ToxSci Advance Access originally published online on April 13, 2007
Toxicological Sciences 2007 98(1):216-230; doi:10.1093/toxsci/kfm087
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Cyclophosphamide-Induced Apoptosis in COV434 Human Granulosa Cells Involves Oxidative Stress and Glutathione Depletion
,
,
,1
* Department of Community and Environmental Medicine
Department of Medicine, University of California, Irvine, California 92617
Department of Environmental Health, School of Public Health, Shanghai Jiaotong University, Shanghai, China
Department of Developmental and Cell Biology, University of California, Irvine, California 92617
1 To whom correspondence should be addressed at Center for Occupational and Environmental Health, 5201 California Avenue, Suite 100, Irvine, CA 92617. Fax: (949) 824-2345. E-mail: uluderer{at}uci.edu.
Received November 2, 2006; accepted April 10, 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The anticancer drug cyclophosphamide induces granulosa cell apoptosis and is detoxified by glutathione (GSH) conjugation. We previously showed that both cyclophosphamide treatment and GSH depletion induced granulosa cell apoptosis in rats, but the role of GSH in apoptosis in human ovarian cells has not been studied. Using the COV434 human granulosa cell line, we tested the hypotheses that (1) GSH depletion or treatment with 4-hydroperoxycyclophosphamide (4HC), a preactivated form of cyclophosphamide, induces apoptosis, (2) GSH depletion potentiates 4HC-induced apoptosis, and (3) 4HC-induced apoptosis is mediated by GSH depletion and oxidative stress. Cells were treated with buthionine sulfoximine (BSO), a specific inhibitor of GSH synthesis, with or without follicle stimulating hormone (FSH) or serum. A significant increase in the number of apoptotic cells, assessed by terminal deoxynucleotidyl transferase–mediated deoxy-uridine triphosphate nick-end labeling (TUNEL) and Hoechst 33342 staining, occurred with BSO treatment. Treatment with 4HC dose-dependently induced apoptosis by TUNEL, Hoechst staining, and caspase 3 activation. Treatment with 4HC caused an increase in reactive oxygen species generation, measured by dichlorofluorescein fluorescence, oxidative DNA damage, measured by 8-hydroxyguanosine immunostaining, and an oxidation of the redox potential for the oxidized glutathione/reduced glutathione couple. Total intracellular GSH declined after 4HC treatment, preceding the onset of cell death. Treatment with antioxidants inhibited 4HC-induced apoptosis. Combined treatment with BSO and 4HC caused greater induction of apoptosis than either treatment alone. These findings are consistent with roles for oxidative stress and GSH depletion in mediating the induction of apoptosis in COV434 cells by cyclophosphamide.
Key Words: granulosa cell; glutathione; cyclophosphamide; apoptosis; 4-hydroperoxycyclophosphamide; oxidative stress; ovary; antioxidant.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The anticancer drug cyclophosphamide causes amenorrhea and premature ovarian failure in women (Howell and Shalet, 1998
; Kumar et al., 1972). Studies in rodents suggest that cyclophosphamide disrupts ovarian function by destroying ovarian follicles. Cyclophosphamide treatment destroys ovarian follicles in mice and rats (Davis and Heindel, 1998; Plowchalk and Mattison, 1991, 1992; Shiromizu et al., 1984). Although mice are more sensitive to the effects of cyclophosphamide on the immature primordial and primary follicles, cyclophosphamide targets granulosa cells of more mature, antral follicles in both mice and rats (Davis and Heindel, 1998; Plowchalk and Mattison, 1991, 1992; Shiromizu et al., 1984). We have shown that the destruction of granulosa cells by cyclophosphamide in rats treated in vivo occurs by the induction of apoptosis (Lopez and Luderer, 2004). Both the death receptor and the mitochondrial apoptotic pathways have been implicated in the induction of apoptosis by cyclophosphamide in developing embryos (Little and Mirkes, 2002; Mirkes and Little, 2000) and hair follicles (Lindner et al., 1997). However, the mechanism by which cyclophosphamide induces granulosa cell apoptosis has not been investigated. The mitochondrial apoptotic pathway is initiated by damaging stimuli such as oxidative stress or DNA damage (Coppola and Ghibelli, 2000). Cyclophosphamide's therapeutic mechanism of action as an alkylating agent induces DNA damage (Gamcsik et al., 1999). Cyclophosphamide treatment has also been shown to cause oxidative stress (Berrigan et al., 1987; Murata et al., 2004; Venkatesan and Chandrakasan, 1995). In the present studies we focused on oxidative stress as a mediator of cyclophosphamide-induced granulosa cell apoptosis.
The tripeptide glutathione (GSH) is the most abundant intracellular nonprotein thiol. GSH serves numerous physiological functions including detoxification of reactive oxygen species (ROS), acting as a cofactor for enzymes, serving as a storage form for cysteine, and regulating protein and DNA synthesis by altering redox status (Dalton et al., 2004). Moreover, GSH conjugation is a critical Phase II biotransformation mechanism for exogenous toxicants. GSH depletion induces apoptosis in some cell types in culture (Anderson et al., 1999; Celli et al., 1998; Higuchi and Matsukawa, 1999; Merad-Boudia et al., 1998). In other cell types, GSH depletion sensitizes cells to proapoptotic stimuli (Schnelldorfer et al., 2000; Wright et al., 1998). In vivo suppression of GSH synthesis with buthionine sulfoximine (BSO) significantly increased antral follicle atresia, an apoptotic process of degeneration, in adult rat ovaries (Lopez and Luderer, 2004). In cultured preovulatory rat follicles, GSH depletion with BSO partially reversed the antiapoptotic effect of follicle stimulating hormone (FSH) on granulosa cells (Tsai-Turton and Luderer, 2006).
Cyclophosphamide requires metabolic activation via oxidation by cytochrome P450 enzymes (CYP2A6, -2B6, -2C8, -2C9, and -3A4) to 4-hydroxycyclophosphamide (Chang et al., 1993; Dirven et al., 1994; Gamcsik et al., 1999). 4-Hydroxycyclophosphamide undergoes ring opening to aldophosphamide, which spontaneously decomposes to the reactive metabolite phosphoramide mustard (Dirven et al., 1994; Gamcsik et al., 1999). Phosphoramide mustard is thought to be the active metabolite both in terms of anticancer activity and in terms of ovarian toxicity (Gamcsik et al., 1999; Plowchalk and Mattison, 1991). Cyclophosphamide itself, phosphoramide mustard, and other metabolites are detoxified by conjugation with GSH (Dirven et al., 1994; Gamcsik et al., 1999).
COV434 cells are derived from a human granulosa cell tumor; however, they possess many of the characteristics of normal granulosa cells, including cyclic-adenosine monophosphate production in response to FSH stimulation, estradiol synthesis in response to FSH and androstenedione, and formation of intercellular connections with cumulus cells surrounding a human oocyte (Zhang et al., 2000). Herein we describe experiments conducted using these cells to test the hypotheses that (1) GSH depletion and cyclophosphamide treatment induce apoptosis in human granulosa cells, (2) GSH depletion sensitizes human granulosa cells to the induction of apoptosis by cyclophosphamide, and (3) cyclophosphamide treatment causes GSH depletion and oxidative stress in granulosa cells, which mediate the induction of apoptosis.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Materials.
All chemicals were purchased from Sigma-Aldrich (St Louis, MO) or Fisher Scientific (Pittsburgh, PA), unless otherwise noted. Tissue culture reagents and Mycotect Kit were from Invitrogen (Carlsbad, CA). Recombinant human FSH (FSH; IOD/BIO Lot#AFP-8468A, 6650 IU/mg) was obtained from Dr A. F. Parlow, National Hormone and Peptide Program, NIDDK. WST-1 reagent, diaminobenzidine substrate and buffer, and In situ Cell Death Detection Kit POD were from Roche Applied Science (Indianapolis, IN). The DNeasy DNA extraction kit was from Qiagen (Valencia, CA). The 4-hydroperoxycyclophosphamide (4HC) was a gift of Dr Susan Ludeman, Duke University Medical Center, Durham, NC. Additional 4HC was purchased from IIT GmbH, part NIOMECH (Bielefeld, Germany).
COV434 human granulosa cell culture.
COV434 cells (gift of Dr Peter Schrier, University Hospital of Leiden, Netherlands) are derived from a human granulosa cell tumor (van den Berg-Bakker et al., 1993), but possess many characteristics of normal granulosa cells (Zhang et al., 2000). A preliminary dose–response experiment showed that culture for 24 h with concentrations of BSO from 100 to 1000µM suppressed intracellular GSH to less than 15% of control levels, whereas 5 and 20µM concentrations suppressed GSH to 15% and 26% of control levels, respectively (data not shown). Depletion of GSH to less than 10% of control levels caused apoptosis in various cell types (Celli et al., 1998; Higuchi and Matsukawa, 1999; Merad-Boudia et al., 1998), whereas depletion of GSH to about 27% of control levels did not cause apoptosis in another study (Jungas et al., 2002). Therefore, to increase the likelihood of inducing apoptosis, 100µM BSO was used for the subsequent experiments.
COV434 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) F12 with GlutaMAX and 9% fetal bovine serum (complete medium). Cells were routinely assessed for mycoplasma contamination using the Mycotect Kit and were found to be free of contamination. During maintenance, cells were passaged every 4–5 days. Cells from passages 34–77 were used for the experiments reported herein. For experiments, cells were harvested by trypsinization, counted using a hemacytometer, and were plated in complete medium. For DNA or protein extraction COV434 cells were plated 7 x 106 cells per 75-cm2 flask for 24 h treatment or 5 x 106 cells/flask for 48–96 h treatment. For GSH assay, cells were plated 2.5 x 106 cells per 25-cm2 flask. Cells were plated 5 x 104 cells per chamber of eight-chamber Lab-Tek II CC2 tissue culture slides (Nunc, Roskilde, Denmark) for TUNEL, Hoechst 33342 staining, or immunofluorescence. Cells were plated 2 x 104 cells per well of a 96-well plate for WST-1 cell viability assay and 3 x 104 cells per well for ROS assay.
For the experiments testing the effects of GSH depletion on apoptosis in COV434 cells, the complete medium was removed after 24 h, cells were washed twice with phosphate buffered saline (PBS), and the treatment media were added. All or a subset of the following treatment groups were included: (1) serum-free control medium alone (DMEM F12), (2) DMEM F12 plus 100µM BSO, (3) 50 ng/ml FSH, (4) 200 ng/ml FSH, (5) 50 ng/ml FSH plus 100µM BSO, (6) 200 ng/ml FSH plus 100µM BSO, (7) DMEM F12 with 9% FBS, and (8) DMEM F12 with FBS plus 100µM BSO. For experiments testing the effects of 4HC on apoptosis, COV434 cells were cultured for 24 h in complete medium before the addition of 4HC dissolved in medium. Final concentrations of 4HC in the culture dishes were 0, 1, 10, or 50µM. Additional experiments included groups that were cotreated with 4HC and 100µM BSO or with 4HC and one of three antioxidants. The concentrations of antioxidants were 5mM glutathione ethyl ester (GEE), a cell-permeable form of GSH, 0.5mM dithiothreitol (DTT), a thiol antioxidant, or 1mM ascorbic acid (AA) (Higuchi and Matsukawa, 1999; Merad-Boudia et al., 1998; Puri and Meister, 1983; Tsai-Turton and Luderer, 2006). Plasma concentrations of 1–10µM of the 4-hydroxycyclophosphamide metabolite of cyclophosphamide, to which 4HC spontaneously decomposes, have been reported in patients receiving infusions of cyclophosphamide prior to bone marrow transplantation (Ren et al., 1998).
GSH assay.
After treatment, cells were harvested by trypsinization, counted with a hemacytometer and trypan blue, and centrifuged at 300 x g for 5 min. Cell pellets were lysed using a Kontes pellet pestle homogenizer in buffer containing 20mM Tris base, 1mM ethylenediaminetetraacetic acid (EDTA), 250mM sucrose, 2mM L-serine, and 20mM borate for 30 s on ice. After removing an aliquot of suspension for protein assay with a bicinchoninic acid (BCA) assay kit (Pierce, Rockford, IL), 1/4 volume 5% sulfosalicylic acid was added. After 15 min incubation on ice, cell suspensions were centrifuged at 15,800 x g for 10 min at 4°C. Total GSH and GSSG concentrations in the supernatants were measured using a modification (Luderer et al., 2001; Tsai-Turton and Luderer, 2005, 2006) of an enzymatic recycling assay developed by Griffith (1980). The concentration of reduced glutathione (rGSH) was calculated as the total GSH concentration in the sample minus two times the GSSG concentration. The interassay coefficient of variation (CV), calculated from a pool that was run in every assay, was 16.8% for total GSH, and 4.5% for GSSG. The intra-assay CVs ranged from 4.0% to 13.4% for total GSH and from 3.0% to 4.9% for GSSG.
The GSH reduction potential was calculated using the Nernst Equation:
 |
 |
; Shen et al., 2005).
Terminal deoxynucleotidyl transferase-mediated deoxy-uridine triphosphate nick-end-labeling (TUNEL).
At the end of the treatment period, cells were washed twice with ice-cold PBS, fixed for 30 min with 4% paraformaldehyde on ice, washed again with cold PBS, and subjected to TUNEL. TUNEL was performed using the In Situ Cell Death Detection Kit POD essentially as described (Luderer et al., 2003), with the following modifications. The terminal transferase was diluted one part terminal transferase to nine parts fluorescein-conjugated deoxy-uridine triphosphate label solution and 10 parts TUNEL dilution buffer. The peroxidase converter solution was diluted 2:1 peroxidase solution to TUNEL dilution buffer.
TUNEL-positive cells and TUNEL-negative cells (counterstained with hematoxylin) were counted in three x400 fields per chamber. The average percentage of TUNEL-positive cells was calculated for each chamber and was used for data presentation and for statistical analyses.
Hoechst 33342 staining for apoptosis.
The bisbenzamide dye, Hoechst 33342, was used for detection of apoptotic cells. At the end of the treatment period, cells were washed and fixed as for TUNEL. Slides were then immediately stained with 50 µg/ml Hoechst 33342 dye diluted in PBS for 10 min in the dark, rinsed with distilled water, and allowed to dry in the dark. Slides were coverslipped with Mowiol mounting agent and viewed using an Olympus BX60 microscope equipped for fluorescence microscopy at an excitation wavelength of 355 nm and an emission wavelength of 465 nm. Cells with intensely fluorescent and/or fragmented nuclei were considered apoptotic (Zhivotovsky and Orrenius, 2001). Apoptotic cells were counted in three x200 fields per chamber, and the average was used for data presentation and statistical analyses.
Protein extraction, gel electrophoresis, and immunoblotting.
Cells were harvested by trypsinization and collected into complete medium to deactivate trypsin. After centrifugation of trypsinized cells and growth medium containing detached cells at 300 x g for 5 min, cell pellets were washed with PBS, then lysed on ice in RIPA buffer (PBS, 1% Nonidet-P-40, 0.1% deoxycholate, 0.1% sodium dodecyl sulfate) with protease inhibitors. Lysates were incubated on ice for 30 min and centrifuged at 15,800 x g for 10 min at 4°C. Protein concentrations in supernatants were measured using a BCA assay kit (Pierce, Rockford, IL), and 40 µg of protein per lane was separated by electrophoresis in 12% Tris–HCl polyacrylamide gels (BioRad, Hercules, CA). The proteins were transferred to polyvinylidine difluoride membranes, and immunoblotting was performed essentially as described (Luderer et al., 2001) using an activated caspase 3 antibody (Asp175, Cell Signaling, Beverly, MA) at a dilution of 1:100, followed by reprobing with ß-actin antiserum (Sigma-Aldrich) as a loading control.
DNA extraction and gel electrophoresis.
At the end of the treatment period, cells were collected by scraping with a cell scraper, followed by centrifugation at 300 x g to pellet the cells. Pellets were then washed with PBS, centrifuged again, and were stored at –70°C until DNA extraction using the DNeasy Kit according to the manufacturer's instructions. Ten micrograms of DNA from each sample was separated by electrophoresis in 2% agarose gels made with modified Tris–acetate–EDTA buffer with 0.5 µg/ml ethidium bromide. Visualization was accomplished by ultraviolet transillumination.
Cell viability WST-1 assay.
Cell viability was based on the cleavage of the tetrazolium salt WST-1 to formazan by mitochondrial dehydrogenases of viable COV434 granulosa cells. At the end of the treatment periods, 10 µl of WST-1 reagent was added to each well, and the plate was incubated for 2 h at 37°C in a humidified atmosphere of 95% air, 5% CO2. The absorbance at 440 nm was then read using a VersaMax Tunable Microplate Reader (Molecular Devices, Sunnyvale, CA). The absorbance of the average of three blank wells containing medium and WST-1 reagent with no cells was subtracted from each absorbance reading, and the resulting values were used for data analysis and presentation.
Immunofluorescence.
Oxidative DNA damage was assessed using 8-hydroxy-2'-deoxyguanosine (8-OHdG) immunofluorescence. Treated cells were washed with PBS, fixed and permeabilized with 70% methanol for 5 min on ice, followed by 100% methanol for 20 min at –70°C, and washed again with PBS. Slides were blocked for 30 min with 10% normal goat serum in PBS, washed in PBS, incubated for 1 h with monoclonal anti-8-OHdG antibody (1:500; Clone 4E9, Trevigen, Gaithersburg, MD), washed, incubated with fluorescein-conjugated goat anti-mouse IgG H + L (1:100; #115-095-146; Jackson ImmunoResearch, West Grove, PA) for 1 h, washed, coverslipped with Vectashield mounting medium, and viewed with an Olympus BX60 microscope equipped for epifluorescence. The number of positively stained cells was counted in two x200 fields per well, and the average was used for data presentation and statistical analyses.
ROS detection by dichlorofluorescein diacetate assay.
Dichlorofluorescein diacetate (DCF-DA) is taken up by cells, is cleaved to DCF by intracellular esterases, and is oxidized in the presence of ROS to DCF (Royall and Ischiropoulos, 1993). Modifications of a previously described DCF-DA microplate assay were used to measure ROS after 4HC treatment (Wang and Joseph, 1999). Cells were cultured 3 x 104 cells per well of a 96-well plate in 100 µl of complete medium for 24 h. For measurement of the acute effect of 4HC treatment on ROS generation, the wells were washed with PBS, loaded for 30 min at 37°C with 100µM DCF-DA in medium, washed again, and treated with 0, 1, 10, or 50µM 4HC or 500µM hydrogen peroxide as a positive control in Hanks Balanced Salt Solution while fluorescence intensity was measured every 5 min for 2 h using a Biotek FL600 spectrofluorometer microplate reader (Biotek Instruments, Winooski, VT) with excitation wavelength set at 485 nm, emission wavelength set at 530 nm, and temperature set at 37°C. For measurement of ROS at a later time point, 10 µl of medium containing 4HC to give a final concentration of 0, 1, 10, or 50µM was added to each well after the 24-h preincubation period, and the plate was returned to the incubator for 6 h. After washing with PBS, 100 µl of medium containing 100µM DCF-DA was added. After incubation at 37°C for 30 min, the DCF-DA containing medium was removed, cells were washed with PBS, and maintained in PBS while fluorescence intensity was read as above.
Estradiol radioimmunoassay.
For the experiments assessing the effect of FSH treatment on estradiol synthesis, 7 x 106 cells were plated per 75-cm2 flask for 24 h treatment or 5 x 106 cells/flask for 72 h treatment. After 24 h in complete medium, the cells were washed and treated with one of the following in DMEM F12 medium: (1) medium alone, (2) 200 ng/ml FSH, or (3) FSH plus 10µM androstenedione (Zhang et al., 2000). At the end of the treatment period, an aliquot of medium was removed and stored at –70°C until estradiol assay. Estradiol was measured using reagents from the Estradiol Double Antibody Radioimmunoassay Kit from Diagnostic Products Corporation (Los Angeles, CA), except that a standard curve was generated in DMEM F12 medium (0, 9, 18, 72, 145, 289, and 578 pg/ml). All of the samples were measured in a single assay. The intraassay CV was 8.5%.
Statistical analyses.
Data from replicate experiments were pooled for statistical analyses. Differences among treatment groups were evaluated by analysis of variance (ANOVA) for each time point. The proportions of TUNEL-positive cells were subjected to arcsine square root transformations (Pasternack and Shore, 1982) and then analyzed by ANOVA. If the overall p-value indicated statistical significance (p < 0.05), post hoc least significant difference (LSD) tests were applied to assess differences among groups based on a priori hypotheses. Levene's test was used to assess homogeneity of variances. If variances were not homogeneous, the natural logarithm transformation was applied. If variances were still not homogenous, the Kruskal–Wallis test was used to assess the overall effect of treatment. If the latter was statistically significant at p < 0.05, the Mann–Whitney test was then used for intergroup comparisons. SPSS Version 13.0 for Macintosh was used for all statistical analyses.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Effect of FSH and BSO Treatment on Intracellular GSH Concentrations in Cultured COV434 Human Granulosa Cells
COV434 cells are responsive to FSH stimulation in our hands, as indicated by the increase in estradiol synthesis when treated with 10µM androstenedione and 200 ng/ml FSH (Fig. 1). FSH treatment appeared to modestly increase intracellular GSH levels at 24 h, but not at later time points (Fig. 2A). This contrasts with the pronounced stimulatory effect of FSH we have observed on GSH synthesis in cultured rat preovulatory follicles (Tsai-Turton and Luderer, 2006
). Combined treatment with FSH and androstenedione also did not increase GSH concentrations relative to cells cultured in medium alone (data not shown). Treatment with BSO suppressed intracellular GSH levels to 15% of control levels after 24 h and to less than 5% of control levels at later time points (p < 0.001, effect of treatment; p = 0.117, effect of time by two-way ANOVA; Fig. 2A). Figure 2B shows that intracellular GSH concentrations were significantly decreased compared to 0 h control levels by 2 h after the onset of treatment with 100µM BSO and were suppressed to 50% of 0 h levels after 8 h of BSO treatment.
|
|
Effect of FSH and BSO Treatment on Apoptosis in Cultured COV434 Human Granulosa Cells
GSH depletion with BSO has been shown to induce apoptosis in some (Anderson et al., 1999
; Celli et al., 1998; Higuchi and Matsukawa, 1999; Merad-Boudia et al., 1998), but not all (Schnelldorfer et al., 2000; Wright et al., 1998), cell types that have been investigated. The effect of GSH depletion on apoptosis in COV434 cells was investigated using TUNEL, Hoechst 33342 staining, and DNA gel electrophoresis. Cotreatment with FSH and BSO or treatment with BSO alone led to a time-dependent and statistically significant increase in the percentage of TUNEL-positive cells, with nearly 100% of BSO-treated cells TUNEL positive after 96 h of treatment (p < 0.004, effect of BSO by two-way ANOVA; p < 0.008, effect of treatment by Kruskal–Wallis test at 24, 48, 72, and 96 h; Figs. 3A and 3B). The effect of FSH treatment was not statistically significant at any time point.
|
Staining of COV434 cells with Hoechst 33342 dye verified that cell death was apoptotic and not necrotic, as judged by intense nuclear staining and fragmentation by Hoechst staining (Zhivotovsky and Orrenius, 2001
). For these experiments, two additional treatment groups of cells cultured with medium supplemented with 9% FBS with or without BSO were included to assess whether the addition of serum would be more protective against apoptosis induced by GSH depletion than FSH treatment. There was a statistically significant increase in the number of apoptotic cells by Hoechst staining already after 24 h of culture with BSO, with or without concomitant FSH or serum treatment, and the number of Hoechst-positive cells increased further after 48 h of BSO treatment (p < 0.004, effect of BSO by three-way ANOVA at both time points; Figs. 3C and 3D). There was also a statistically significant effect of FSH treatment, with fewer Hoechst-positive cells in the 200 ng/ml FSH-treated groups compared to cells cultured with DMEM F12 alone (p = 0.001, effect of FSH dose; Fig. 3D). Culture in medium with serum also appeared to be slightly protective against apoptosis, but this was statistically significant only at 48 h (p = 0.001, effect of serum). Surprisingly, in light of the TUNEL and Hoechst results, no oligonucleosomal DNA fragmentation was observed by agarose gel electrophoresis at 24, 48, 72, or 96 h of culture in any treatment group (not shown).
Effect of FSH and BSO Treatment on COV434 Cell Viability
Treatment with BSO for 24, 48, 72, and 96 h significantly decreased cell viability, as assessed by WST-1 assay (p < 0.003, effect of treatment at 24, 48, 72, and 96 h by Kruskal–Wallis test; Fig. 4). Intergroup comparisons revealed significantly lower cell viability in nearly every BSO-treated group at every time point, compared to the respective non-BSO–treated control. At 24 h, FSH appeared to increase the number of viable cells (p = 0.033 and p = 0.061, 50 ng/ml FSH and 200 ng/ml FSH, respectively, vs. DMEM control by Mann–Whitney test). However, there was no effect of treatment with 50 or 200 ng/ml FSH or modulation of the effect of BSO by FSH on cell viability at any of the later time points.
|
4HC Treatment Induces Apoptosis in COV434 Granulosa Cells
Preliminary experiments with 10 to 500µM cyclophosphamide for 48–96 h did not increase apoptosis in COV434 cells as judged by only a minimal, nonstatistically significant, decrease in cell viability at the highest dose, suggesting that these cells do not efficiently metabolize cyclophosphamide to its active metabolites (data not shown). We therefore used 4HC, a preactivated form of cyclophosphamide (Flowers et al., 2000
; Gamcsik et al., 1999) for the subsequent experiments. Treatment for 12 or 24 h with 4HC induced apoptosis in COV434 cells in a dose- and time-dependent manner, as measured both by TUNEL (p < 0.001, effect of 4HC dose by Kruskal–Wallis test at both time points; Fig. 5A) and by Hoechst staining (p < 0.01, effect of dose at both time points; Fig. 5B). Treatment of COV434 cells with 4HC caused a dose-dependent appearance of activated (cleaved) caspase 3 protein at 12 and 24 h (Fig. 5C). 4HC treatment also led to statistically significant decreases in cell viability, as measured by WST cleavage (p < 0.001, effect of dose by Kruskal–Wallis test at both time points; Fig. 5D). Post hoc comparisons revealed that the lowest dose of 4HC (1µM) induced a statistically significant decrease in cell viability already at 12 h (Fig. 5D).
|
Oligonucleosomal DNA fragmentation was not evident on DNA gel electrophoresis of DNA extracted from COV434 cells treated with 0–50µM 4HC for 24 h (data not shown).
4HC Treatment Increases ROS Formation in COV434 Cells
ROS generation, measured as the slope of the increase in DCF fluorescence, was continuously monitored during 2 h treatment with 0, 1, 10, or 50µM 4HC or 500µM hydrogen peroxide as a positive control. There was a statistically significant effect of treatment on ROS generation during the 2-h incubation (p < 0.001, by 1-way ANOVA; Fig. 6A). ROS levels in the 50µM 4HC and H2O2 groups were significantly greater than in the 0µM 4HC controls. ROS were also measured using DCF-DA after 6 h incubation with 0, 1, 10, or 50µM 4HC. DCF fluorescence increased significantly in groups treated with 1, 10, and 50µM 4HC for 6 h compared to 0µM controls (p < 0.001, effect of 4HC by Kruskal–Wallis test, Fig. 6B).
|
4HC Treatment Causes Oxidative DNA Damage in COV434 Cells
To test whether the rise in ROS caused by treatment with 4HC causes oxidative DNA damage, we used 8-OHdG immunostaining. The number of 8-OHdG–positive cells increased in a dose-dependent manner after 12 h of 4HC treatment (p < 0.001, effect of dose by Kruskal–Wallis test, Fig. 7).
|
4HC Treatment Depletes Total Intracellular GSH in COV434 Cells
Treatment of cultured hepatic sinusoidal endothelial cells with 4HC or hepatocytes with cyclophosphamide caused a rapid decline in intracellular GSH concentrations that preceded the onset of cell death (DeLeve, 1996
). Prevention of this decline in GSH concentrations also protected against cell killing by cyclophosphamide (DeLeve, 1996). To determine whether the induction of COV434 cell apoptosis by 4HC also involves the depletion of GSH, we measured intracellular GSH concentrations at different times after 4HC treatment. Treatment of COV434 cells with 50µM 4HC caused a rapid decline in total intracellular GSH concentrations by 2 h compared to control 0µM cells (Fig. 8A; p = 0.002, p < 0.001, effect of 4HC by Kruskal–Wallis test at 2 and 4h, respectively). Treatment with 10µM 4HC caused a small, statistically significant decrease in GSH levels at 4 h only (Fig. 8A). The decline in GSH concentrations at 2 h in the 50µM group preceded an increase in cell death at 4 h, which was not statistically significant due to high variability in the percentage of dead cells in the 50µM group (Fig. 8E). GSSG concentrations also declined by 2 h of treatment with 50µM 4HC, but the effect of treatment was not statistically significant (Fig. 8B). The ratio of rGSH/GSSG decreased significantly after 2 and 4 h of treatment with 50µM 4HC (p = 0.002, effect of 4HC at 2 and 4 h by Kruskal–Wallis test; Fig. 8D). The GSH reduction potential (Eh) became significantly more positive (oxidized) after 2 and 4 h of 50µM 4HC treatment (p = 0.001, effect of 4HC by Mann–Whitney test at 2 and 4 h; Fig. 8E).
|
To assess whether longer duration of treatment with lower doses of 4HC would cause a further decline in GSH concentrations, cells were treated with 1 or 10µM 4HC for 8 or 12 h. There was no effect of 4HC treatment on total intracellular GSH concentrations at either time point (Fig. 9A), despite a small, but statistically significant increase in cell death (p < 0.001, effect of 4HC dose; Fig. 9B).
|
Supplementation of GSH or Treatment with Other Antioxidants Suppresses 4HC-Induced Apoptosis
To test whether the depletion of GSH or the generation of ROS mediates the induction of apoptosis by 4HC, we cotreated cells with 50µM 4HC and a cell-permeable form of GSH, GEE, or with the thiol antioxidant DTT or with AA for 24 h. Treatment with AA, DTT, or GEE alone significantly increased apoptosis, measured by TUNEL, in the absence of 4HC treatment (Fig. 10). This effect was greatest for AA and minimal for GEE. As expected, 50µM 4HC induced apoptosis in nearly 100% of the cells by 24 h. This effect was significantly suppressed by all three antioxidants, but most effectively by GEE and DTT (Fig. 10; p < 0.01 4HC + antioxidant groups vs. 4HC by Mann–Whitney test).
|
Depletion of GSH with BSO Sensitizes COV434 Cells to 4HC-Induced Apoptosis
To test whether GSH depletion enhances the toxicity of 4HC to COV434 cells, the cells were treated with 0, 1, or 10µM 4HC alone or in combination with 100µM BSO. Treatment with 4HC alone caused a dose-dependent and statistically significant increase in apoptosis that became more pronounced with increasing duration of treatment, as judged by both TUNEL (p < 0.001, effect of 4HC dose by three-way ANOVA; Fig. 11A) and Hoechst staining (p < 0.001, effect of 4HC dose; Fig. 11B). Cotreatment with BSO increased apoptosis as assessed by TUNEL (p < 0.001, effect of BSO, Fig. 11A) and Hoechst staining (p < 0.001, effect of BSO by three-way ANOVA) in 4HC treated cells as well as in control cells not treated with 4HC. For each 4HC dose and time point, there were more apoptotic cells with BSO than without. The effect of 4HC on cell viability (Fig. 11C) was similarly enhanced by cotreatment with BSO (p
0.001, effect of 4HC dose, effect of BSO, effect of time by three-way ANOVA).
|
To assess whether the effect of GSH depletion would be more pronounced if GSH levels were already suppressed when 4HC was added to the cultures, cells were treated for 8 h with 100µM BSO, followed by treatment with 1 or 10µM 4HC for 12 or 24 h (Fig. 12). The effects of treatment with 4HC and with BSO on the number of apoptotic cells as judged by TUNEL and Hoechst staining were highly statistically significant (p < 0.001, effect of 4HC, effect of BSO by three-way ANOVA). The number of Hoechst-positive cells at both time points was greater in the groups treated with both 4HC and BSO than the sum of the number of positive cells in the groups treated with either compound alone (p = 0.012, 4HC x BSO interaction). Similarly, pretreatment with BSO significantly enhanced the percentage of cells that were TUNEL positive at both doses of 4HC (p < 0.001, 4HC x BSO interaction).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cyclophosphamide treatment causes temporary amenorrhea or permanent ovarian failure in women (Howell and Shalet, 1998
; Kumar et al., 1972). In rodents cyclophosphamide destroys primordial and primary ovarian follicles as well as preantral and antral follicles (Davis and Heindel, 1998; Plowchalk and Mattison, 1991, 1992; Shiromizu et al., 1984). In primordial and small primary follicles, cyclophosphamide appears to target the oocytes for apoptotic destruction, whereas in larger follicles, cyclophosphamide induces granulosa cell apoptosis followed by death of the oocyte (Desmeules and Devine, 2006; Lopez and Luderer, 2004). In the current study we have used a human granulosa cell tumor line and a preactivated form of cyclophosphamide, 4HC, to investigate the mechanism by which cyclophosphamide induces apoptosis in granulosa cells and to examine the role of GSH in granulosa cell apoptosis. We demonstrated that 4HC dose-dependently induced apoptosis in COV434 human granulosa cells and that the induction of apoptosis by 4HC was associated with caspase 3 activation. 4HC-induced apoptosis was preceded by increased ROS levels, decreased ratio of reduced to oxidized glutathione (GSSG), and by oxidation of the GSSG/2rGSH reduction potential. Together with the observation that 4HC caused oxidative DNA damage, these results show that 4HC treatment induces oxidative stress in COV434 cells. Cotreatment with antioxidants protected against 4HC-induced apoptosis. We also showed that 4HC-induced apoptosis was preceded by depression of intracellular GSH concentrations, that GSH depletion using BSO induced apoptosis, and that combined treatment with BSO and 4HC more than additively induced apoptosis. Together these findings provide strong evidence that GSH depletion mediates 4HC-induced apoptosis in COV434 cells via the induction of oxidative stress.
The observations in the present study that 4HC treatment (Fig. 5) and GSH depletion with BSO (Fig. 3) induced apoptosis in a human granulosa cell line are consistent with our previous observations in rat granulosa cells. Treatment of adult female rats with cyclophosphamide caused granulosa cell apoptosis in secondary and antral follicles within 24 h (Lopez and Luderer, 2004). This apoptosis after in vivo treatment with cyclophosphamide was associated with activation of caspase 9 and caspase 3 in granulosa cells (unpublished data). Similarly, in the present experiments treatment of human COV434 granulosa cells with 4HC was associated with caspase 3 activation. In vivo treatment with BSO led to a significant increase in antral follicle atresia (Lopez and Luderer, 2004), and treatment of cultured preovulatory rat follicles with BSO increased granulosa cell apoptosis (Tsai-Turton and Luderer, 2006). GSH depletion alone has also been reported to induce apoptosis in some cultured cell lines (Anderson et al., 1999; Celli et al., 1998; Higuchi and Matsukawa, 1999; Merad-Boudia et al., 1998).
Unlike COV434 cells and the other cell lines mentioned above, GSH depletion alone did not induce apoptosis in pancreatic cancer cells (Schnelldorfer et al., 2000), HL-60 or PW cells overexpressing Bcl-2 (Wright et al., 1998), HeLa cells (Jungas et al., 2002), or nonsmall cell lung cancer cells (Honda et al., 2004), but GSH depletion sensitized these cells to other proapoptotic stimuli. In the current study, we observed that GSH depletion with BSO slightly, but statistically significantly, enhanced the induction of apoptosis by 4HC when BSO and 4HC treatment were initiated at the same time (Fig. 11). Because it takes several hours after BSO treatment for intracellular GSH levels to fall (Fig. 2B), we tested whether pretreatment with BSO for 8 h prior to initiating 4HC treatment would augment this potentiating effect of BSO. Pretreatment with BSO, followed by the addition of 4HC indeed led to greater potentiation of 4HC-induced apoptosis (Fig. 12) than was observed when the treatments were initiated simultaneously.
Oligonucleosomal DNA fragmentation, yielding the characteristic DNA ladder pattern on DNA gel electrophoresis, is often considered a hallmark of apoptosis. However, despite TUNEL and Hoechst 33342 staining and caspase 3 activation, we did not observe DNA laddering in COV434 cells treated with 4HC or BSO alone. The only other study of which we are aware in which apoptosis was studied in COV434 cells assessed apoptosis by morphological changes and by flow cytometry analysis after staurosporine treatment (Zhang et al., 2005). Oligonucleosomal DNA fragmentation was not evaluated in that study. Others have reported that oligonucleosomal DNA fragmentation is not required for apoptotic cell death, whereas high molecular weight DNA fragmentation, resulting from the excision of nuclear loop domains, appears to be a necessary step for the initiation of apoptosis (Oberhammer et al., 1993; Solovyan et al., 1998). In a human T-lymphoblastic leukemia cell line, treatment with the antineoplastic drug 1-ß-D-arabinofuranosylcytosine (ara-C) caused high molecular weight DNA fragmentation, which was followed by lower molecular weight DNA fragmentation that did not exhibit the characteristic oligonucleosomal fragmentation pattern (Badran et al., 2003). In these experiments, lower molecular weight DNA fragmentation was confirmed by TUNEL and by the detection of a subpopulation of cells with less than 2N DNA complement by flow cytometry (Badran et al., 2003). Moreover, significant increases in caspase 3 enzymatic activity were also detected in these cells (Badran et al., 2003). We conclude that cell death in the present experiments was indeed apoptotic, but that COV434 cells may not undergo oligonucleosomal DNA fragmentation during apoptosis. Unlike the similarities noted above between granulosa cell apoptosis in rodent models and COV434 cell apoptosis in the present study, the lack of oligonucleosomal DNA fragmentation in COV434 cells constitutes a difference between these two models. Rat granulosa cells exhibit oligonucleosomal DNA fragmentation during apoptosis in vivo and in vitro (Tilly et al., 1992; Trbovich et al., 1998; Zeleznik et al., 1989). The lack of a prominent antiapoptotic effect of FSH treatment is another difference between the COV434 cells and cultured primary rat follicles (Chun et al., 1994, 1996; Tsai-Turton and Luderer, 2006) and primary human luteinized granulosa cells (Matsubara et al., 2000). These differences suggest that primary human granulosa cells may respond differently to GSH depletion or 4HC treatment than COV434 cells, and we plan to explore this in future experiments.
Treatment of COV434 cells with 4HC significantly depressed intracellular GSH concentrations. This is consistent with previous reports that cyclophosphamide or 4HC treatment depleted GSH concentrations in other cell types. Treatment of hepatic sinusoidal epithelial cells with 4HC was reported to cause 95% depletion of GSH and a significant increase in cell death within 6 h (DeLeve, 1996). Treatment of cultured hepatocytes with cyclophosphamide also resulted in GSH depletion and cell death, and prevention of the decline in GSH protected against cell killing (DeLeve, 1996). In vivo treatment of mice with cyclophosphamide has also been reported to result in rapid depletion of bone marrow, liver, and blood GSH concentrations that was maximal at 24 h (Adams et al., 1985). A mechanistic role for active GSH extrusion from cells has been proposed by Coppola and Ghibelli (2000) in the mitochondrial apoptotic pathway. These workers showed that GSH extrusion from cells preceded cytochrome c release, that inhibition of GSH extrusion with methionine or cystathionine inhibited apoptosis induced by puromycin, and that GSH depletion with BSO also caused cytochrome c release (Coppola and Ghibelli, 2000; Ghibelli et al., 1999). Taken together, our observations that depleting GSH with BSO-induced apoptosis, that 4HC caused depletion of GSH prior to the onset of apoptosis, that GSH supplementation with GEE prevented 4HC-induced apoptosis, and that GSH depletion with BSO enhanced the apoptotic effect of 4HC suggest that GSH depletion mediates the induction of apoptosis by 4HC in COV434 cells.
In the present study we observed that 4HC treatment of COV434 cells caused declines in the ratio of reduced GSH to GSSG and oxidation of the GSSG/2rGSH reduction potential by 2 h (Fig. 8), increased ROS by 2 h (Fig. 6), and caused oxidative DNA damage by 12 h of treatment (Fig. 7). Furthermore, cotreatment with the antioxidants GEE, DTT, or AA prevented 4HC-induced apoptosis (Fig. 10). These findings provide strong evidence that 4HC initiates apoptosis in granulosa cells by inducing oxidative stress. The increase in ROS appears to be due to the depletion of GSH, since levels of GSSG did not increase, but rather declined, with 4HC treatment. Others have observed increased levels of lipid peroxidation in bronchoalveolar lavage fluid (Venkatesan and Chandrakasan, 1995) and in the liver (Berrigan et al., 1987) after in vivo treatment with cyclophosphamide. Murata et al. (2004) reported that treatment of HL-60 cells for 2 h with a dose of 4HC that also induced apoptosis led to a significant increase in 8-OHdG formation. In contrast, these workers reported no apoptosis or significantly increased 8-OHdG formation when the cells were treated with cyclophosphamide. There are several possible explanations for these apparently discordant observations. One explanation is that 4HC decomposition liberates hydrogen peroxide as well as 4-hydroxycyclophosphamide, which spontaneously converts to phosphoramide mustard and other metabolites of cyclophosphamide, whereas treatment with cyclophosphamide itself does not generate hydrogen peroxide. Murata et al. (2004) provided evidence that hydrogen peroxide was generated upon treatment of HL-60 cells with 4HC, but did not assess hydrogen peroxide generation after cyclophosphamide treatment. Another, not mutually exclusive explanation is that HL-60 cells do not possess the cytochrome P450 enzymes required to metabolize cyclophosphamide to 4-hydroxycyclophosphamide. The latter explanation seems to be plausible, as CYP3A4, -2B6, -2C8, and -2C9 were undetectable by reverse transcription-polymerase chain reaction in untreated HL-60 cells (Kawai et al., 2003; Nagai et al., 2002). On the other hand, the presence of mRNA for CYP2A6 in HL-60 cells (Nagai et al., 2002) suggests that they may possess at least one of the cytochrome P450 enzymes that is active in metabolizing cyclophosphamide (Chang et al., 1993). The cytochrome P450 expression profiles of COV434 cells have not been investigated, but the lack of significant cell death upon treatment with cyclophosphamide in our preliminary experiments suggests that COV434 cells lack the ability to metabolize cyclophosphamide.
In summary, we have shown that that treatment of COV434 human granulosa cells with 4HC caused oxidative stress, as indicated by oxidative DNA damage, DCF fluorescence, and decreased ratio of reduced GSH to GSSG, and induced apoptosis subsequent to depressing total intracellular GSH concentrations. We further showed that GSH depletion by inhibition of its synthesis using BSO induced apoptosis, that BSO enhanced the apoptotic effect of 4HC treatment, and that antioxidants prevented the apoptotic effect of 4HC on COV434 granulosa cells. These findings are consistent with roles for oxidative stress and GSH depletion in mediating the induction of apoptosis in COV434 granulosa cells by cyclophosphamide. Future studies will investigate whether these findings are also applicable to the induction of apoptosis by cyclophosphamide in primary human granulosa cells.
|
ACKNOWLEDGMENTS |
|---|
We are grateful to Dr Susan Ludeman for the gift of the 4HC and to Dr Peter Schrier for the gift of the COV434 cells. This work was funded by National Institutes of Health grant ES10963 to U.L. and by the Center for Occupational and Environmental Health, University of California Irvine. B.T.L. received a UC Irvine Undergraduate Research Opportunities Program grant to carry out some experiments described herein.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Adams DJ, Carmichael J, Wolf CR. Altered mouse bone marrow glutathione and glutathione transferase levels in response to cytotoxins. Cancer Res. (1985) 45:1669–1673.
Anderson CP, Tsai JM, Meek WE, Liu R-M, Tang Y, Forman HJ, Reynolds CP. Depletion of glutathione by buthionine sulfoximine is cytotoxic for human neuroblastoma cell lines via apoptosis. Exp. Cell Res. (1999) 246:183–192.[CrossRef][ISI][Medlin