls, even in the presence of antioxidants. Differences between CSA and CSA plus antioxidants were not statistically significant at the 0.05 level. To compensate for the decreased cell density in the presence of CSA, nucleosomes were normalized to the protein content of the sample in panels B and C. ***P ⬍ 0.001. C) Column bars plot Bcl-2 protein levels in HUVECs incubated for 21 h in control conditions, in the presence of 200 ␮M CSA or 200 ␮M CSA plus 300 ␮M melatonin (MLT). Note the decrease in Bcl-2 levels in the presence of both CSA as well as CSA plus melatonin. Results are the mean of four independent experiments performed in duplicate. **P ⬍ 0.01, one-way ANOVA 734

Vol. 15

March 2001

The FASEB Journal

LONGONI ET AL.

mechanisms independent of the drug’s ability to induce oxidative stress. Effects of acute CSA treatment on endothelial cell density and morphology This idea was further tested by observing the effects of CSA on HUVEC vitality. Treatment of HUVECs with 200 ␮M CSA for 21 h causes a decrease in cell density (Fig. 4B) compared with control cultures (Fig. 4A). Moreover, the acute treatment also affects cellular morphology, with the appearance of round cells (arrows in Fig. 4B) quite different from the typical polygonal shape of HUVECs (Fig. 4A). Simultaneous treatment of cells with CSA and antioxidants such as Trolox (Fig. 4C) or melatonin (Fig. 4D) failed to prevent the effect of acute CSA treatment on HUVEC density. A similar lack of protection has also been noted with other antioxidants, such as ascorbic acid or N-acetylserotonin (data not shown). The failure of antioxidants to afford protection against the effects of acute treatment indicates that CSA triggers apoptotic cell death of HUVECs via an oxidative stress-independent mechanism. Effects of chronic CSA treatment on endothelial cell density and morphology As illustrated in Fig. 5, chronic treatment of HUVECs with CSA (2 ␮M for 8 days) did not appreciably affect cell density (Fig. 5B) compared with control cultures (Fig. 5A). However, HUVECs treated with CSA have an altered morphology (Fig. 5B), with a transition to an elongated

Figure 5. Morphological alterations in cultured endothelial cells exposed to chronic cyclosporin A treatment. Phasecontrast micrographs of human endothelial cells cultured for 8 days in either control medium (A) or medium plus 2.5 ␮M CSA. Note the change in cell morphology from a polygonal to an elongated shape in response to CSA. In this particular experiment, the small differences in cell density of panels B and C with those of panels A and D mirrored the differences in cell density before treatment; in three experiments cell density in CSA 2.5 ␮M, alone or with antioxidants, was not statistically different from controls (P⬎0.05, 1-way ANOVA). Antioxidants like Trolox (C) and vitamin C (D) prevent the change in cell morphology induced by CSA.

shape, when compared with controls (Fig. 5A). These morphological changes induced by chronic treatment with low CSA concentrations were suppressed by simultaneous treatment with Trolox (Fig. 5C) or vitamin C (Fig. 5D). A cobblestone morphology was also preserved with melatonin (data not shown). This indicates that chronic treatment with CSA may cause oxidative stress, which in turn affects the expression of specific proteins involved in maintaining the endothelial phenotype. Oxidative stress on exposure to therapeutic levels of CSA

Figure 4. Growth inhibition and changes in cell morphology induced by acute CSA treatment: effects of Trolox and melatonin. Pictures are phase-contrast micrographs taken 21 h after the addition to confluent human endothelial cell cultures of medium (A); CSA 200 ␮M (B); 200 ␮M CSA plus 100 ␮M Trolox (C); 200 ␮M CSA plus 300 ␮M melatonin (D). Note the decrease in cell density and the morphological alterations (B, arrowheads) and the lack of protection against CSA-induced cell loss by simultaneous treatment with antioxidants. Differences between control and CSA 200 ␮M treated cells were statistically significant (P⬍0.001, one-way ANOVA). OXIDATIVE STRESS IN ENDOTHELIAL CELLS

In agreement with this notion, confocal microscopy of EC exposed to CSA concentrations of clinical importance revealed an increased production of oxidants, albeit to a lower extent than in response to acute treatment. Figure 6 shows the increase in RHD 123 fluorescence in cells treated with 2.5 ␮M CSA for 15 min (Fig. 6B) compared with control conditions (Fig. 6A) before CSA; 2-way ANOVA indicates that the increase in fluorescence is statistically different from controls 10 and 15 min after CSA application. The average relative increase was 37.0 ⫾ 4.1%, calculated over 11 cells. Similar relative increase were measured in 70 cells from three independent experiments. In addition to the increased generation of oxidants monitored by DHR 123, chronic treatment of cells with 2.5 ␮M CSA significantly increased the release of malonaldehyde and 4-hydroxyalkenals, as shown in Table 1. 735

Figure 6. Generation of ROS in human endothelial cells by therapeutic concentrations of CSA. ROS formation was detected by confocal microscopy imaging of the rhodamine 123 fluorescent signal. After collecting the control image (A), 2.5 ␮M CSA was added and the image in panel B was acquired after 15 min in the presence of the drug. Note the increase in fluorescence in the presence of CSA. Average basal level of fluorescence was 9.5 ⫾ 1.1 counts/pixel. Similar results were obtained in two additional experiments.

Despite the generation of oxidants in response to therapeutic levels of CSA (2.5 ␮M), no morphological signs of apoptosis were apparent (data not shown). Moreover, EC chronically treated (8 days) with 0.5–2.5 ␮M CSA did not undergo nucleosomal-size DNA fragmentation, as shown in Fig. 7A. On the other hand, HUVECs surviving the chronic challenge with clinical levels of CSA show an increased expression of the antiapoptotic protein Bcl-2. Figure 7B shows that Bcl-2 protein expression significantly increased in cells exposed to chronic CSA treatment (0.5–2.5 ␮M for 8 days). Simultaneous treatment of EC with the antioxidant melatonin (300 ␮M), which was ineffective in preventing the DNA degradation induced by 200 ␮M CSA (see Fig. 4), suppresses the increase of Bcl-2 protein expression in response to 0.5–2.5 ␮M CSA, suggesting that low levels of oxidants may act as signaling molecules that trigger the adaptation process (Fig. 7B).

Figure 7. DNA fragmentation and Bcl-2 analysis upon treatment with therapeutic CSA concentrations. A) Column bars plot nucleosome levels in HUVECs cultured for 8 days in control medium or in the presence of 0.5 or 2.5 ␮M CSA. Note the lack of nucleosome increase in response to therapeutic concentrations of CSA. Results are the mean ⫾ se of three different experiments performed in duplicate. B) Column bars plot Bcl-2 protein levels in HUVECs cultured for 8 days in control medium or in the presence of CSA 0.5 or 2.5 ␮M, either alone or in the presence of 300 ␮M melatonin. CSA 0.5–2.5 ␮M induced an increase in Bcl-2 levels. Bcl-2 levels were similar to control values when HUVECs were simultaneously treated with CSA and melatonin (MLT) 300 ␮M. Treatment with MLT alone did not change control levels of Bcl-2 protein (data not shown). ***P ⬍ 0.001

Fig. 8 show that HUVEC cultures acutely challenged for 21 h with 200 ␮M CSA have a significantly higher survival rate when chronically treated for 6 days with 2.5 ␮M CSA (Fig. 8C) than cultures not adapted to CSA prior to the acute challenge with the drug (Fig. 8B). Note that the cell density in Fig. 8C was significantly lower than in cultures treated for 6 days with control medium (Fig. 8A) or 2 ␮M CSA (not shown). The cell density mirrored the changes in the level of the Bcl-2 protein. This suggests that the Bcl-2 up-regulation induced by oxidants during the chronic treatment with therapeutic doses of CSA allows the cells to survive the acute challenge with pharmacological doses of CSA.

DISCUSSION Chronic exposure to clinical concentrations of CSA induces adaptation and protects HUVECs against acute challenge with pharmacological CSA concentrations To verify that clinical concentrations of CSA trigger an adaptation process, HUVECs were challenged with 200 ␮M CSA after 6 day exposure to 2.5 ␮M CSA. Data in

The most significant novel finding of this paper is that CSA acts on HUVECs through two independent mechanisms, which operate at different concentrations and durations of exposure to the drug. The first mechanism operates for long exposures to clinical concentrations through an increase in cellular oxidants, which in turn causes a change in cell morphology and an up-regula-

TABLE 1. Malonaldehyde and 4-hydroxyalkenals content of culture medium in nmoles ml⫺1 a Time (days)

Control CSA 2.5 ␮M

0

1

2

4

6

11.7 ⫾ 0.5 9.5 ⫾ 0.4

13.1 ⫾ 0.8 20.4 ⫾ 1.1

14.2 ⫾ 1.2 26.9 ⫾ 5.3

13.2 ⫾ 0.8 37.3 ⫾ 8.7

21.3 ⫾ 6.9 52.8 ⫾ 3.9

a Data report the release by HUVEC of lipid peroxidation by-products in the medium. Medium was collected from control and CSA-treated HUVEC cultures at different time points and analyzed for malonaldehyde and 4-hydroxyalkenals as described in Materials and Methods. Time 0 indicates the medium content of malonaldehyde and 4-hydroxyalkenals just before adding CSA. CSA induces a significant increase in lipid peroxidation over control (P⬍0.05, Paired t test).

736

Vol. 15

March 2001

The FASEB Journal

LONGONI ET AL.

Figure 8. Adaptive responses induced by pretreatment with clinical doses of CSA in endothelial cells. Cells were either exposed to the chronic (6 days) pretreatment with clinical concentrations of CSA (2.5 ␮M) or to the control medium. Phase-contrast micrographs taken at the 8th day of culture show that the acute (21 h) challenge dose (200 ␮M) of CSA induced a clear decrease in cell density in cells pretreated with control medium (B) compared to control cells (A) (P⬍0.001, 1-way ANOVA). Protection from acute challenge is shown in panel C for cells adapted by pretreatment with CSA 2.5 ␮M whose density is intermediate between control and CSA 200 ␮M treated cells not pretreated with CSA 2.5 ␮M for 6 days (P⬍0.001 vs. panels A and B, 1-way ANOVA). D) average Bcl-2 content from three experiments corresponding to the conditions illustrated in panels A–C. Bcl-2 content of panel C is significantly different from that of panel B (*P⬍0.01, 1-way ANOVA).

tion of the antiapoptotic factor Bcl-2, suggesting the adaptation of HUVECs to oxidants. In sharp contrast, the second mechanism operates for short exposures to pharmacological concentrations and causes Bcl-2 down-regulation and apoptotic cell death by mechanisms independent of the concomitant CSA-induced oxidative stress. We will next discuss the role that the properties of either CSA or HUVECs may play in eliciting the spectrum of responses to the drug. We will also discuss the possible clinical significance of adaptation to clinical concentrations of CSA through Bcl-2 up-regulation. Pharmacological doses of CSA triggers oxidative stress-independent apoptosis in HUVECs Pharmacological concentrations of CSA have been observed to cause oxidative stress in renal (31) and hepatic (5) tissues and in glioma cells (19). As oxidative stress is a well-known cause of apoptosis (17, 18), we investigated the ability of pharmacological doses of CSA to induce apoptosis in HUVECs. As shown in Figs. 2 and 3, acute exposure to nonclinical levels of CSA concentrations induces both the morphological and biochemical hallmarks of apoptosis. Thus, CSAinduced apoptosis probably plays a role in the decrease in HUVEC density documented in Fig. 4B. However, the observation that simultaneous treatment with antioxidants does not suppress the biochemical markers of apoptosis (Fig. 3A) or the loss of HUVECs (Fig. 4C, D), while blocking CSA-induced lipid peroxidation (16), suggests that pharmacological concentrations of CSA induce apoptosis by mechanisms that operate independently of the simultaneously occurring oxidative stress. Both CSA and its complexes with cyclophylins (32, 33) are known to exert several biological effects in target cells (for a review, see ref 34), such as modulation of the protein phosphatase calcineurin (35) or of OXIDATIVE STRESS IN ENDOTHELIAL CELLS

the permeability transition pore at the mitochondrial level (36). Considering that CSA has previously been reported to protect cells from apoptosis because of its ability to block the mitochondrial permeability transition pore (37), it seems unlikely that this mitochondrial action is responsible for CSA-induced apoptosis in HUVECs. Rather, the observation that apoptosis is independent of oxidative stress, despite the occurrence of lipid peroxidation in response to acute treatment with pharmacological concentrations of CSA (16), suggests that perhaps CSA protects HUVECs from oxidative stress-induced apoptosis. On the other hand, inhibition of calcineurin (38), a calcium-activated protein phosphatase that plays a key role in the immunosuppressive action of CSA, may have a causal role in CSA-induced apoptosis in HUVECs. Despite the lack of any direct evidence for a specific role of calcineurin in inducing oxidative stress in HUVECs, it seems reasonable to postulate that, in general, the interaction of the CSA/cyclophilins complexes with specific proteins may trigger an apoptotic program independently of oxidative stress. Further studies are required to define the molecular basis underlying the ability of CSA to induce both the increase in cellular oxidants and apoptosis and to elucidate the possible relationships between these two pathways of CSA action in HUVECs. Clinical concentrations of CSA trigger oxidative stress-dependent adaptation in HUVECs The observation that CSA-induced oxidants do not play a role in triggering apoptosis does not mean they are devoid of any effects in HUVECs. Indeed, antioxidants suppress the effects of clinical concentrations of CSA, including the up-regulation of Bcl-2 (Fig. 7B) and the modification of cell morphology from a cobblestone to an elongated shape (Fig. 5B). This suggests that clinical 737

levels of CSA affect HUVECs through mechanisms regulated by oxidants and may also indicate that cellular oxidants are the second messengers mediating the effects of clinical levels of CSA in HUVECs. It is important to note that the increase in oxidants takes place on a time scale of minutes, attaining a steady level as long as CSA is present in the medium, whereas the increase in Bcl-2 levels and the changes in cell morphology occur on a slow time scale, because they are present after 4 days but not after 24 h. This suggests that changes in Bcl-2 levels and cell shape represent a true adaptive response to a chronic increase in oxidants rather than a short-term response to a low level of oxidants. The notion of an adaptive response to oxidants via an increase in Bcl-2 is further supported by the data in Fig. 8, which show an improved cell survival to acute challenge with 200 ␮M CSA after 6 days pretreatment with 2.5 ␮M CSA that increases Bcl-2 levels. Relationship between the level of CSA-induced oxidants, apoptosis and adaptation Previous work with mammalian fibroblasts (see ref 39 for a recent review) indicates that the switch-over between adaptation and apoptosis may occur for increase of external H2O2 between three- and sixfold. Moreover, recent data obtained in bovine EC exposed to carbon monoxide (CO) indicate that either adaptation or apoptosis develops within a less than 10-fold increase in environmental CO levels (40). These data suggest that EC share with other mammalian cells the ability to trigger either an apoptotic or an adaptive response, the switch operating over a narrow range of concentrations of either oxidants or oxidant generators. According to measurements based on laser scanning confocal microscopy, CSA induces an increase in the fluorescent signal of 30% and 100% over basal values in HUVECs exposed to CSA concentrations of 2.5 and 200 ␮M, respectively. Assuming a linear relationship between oxidant levels and cellular fluorescence, a twofold increase in cellular oxidants may be lower than the threshold for triggering an apoptotic response. On the other hand, this increase is enough to cause lipid peroxidation (16), which paves the way to apoptosis, but CSA-induced inhibition of the mitochondrial permeability transition pore may block the apoptotic response to CSA-induced oxidative stress. Furthermore, the possibility that the failure of oxidative stress to trigger apoptosis in HUVECs is related to differences between the factors regulated in CO-treated EC (i.e., superoxide dismutase and caspase-1) and in CSAtreated EC (Bcl-2) or to interspecies differences (human vs. bovine) may not be ruled out at this time. It must also be considered that CSA probably modulates the level of several pro- and antiapoptotic factors in EC, perhaps including caspases in addition to Bcl-2, and further studies will be required to address this issue. At 738

Vol. 15

March 2001

clinical concentrations of CSA, a 30% increase over basal values of cellular oxidants may be enough to modulate Bcl-2 levels in a way similar to that reported for other mammalian fibroblasts and endothelial cells. In conclusion, whereas the responses of HUVECs to acute treatment with pharmacological concentrations might mirror the presence of different branches in the CSA mechanisms of action, the ability to switch between adaptation and apoptosis for different levels of CSA may rely as much on the different effects of CSA on its target proteins as on the general properties of the response of HUVECs to cellular oxidants, which may operate independently to produce the full spectrum of drug effects in HUVECs. Adaptation to oxidants generated in EC by clinical doses of CSA: friend or foe? The observation that clinical doses of CSA may affect HUVECs through the generation of oxidants without causing apoptosis does not support a role for CSAtriggered apoptosis in the pathogenesis of vascular problems of transplanted patients. On the other hand, the up-regulation of Bcl-2, which occurs at clinical concentrations and indicates an adaptation to oxidants, may play a key pathogenetic role in these patients. The morphological changes observed in response to clinical levels of CSA with the loss of the cobblestone morphology typical of endothelial cells and the development of an elongated ‘fibroblast-like’ phenotype may indicate a loss of specialization that occurs in parallel to adaptation of HUVECs to CSA-induced oxidants. By analogy with the increased resistance of cancer cells to antineoplastic drugs because of their increased content of Bcl-2 (41), the up-regulation by CSA-induced oxidants of Bcl-2 may increase the survival of functionally impaired HUVECs, as suggested by their altered morphology. Future studies are required to define the functional properties of HUVECs chronically exposed to clinical levels of CSA, with special reference to the regulation of 䡠 NO production, which may eventually affect the function of the vasculature. Considering that endothelial cell lysis is an important step in the cascade of events responsible for hyperacute rejection of xenotransplanted organs (42), the adaptive properties of endothelial cells may be useful to maintain the integrity of their membrane and its function as a selective barrier during hyperacute rejection. Recent results indicate that Bcl-2 transfection increases the ability of HUVECs to colonize and form new microvessels when xenotransplanted in immunodeficient mice (43). A potential hazard of including HUVECs transfected by retroviral transduction into synthetic tissues intended for human use is the malignant potential of Bcl-2 overexpressing cells. However, we found no evidence for increased cell proliferation in HUVECs with up-regulated Bcl-2 levels. Furthermore, life-long treatment with CSA does not induce endothelial cell proliferation in transplanted patients, suggesting that Bcl-2

The FASEB Journal

LONGONI ET AL.

overexpression in human endothelial cells may not pose major concerns in clinical application. Our results may thus encourage further development of Bcl-2 overexpressing cells in the engineering of tissue equivalents.

15.

We thank Prof. Kelvin J. A. Davies and Dr. A. Celi for their critical review and helpful suggestions for an earlier version of this paper. We are indebted to Prof. P. L. Marchiafava for constant encouragement and helpful discussions. We are also indebted to Sister Fatima of the S. M. Addolorata Clinic, Pisa, for kindly providing umbilical cords, Dr. A. Lofaro, for measuring the ratio of bound to free CSA in the EC culture medium, Prof. R. Giordani for the lactic dehydrogenase assay, and Dr. A. Ferretti for help with the experiments. This study was supported by the Fondazione Pisana Ricerche in Chirurgia and by the National Research Council (CNR) grants to F.M.

17.

REFERENCES

16.

18. 19.

20.

21.

22. 23.

1. 2. 3.

4.

5.

6. 7.

8.

9. 10. 11. 12.

13. 14.

Kahan, B. D. (1989) Cyclosporin. N. Engl. J. Med. 321, 1725– 1738 Zoja, C., Furci, L., Ghilardi, F., Zilio, P., Benigni, A., and Remuzzi, G. (1986) Cyclosporin-induced endothelial cell injury. Lab. Invest. 55, 455– 462 Myers, B. D., Newton, L., Boshkos, C., Macoviak, J. A., Frist, W. H., Derby, G. C., Perlroth, M. G., and Sibley, R. K. (1988) Chronic injury of human renal microvessels with low-dose cyclosporine therapy. Transplantation 146, 694 –703 Myers, B. D., Sibley, R., Newton, L., Tomlanovich, S. J., Boshkos, C., Stinson, E., Luetscher, J. A., Whitney, D. J., Krasny, D., Coplon, N. S., et al. (1988) The long-term course of cyclosporine-associated chronic nephropathy. Kidney Int. 33, 590 – 600 Wolf, A., Trendelenburg, C. F., Diez-Fernandez, C., Prieto, P., Houy, S., Trommer, W. E., and Cordier, A. (1997) Cyclosporin A-induced oxidative stress in rat hepatocytes. J. Pharmacol. Exp. Ther. 280, 1328 –1334 Wang, C., and Salahudeen, A. K. (1995) Lipid peroxidation accompanies cyclosporin nephrotoxicity: effects of vitamin E. Kidney Int. 47, 927–934 Durak, I., Karabacak, H. I., Buyukkocak, S., Cimen, M. Y., Kacmaz, M., Omeroglu, E., and Ozturk, H. S. (1998) Impaired antioxidant defense system in the kidney tissues from rabbits treated with cyclosporine. Protective effects of vitamins E and C. Nephron 78, 207–11 Kumar, K. V., Naidu, M. U., Shifow, A. A., Prayag, A., and Ratnakar, K. S. (1999) Melatonin: an antioxidant protects against cyclosporin-induced nephrotoxicity. Transplantation 67, 1065–1068 Wang, C., and Salahudeen, A. B. (1994) Cyclosporine nephrotoxicity: attenuation by an antioxidant-inhibitor of lipid peroxidation in vitro and in vivo. Transplantation 58, 940 –946 Griffith, T. M., Edwards, D. H., Lewis, M. J., Newby, A. C., and Henderson, A. H. (1984) The nature of endothelium-derived vascular relaxant factor. Nature (London) 308, 645– 647 Furchgott, R. F., Cherry, P. D., Zawadzki, J. V., and Jothianandan, D. (1984) Endothelial cells as mediators of vasodilation of arteries. J. Cardiovasc. Pharmacol. 6, S336 –S343 Gerkens, J. F. (1989) Cyclosporin treatment of normal rats produces a rise in blood pressure and decreased renal vascular responses to nerve stimulation, vasoconstrictors and endothelium-dependent dilators. J. Pharmacol. Exp. Ther. 250, 1105–1112 Rego, A., Vargas, R., Wroblewska, B., Foegh, M. L., and Ramwell, P. W. (1990) Attenuation of vascular relaxation and cyclic GMP responses by cyclosporin A. J. Pharmacol. Exp. Ther. 252, 165–170 Diederich, D., Skopec, J., Diederich, A., and Dai, F. X. (1994) Cyclosporine produces endothelial dysfunction by increased production of superoxide. Hypertension 23, 957–961

OXIDATIVE STRESS IN ENDOTHELIAL CELLS

24. 25. 26. 27. 28. 29. 30. 31. 32.

33.

34. 35.

36.

37.

38.

Stephan, D., Billing, A., Krieger, J. P., Grima, M., Fabre, M., Hofner, M., Imbs, J. L., and Barthelmebs, M. (1995) Endothelium-dependent relaxation in the isolated rat kidney: impairment by cyclosporine A. J. Cardiovasc. Pharmacol. 26, 859 – 868 Longoni, B., Boschi, E., Demontis, G. C., Marchiafava, P. L., and Mosca, F. (1999) Regulation of Bcl-2 protein expression during oxidative stress in neuronal and in endothelial cells. Biochem. Biophys. Res. Commun. 260, 522–526 Buttke, T. M., and Sandstrom, P. A. (1994) Oxidative stress as mediator of apoptosis. Immunol. Today 15, 7–10 Green, D. R., and Reed, J. C. (1998) Mitochondria and apoptosis. Science 281, 1309 –1312 Mosieniak, G., Figiel, I., and Kaminska, B. (1997) Cyclosporin A, an immunosuppressive drug, induces programmed cell death in rat C6 glioma cells by a mechanism that involves the AP-1 transcription factor. J. Neurochem. 68, 1142–1149 Pyrzynska, B., Mosieniak, G., and Kaminska, B. (2000) Changes of the trans-activating potential of AP-1 transcription factor during cyclosporin A-induced apoptosis of glioma cells are mediated by phosphorylation and alterations of AP-1 composition. J. Neurochem. 74, 42–51 Jaffe, E. A., Nachmann, R. L., Becker, C. G., and Minick, C. R. (1973) Culture of human endothelial cells derived from umbilical veins. Identification by morphologic and immunologic criteria. J. Clin. Invest. 52, 2745–2756 Bergmeyer, H. U. (1974) In Methods in Enzymatic Analysis, Verlag Chemie/Academic Press, New York Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) Protein measurements with the Folin phenol reagent. J. Biol. Chem. 193, 265–275 Armitage, P. (1971) Statistical Methods in Medical Research, Blackwell Scientific Publications Esterbauer, H., and Cheeseman, K. H. (1990) Determination of aldehydic lipid peroxidation products: malonaldehyde and 4-hydroxyalkenals. Methods Enzymol. 186, 407– 421 Kerr, J. F. R., Wyllie, A. H., and Currie, A. R. (1972) Apoptosis: a basic biological phenomenon with wide-ranging implications in tissue kinetics. Br. J. Cancer 26, 239 –257 Wyllie, A. H. (1980) Glucocorticoid-induced thymocyte apoptosis is associated with endogenous endonuclease activation. Nature (London) 284, 555– 6 Steller, H. (1995) Mechanisms and genes of cellular suicide. Science 267, 1445–1462 Tan, D-X., Chen, L-D., Poeggeler, B., Manchester, L. C., and Reiter, R. J. (1993) Melatonin: a potent, endogenous hydroxyl radical scavenger. Endocrine J. 1, 57– 60 Reiter, R. J., Tan, D., Cabrera, J., and D’Arpa, D. (1999) Melatonin and tryptophan derivatives as free radical scavengers and antioxidants. Adv. Exp. Med. Biol. 467, 379 –387 Baliga, R., Ueda, N., Walker, P. D., and Shah, S. V. (1999) Oxidant mechanisms in toxic acute renal failure. Drug Metab. Rev. 31, 971–997 Theriault, Y., Logan, T. M., Meadows, R., Yu, L., Olejniczak, E. T., Holzman, T. F., Simmer, R. L., and Fesik, S. W. (1993) Solution structure of the cyclosporin A/cyclophilin complex by NMR. Nature (London) 361, 88 –91 Pflugl, G., Kallen, J., Schirmer, T., Jansonius, J. N., Zurini, M. G., and Walkinshaw, M. D. (1993) X-ray structure of a decameric cyclophilin-cyclosporin crystal complex. Nature (London) 361, 91–94 Clardy, J. (1995) The chemistry of signal transduction. Proc. Natl. Acad. Sci. USA 92, 56 – 61 Swanson, S. K., Born, T., Zydowsky, L. D., Cho, H., Chang, H. Y., Walsh, C. T., and Rusnak, F. (1992) Cyclosporin-mediated inhibition of bovine calcineurin by cyclophilins A and B. Proc. Natl. Acad. Sci. USA 89, 3741–3745 Crompton, M., Ellinger, H., and Costi, A. (1988) Inhibition by cyclosporin A of a Ca2⫹-dependent pore in heart mitochondria activated by inorganic phosphate and oxidative stress. Biochem. J. 255, 357–360 Marzo, I., Brenner, C., Zamzami, N., Susin, S. A., Beutner, G., Brdiczka, D., Remy, R., Xie, Z. H., Reed, J. C., and Kroemer, G. (1998) The permeability transition pore complex: a target for apoptosis regulation by caspases and bcl-2-related proteins. J. Exp. Med. 187, 1261–1271 Liu, J., Farmer, J. D. Jr, Lane, W. S., Friedman, J., Weissman, I., and Schreiber, S. L. (1991) Calcineurin is a common target of

739

cyclophilin-cyclosporin A and FKBP-FK506 complexes. Cell 66, 807– 815 39. Davies, K. J. A. (1999) The broad spectrum of responses to oxidants in proliferating cells: a new paradigm for oxidative stress. IUBMB Life 48, 41– 47 40. Thom, S. R., Fisher, D., Xu, Y. A., Notarfrancesco, K., and Ischiropoulos, H. (2000) Adaptive responses and apoptosis in endothelial cells exposed to carbon monoxide. Proc. Natl. Acad. Sci. USA 97, 1305–1310 41. Jansen, B., Schlagbauer-Wadl, H., Brown, B. D., Bryan, R. N., van Elsas, A., Muller, M., Wolff, K., Eichler, H. G., and Pehamberger, H. (1998) Bcl-2 antisense therapy chemosensitizes human melanoma in SCID mice. Nat. Med. 4, 232–234

740

Vol. 15

March 2001

42.

Hiroyasu, S., Shiraishi, M., Koji, T., Mamadi, T., Sugawa, H., and Muto, Y. (1999) Analysis of the Fas system and Bcl-2 in rat liver allograft rejection. J. Surg. Res. 84, 204 –211 43. Schechner, J. S., Nath, A. K., Zheng, L., Kluger, M. S., Hughes, C. C. W., Sierra-Honigmann, M. R., Lorber, M. I., Tellides, G., Kashgarian, M., Bothwell, A. L. M., and Pober, J. S. (2000) In vivo formation of complex microvessels lined by human endothelial cells in an immunodeficient mouse. Proc. Natl. Acad. Sci. USA 97, 9191–9197

The FASEB Journal

Received for publication May 1, 2000. Revised for publication August 28, 2000.

LONGONI ET AL.