N-acetylcysteine impairs survival of luteal cells through mitochondrial dysfunction

Authors


Abstract

N-acetylcysteine (NAC) is known as an antioxidant and used for mucus viscosity reduction. However, this drug prevents or induces cell death depending on the cell type. The response of steroidogenic luteal cells to NAC is unknown. Our data shows that NAC can behave as an antioxidant or prooxidant in dependency on the concentration and mitochondrial energization. NAC elevated the flowcytometric-measured portion of hypodiploid (dying) cells. This rise was completely abolished by aurintricarboxylic acid, an inhibitor of topoisomerase II. NAC increased the secretion of nitric oxide and cellular nitrotyrosine. An image analysis indicated that cells pretreated with NAC and loaded with DHR showed a fluorescent structure probably elicited by the oxidative product of DHR, rhodamine 123 that sequesters mitochondrially. Pretreating luteal cells with NAC or adding NAC directly to mitochondrial fractions followed by assessing the mitochondrial transmembrane potential difference (Δψ) by the JC-1 technique demonstrated a marked decrease in Δψ. A protonophore restored Δψ and rotenone (an inhibitor of respiratory chain complex I) inhibited mitochondrial recovering. Thus, in steroidogenic luteal cells from healthy mature corpus luteum, NAC impairs cellular survival by interfering with mitochondrial metabolism. The protonophore-induced recovering of NAC-provoked decrease in Δψ indicates that an ATP synthase-favored route of H+ re-entry to the matrix is essentially switched off by NAC while other respiratory chain complexes remain intact. These data may be important for therapeutic timing of treatments with NAC. © 2010 International Society for Advancement of Cytometry

The corpus luteum is a transient ovarian steroidogenic gland, which develops within few days from the ovulated follicle to establish embryonic development. Intense vascularization is essential for maturation and maintenance of a healthy gland (1, 2). The mature mid-luteal phase corpus luteum secretes large quantities of progesterone, a derivative from its mitochondrially formed precursor, pregnenolone.

Steroidogenesis-dependent oxy-radicals and oxidants are generated during intramitochondrial conversion of cholesterol through the activity of NADPH-dependent cholesterol monooxygenase (side chain cleaving). It catalyzes in a phospholipid-assisted manner the formation of pregnenolone. This process is accompanied by an electron leakage into the mitochondrial matrix (3) and the synthesis of 4-methylpental as byproduct (4, 5). In addition, regulatory constituents driving enhanced steroidogenesis are generated by the cyclooxygenase and lipoxygenase pathways which require reactive oxygen species. For instance, cyclooxygenase has an absolute requirement for hydroperoxide to catalyze oxygenation of lipids with a lipoperoxidation as the first step in forming prostanoids (6). In turn, high concentration of antioxidants and oxidant-detoxifying enzymes are present in the ovary (7) to cope with excessive reactive oxygen species (ROS). They have been also implicated in the glandular regression, since oxidants such as hydrogen peroxide can decrease steroidogenesis, a feature of functional regression. However, the role of endogenous production of ROS by healthy corpus luteum remains largely unknown because the majority of the cellular systems used exogenous sources of ROS (8–10).

Studies on leukocytes, hemapoietic, vascular smooth muscle, and endothelial cells provided evidence that ROS can function as essential second messengers that are necessary for regulation of gene expression and cellular metabolism (11, 12). These cell types differ in their response to antioxidants, particularly to N-acetylcysteine (NAC). This drug, readily forming disulfide bonds and well known for mucus viscosity reduction, has been a subject of numerous clinical and experimental research aimed to clarify its role in modulation of ROS activity (13).

Depending on experimental condition and cell type, inconsistent results have been reported about NAC effects on cellular processes potentially related to ROS. Hepatocytes (14) and bovine artery endothelial cells (15) responded to NAC with an increase in their glutathione content, leukocytes with a decrease (11). At a dose at which NAC was reported to be used as an antioxidant (5 mM), the drug induced apoptosis in human and rat aortic smooth muscle cells but not in endothelial cells (16).

The effect of NAC on steroidogenic luteal cells is unknown. They develop from follicular granulosa cells (17) which exhibit endothelial-like properties (18). This led us to hypothesize that NAC may support luteal cell survival owing to its activity to prevent excessive production of ROS. However, we observed a NAC-provoked rise in the portion of dead cells and a compartmentalized ROS level in luteal cells from midphase corpus luteum using dihydrorhodamine 123 (DHR) as cell-permeable ROS-specific fluorogenic probe. It is oxidized to positively charged fluorescent rhodamine 123 (R 123) that sequesters mitochondrially (19, 20). Luteal cells, which were pretreated with inhibitors of lipid-oxygenating activities, responded to NAC in a strongly dose-dependent fashion with decrease or increase in the oxidation of DHR. Considering these data, we assumed that mitochondria can be targets of NAC. The primary factor governing mitochondrial ROS generation is the redox state of the respiratory chain (21, 22). Thus, we focused on the mitochondrial transmembrane potential in experiments aimed to elucidate the NAC-elicited compartmentalization of R 123, the oxidative product of DHR.

MATERIALS AND METHODS

Cells

Luteal cells were isolated from ovaries sampled by ovariectomy using Holstein heifers (420 ± 30 kg in body weight, 16 ± 2 month old) after cycle synchronization. The corpus luteum was regressed by an injection (i.m.) of a prostaglandin F analog (cloprostenol) followed by an injection of the analog of gonadotropin releasing hormone (depherelin) as described previously (23). The corpus luteum was cut in small pieces and mechanically dispersed in cold Dulbecco's phosphate-buffered saline (DPBS) containing bovine serum albumin (BSA, 1%). After filtration (Falcon, 100 μm pores) and centrifugation (60g, 10 min, 4°C), the cells were resuspended in Mega cell medium (Sigma). Large and small luteal cells were counted using the distribution profile displayed by an analytical counter (Coulter counter) or the portion of large and small cells displayed by a flowcytometric histogram as indicated in Figure 2A. Aliquots were cultured on cell culture plates (Greiner). Concentration of additives and periods of cultivation are indicated in figures. Contamination with endothelial cells and leukocytes was tested by immunocytochemistry using mouse antibodies against CD31 respective CD11a (commonly expressed on leukocytes) and proved to be reactive to bovine epitopes by the manufacturer (Serotec). Anti-mouse antibody fluoresceinisothiocynate conjugate was used as the negative control according to the recommendation of the manufacturer (Serotec). Single cell fluorescence was measured by flow cytometry. Steroidogenesis was assayed using an aliquot of the cellular supernatant by RIA as described previously (23, 24). The supernatant level of NO2 following reduction of NO3 was determined by diazotation using a commercial kit (Cayman).

Visualization of Cellular Compartments

Fluorescence compartmentalization was visualized by a fluorescence light microscope (Nikon, Germany) equipped with a color video camera (Hamamatsu, Germany) and a computer-aided photomultiplier for multiple fluorescence detection. The fluorescence of pyridine nucleotides and rhodamine 123 in living cells was excited at 340 and 480 nm, the emission was recorded at 420 ± 20 nm and 530 ± 10 nm, respectively. To identify changes in the nuclear morphology by NAC, living cells were stained by bisbenzimide (Hoechst 33342, 0.5 μM) in some experiments using an excitation and emission of fluorescence at 340 and 460 nm.

Measurement of Intracellular Reactive Oxygen Species

Subsequently to cultivation, medium was replaced with DPBS, adding dihydrorhodamine 123 (DHR) or hydroethidine (HE) with a final concentration of 0.7 and 20 μM (0.01 and 0.05% dimethyl sulfoxide). In a preliminary experiment, doubling the dose of DHR (1.4 μM final concentration) resulted in almost linear increase in cellular fluorescence. Cellular resuspensions were diluted (10-fold) into the flowcytometric flush buffer (10 mM Hepes, 150 mM saline, pH 7.3). Fluorescence of large and small cells was recorded by an argon laser—equipped flow cytometer (Beckman-Coulter, EPICS-XL) gating the cells in the histogram that displays cellular size and granularity (forward vs. side light scatter) as described previously (25). Fluorescence was excited at 488 nm, recording an emission at 530 nm ± 10 nm and 600 nm ± 10 nm for oxidized DHR and HE, respectively.

Cell Cycle Analysis

The assay followed a described technique (23, 26). Briefly, cells were fixed in ethanol (70%, -20°C, overnight), RNA was digested by RNAase and the DNA was stained with propidium iodide (70 μM). The propidium iodide fluorescence of single cells was measured by flow cytometry using an excitation at 488 nm (argon laser band) and an emission at 600 nm ± 10 nm. The data were recorded by the list mode program and analyzed by the Multicycle program (Phoenix, USA) of the flow cytometer as described previously (23, 27). Further statistical analysis was performed by computer-aided SAS program system.

Assays of DNA Fragmentation, Nitrotyrosine, and Lipohydroperoxide

The DNA was extracted from mixed luteal cells (80% ± 5% large luteal cells, three independent experiments) by a commercial kit (Qiagen) after incubation (4 and 16 h, 37°C) with additives or the vehicle (control). To detect oligonucleosomes, electrophoresis was performed as described elsewhere (28). The content of nitrotyrosine in fixed luteal cells (methanol, -20°C) was assessed by immunocytochemistry and flow cytometry using rabbit anti-nitrotyrosine antibodies (20 μg ml−1, 2 h, room temperature) and anti-rabbit antibodies conjugated with Alexa-488 (unspecific rabbit serum as control). Results are displayed after subtraction of unspecific fluorescence. The concentration of lipohydroperoxides was determined in organic phase lipid extracts from 0.5–1 × 106 cells as described elsewhere (24, 29). Briefly, the assay based on a kit (Cayman) using 13-hydroperoxyoctadienoic acid as standard and ferrous sulphate-ammonium thiocyanate as chromogen. The assay directly determined the unstable hydroperoxides of both saturated and unsaturated lipids (detection of actual transient oxidants that overwhelm defense).

Mitochondrial Transmembrane Potential

Luteal cells were homogenized on ice by a Potter-Elvehjem homogenizer equipped with a Teflon pistil using the buffer (10 mM Hepes pH 7.5, containing 200 mM mannitol, 70 mM sucrose, 1 mM EGTA) provided by a commercial kit (Mito-Iso 1-Kit, Sigma) to prepare mitochondrial fraction. Differential centrifugation (600g, 10 min, 4°C to remove cell debris and nuclei followed by 12,000g, 15 min, 4°C) was employed according to a described method (30). The sediment was resuspended in the storage buffer of the kit (10 mM Hepes pH 7.4, containing 250 mM sucrose, 1 mM ATP, 0.08 mM ADP, 5 mM sodium succinate, 2 mM K2HPO4, 1 mM dithiothreitol) and stored on ice. The protein concentration was adjusted to 1 mg ml−1 with storage buffer. The increase in JC-1 (5,5′,6, 6′-tetrachloro-1,1′,3,3′—tetraethylbenzimidazol carbocyanine iodide) aggregates with the protein concentration of the mitochondrial fraction was tested in a preliminary experiment. Finally, aliquots (30 μl) added into the buffer mentioned below were used to assess mitochondrial inner membrane integrity by observing the presence of the electrochemical proton gradient respective electrochemical potential (Δψ) over the inner mitochondrial membrane. The corresponding assay measured the uptake of the fluorescent carbocyanine dye JC-1 into the mitochondria. In dependence on Δψ and dye partitioning, JC-1 aggregates are formed (19, 31) exhibiting a strong red-orange fluorescence. The buffer (270 μl) of a 300 μl reaction mixture consisted of (final concentrations) 20 mM 4-morpholinepropanesulfonic acid, pH 7.4, containing 110 mM KCl, 10 mM ATP, 10 mM MgCl2, 10 mM sodium succinate, 1 mM EGTA, and JC-1 (1.53 μM, 0.1% dimethyl sulfoxide). A constant volume (total 3 μl) of additives (N-acetylcysteine, rotenone) was used to obtain final concentrations indicated. After incubation (7 min in the dark) at room temperature, the fluorescence was measured by a cytoplate reader (Millipore, Germany) using an excitation at 480 ± 20 nm, recording the emission at 600 ± 10 nm. The fluorescence produced per mg protein was calculated by the ratio (fluorescence (sample) - fluorescence (blank)) × dilution factor/volume of mitochondrial sample (ml) × mg protein/ml and expressed relatively to the basal fluorescence.

The assessment of Δψ in luteal cells was accomplished on the analogy of a described technique (19) with some modifications regarding the use of luteal cells. Routinely, the cells were preincubated (15 min, 37°C) with NAC in DPBS medium, and then the effectors were added (5 min, 37°C). The final concentration of JC-1 was 2 μg ml−1 and that of organic solvent (DMSO) never exceeded 0.05% (v/v). The cells were washed with the flowcytometric flush buffer. The cytofluorimetry (EPICS—XL, Beckman Coulter, equipped with an argon laser) of luteal cells (gated according to the size × granularity histogram) was performed by setting the photomultipliers at 250 V. The compensation of green vs. orange was 25%. The fluorescence was excited at 488 nm recording emissions at 525 ± 20 and 575 ± 20 nm. The data were acquired in list mode and analyzed with the internal software of the flow cytometer.

RESULTS

Responses of Steroidogenesis and Markers of Reactive Oxygen Species to NAC

To inform about possible generation of reactive oxygen species in luteal cells isolated from healthy corpus luteum, formation of nitric oxide and progesterone were assayed in luteal cells from early and midphase gland. In these experiments, contaminating CD11a- and CD31-positive cells were not detectable. The portion of large and small luteal cells from corpus luteum at postovulatory days 4 ± 1 and 14 ± 2 were (% ± SD) 5 ± 3 and 79 ± 10; 52 ± 10 and 21 ± 6, respectively. The corresponding wet weights of the excised glands were (g ± SD) 1.9 ± 0.6 and 6.6 ± 1.2, respectively. The luteal cells from the early (day 4 ± 1) and midphase (day 14 ± 2) corpora lutea secreted progesterone into the medium with a similar rate (Fig. 1). NAC did not exert substantial steroidogenic effects while linoleate (used as a positive control) significantly elevated the progesterone level (Fig. 1). These data indicate that neither functional nor structural regression occurred. Luteal cells from early and midphase glands secreted significantly nitric oxide into the culture medium (0.39 ± 0.08 and 0.50 ± 0.12 μM/h/106 cells).

Figure 1.

Secretory rates of progesterone. Luteal cells from early (n = 4, empty bars) and midphase (n = 6, filled bars) corpora lutea were cultured without additives (basal rate) and with N-acetylcysteine (NAC) or linoleate (Lino) as a positive control (functional-intact cells). The progesterone level of the medium was measured by RIA. Bars stand for mean ± SD, asterisks for P < 0.05 vs. basal rate (ANOVA and post hoc Newman-Keuls test).

To assess the basal ROS level in early and midphase corpus luteum, luteal cells from these stages were loaded with DHR or dihydroethidine to get some information about the source of ROS. The oxidative products of these fluorogenic probes were measured in single cells by flow cytometry gating large and small luteal cells in a size vs. granularity histogram (Fig. 2A). DHR (virtually nonfluorescent) was substantially oxidized to fluorescent R 123 (Fig. 2B). Dihydroethidine was not markedly oxidized to its fluorescent product (data not shown) probably because luteal cells rapidly convert superoxide radicals to peroxides. Supporting evidence was obtained by the significant content of lipohydroperoxide in luteal cells from the midphase gland (Fig. 3) whereas lipohydroperoxide level of luteal cells from early corpora lutea varied at the detection limit (1 μM) of the assay.

Figure 2.

Flowcytometric analysis of the intracellular oxidation of dihydrorhodamine in correlation to luteal cell size and corpus luteum stage. (A), demonstrating a flowcytometric histogram, gating large cells (framed population) from midphase gland by light scattering. FSC vs. SSC denote forward vs. 90° side scatter using an excitation at 488 nm (argon laser band). (B), illustrating the intracellular fluorescence arising from oxidation of dihydrorhodamine (0.7 μM) to rhodamine 123 by gated luteal cells. The broken line denotes the basal fluorescence. (C), showing the flowcytometrically quantified rhodamine 123 (filled bars) vs. basal (empty bars) fluorescence in large, (D) in small luteal cells from early (n = 4) and midphase (n = 6) corpora lutea. Bars stand for mean ± SEM, asterisks for P < 0.05 vs. basal fluorescence (early or midphase stage) using ANOVA and post hoc Dunn's test because of significant variance differences. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Figure 3.

Dose- and time-dependent elevation of the lipohydroperoxide level by N-acteylcysteine. Luteal cells from midphase corpora lutea were incubated without (vehicle-treated control) and with N-acetylcysteine (NAC) in doses (A) and periods (B) indicated. Lipohydroperoxide (LOOH) was assayed in organic phase extract as described under methods. Circles stand for mean, bars for SEM, asterisks for P < 0.05 vs. control (n = 3 independent experiments). Control concentration of LOOH was 2.08 ± 0.53 μM/106 cells.

The fluorescence derived from DHR, which is oxidized by strong oxidants, such as peroxynitrite, to fluorescent R 123, was found to be more intense in large (Fig. 2C) than in small luteal cells (Fig. 2D). N-acetylcysteine affected the intracellular generation of R 123 in dependence on the dose and the luteal stage (Fig. 4). Evidence for peroxynitrite as a component of fluorescence production provided the detection of nitrotyrosine in luteal cells (Fig. 5). Its concentration increased with the dose of NAC (Fig. 5).

Figure 4.

Dose-dependent effect of N-acetylcysteine on the oxidation of dihydrorhodamine to rhodamine 123. Luteal cells from early (broken line) and midphase (line) corpora lutea were exposed to the vehicle (basal rate) or to N-acetylcysteine (NAC) in doses indicated. Rhodamine 123 (R 123) fluorescence in large luteal cells was measured by flow cytometry. Circles denote mean, bars SD, asterisks P < 0.05 vs. basal value (n = 4, ANOVA, Newman-Keuls test).

Figure 5.

N-acetylcysteine induces nitrotyrosine. Immunoreaction was performed in luteal cells from midphase corpus luteum exposed (30 min) to N-acetylcysteine (NAC) in doses indicated or to the vehicle of NAC (basal value). Next, the cells were fixed and incubated with rabbit anti-nitrotyrosine antibodies followed by fluorescent anti-rabbit antibodies and the flowcytometric measurement of fluorescence (empty and filled bars correspond to small and large luteal cells). Control of unspecific fluorescence was obtained by replacing the specific antibody with unrelated rabbit antibodies and subtracted from the specific fluorescence. Bars stand for mean ± SEM (experiments run in triplicate with duplicate determinations), asterisks for P < 0.05 vs. basal value.

Impact of N-Acetylcysteine on Putative Pathways Involved in ROS Production

Despite the different level, the relative increase in fluorescence (basal fluorescence of vehicle-treated cells versus fluorescence production subsequent loading the cells with DHR) was similar in small and large luteal cells. Therefore, we focused on large luteal cells in these analyses.

To investigate pathways associating with NAC-inducible ROS production, luteal cells from midphase corpus luteum were exposed to inhibitors known to block the activity of the corresponding key enzymes (Fig. 6). The inhibitors, their corresponding final concentrations, and inhibited activities were as follows: apocynin (500 μM) and NADPH oxidase, allopurinol (100 μM) and xanthine oxidase, aurintricarboxylic acid (5 and 10 μM) and endonucleases, indomethacin (50 and 100 μM) and prostaglandin endoperoxide H synthases (PGHS), nordihydroguaiaretic acid (NDGA, 10 and 20 μM) and lipoxygenases (NDGA is also known as a hydroxyl radical quencher), ebselen (5 and 50 μM) and PGHS (ebselen is also known as a powerful peroxynitrite scavenger and inhibitor of the cyclooxygenase activity). In some experiments, superoxide anion was scavenged by polyethyleneglycol-conjugated superoxide dismutase (100 μg ml−1). The activity of flavoproteins, including NAD(P)H oxidase, was blocked by diphenylene iodonium (6 and 10 μM). Then, the cells were exposed to NAC in doses shown in Figure 6. Subsequently loading the cells with DHR, the significant responses of fluorescence production to NAC are demonstrated in Figure 6. The results indicate dose-dependent interactions of NAC with the activity of cyclooxygenase (Fig. 6A), the production of peroxynitrite (Fig. 6B), the hydroxyl radical (Fig. 6C), and the activity of endonucleases (Fig. 6D). Neither superoxide dismutase nor inhibitors of superoxide-generating enzymes exerted substantial effects (data not shown). In higher dose, apocynin and allopurinol even tended to increase the oxidation of DHR while dihydroethidine fluorescence remained virtually unchanged (data not shown). Considering the substantial level of nitric oxide, the presence of nitrotyrosine (Fig. 5), and the ebselen-evoked decrease in the oxidation of DHR (Fig. 6B), the results strongly suggest peroxynitrite as a major contributor to the oxidative attacks against DHR. The decrease in the oxidation of DHR by nordihydroguaiaretic acid (Fig. 6C) indicates that the hydroxyl radical participates in the oxidation of DHR. This action is supported by NAC in a dose exceeding 500 μmol l−1 (Fig. 6C).

Figure 6.

Modification of the dihydrorhodamine-oxidizing activity in luteal cells. Luteal cells were isolated from midphase corpus luteum. They were exposed to indomethacin (Indo), ebselen (Ebs), nordihydroguaiaretic acid (NDGA), or aurintricarboxylic acid (ATA), next to N-acetylcysteine (NAC) in doses indicated on the abscissa. Then, the cells were loaded with dihydrorhodamine (0.7 μM) and the intracellular fluorescence (measured by flow cytometry) was expressed relative to the vehicle-treated control (relative fluorescence intensity, RFI). Bars denote mean ± SEM (n = 3–5). Asterisks stand for significant difference from the vehicle-treated control (P < 0.05). Means without common letter differ significantly (ANOVA, post hoc Newman-Keuls test, P < 0.05).

NAC Elicits Compartmentalized Rhodamine 123 Fluorescence

To determine the intracellular distribution of the NAC-inducible fluorescence arising from the oxidation of DHR, the R 123 fluorescence of luteal cells from the midphase corpus luteum was visualized by microscopy (Fig. 7). Fluorescent structures were evident subsequent exposure of NAC to these luteal cells (Fig. 7). Luteal cells from the early corpus luteum did not develop punctated fluorescence (images corresponded to Fig. 7A), indicating stage-dependent sensitivity to NAC.

Figure 7.

Imaging of luteal nuclei and rhodamine 123 fluorescence after intracellular oxidation of dihydrorhodamine. Luteal cells from midphase corpus luteum were incubated for 30 min (A) with the vehicle, (B) with N-acetylcysteine (5 mM), (C) with N-acetylcysteine (5 mM) then dihydrorhodamine (15 min, 0.7 μM). During the last 5 min the nuclei were stained with bisbenzimide H 33342 (Hoechst 33342, 0.5 μM). Dual excitation was performed at 340 nm (emission record at 440 nm, H 33342 fluorescence) and at 480 nm (emission record at 530 nm, rhodamine 123 fluorescence) to visualize nuclei and extra-nuclear pyridine nucleotides as control for the extra-nuclear punctated fluorescence arising from rhodamine 123, the oxidative product of dihydrorhodamine 123. Bars stand for 10 μm.

Decrease in Luteal Cell Viability by NAC

The NAC-inducible oxidation of DHR which is linked to cellular fluorescent structures, may associate with the survival of short-time cultured luteal cells. To test such a relationship, large and small luteal cells from midphase corpus luteum were exposed to the drug in doses indicated in Figure 8. The analysis of the DNA content in single cells subsequent gating large luteal cells by flow cytometry revealed a substantial dose-dependant rise in the hypodiploid (sub G0/G1) counts (representing died luteal cells). However, we did not observe oligonucleosomal fragmentation of DNA (data not shown). Notably, the NAC-provoked rise in hypodiploid cells was suspended by coincubating the cells with aurintricarboxylic acid (Fig. 8D), indicating that NAC is not an inducer of unspecific cytotoxicity. Neither any other inhibitors used above nor inhibitors of nitric oxide synthases (L-NAME, L-NMMA in doses of 0.5 and 1 mM) exerted significant effect on the portion of hypodiploid luteal cells (data not shown).

Figure 8.

Aurintricarboxylic acid prevents induction of hypodiploid luteal cells by N-acetylcysteine. Luteal cells from midphase corpus luteum were exposed to N-acetylcysteine (NAC) in doses indicated or cocultured with aurintricarboxylic acid (ATA, 10 μM). Then, the cells were fixed, RNA was digested, and the DNA was stained by propidium iodide (70 μM). Fluorescence was measured by flow cytometry in gated large luteal cells. The cursors in the flowcytometric fluorescence histogram denote H, sub-G0/G1 phase (hypodiploid, died) cells, C, G0/G1 phase diploid (resting) cells, Q, DNA synthesizing cells, and I, G2/M mitotic cells. The data represent one of three independent experiments.

Mitochondrial Depolarization in Luteal Cells by NAC

NAC evoked a decrease in the JC-1 aggregation as assessed by flow cytometry of single luteal cells. This response was essentially reversed by a protonophore in dependency on the dose of NAC (Fig. 9). The chelator of free calcium, EGTA, attenuated the NAC-evoked depolarization in contrast to the calcium ionophore, A23187 (Fig. 9), suggesting NAC effects on Δψ include also a calcium component. The dose of these additives intended to get some information on the mechanism of the NAC action represents final concentrations which did not disturb the viability (exclusion of trypan blue and propidium iodide, 7 μM) and morphology of living luteal cells as assessed by the flowcytometric histogram corresponding to the cellular size and granularity (forward vs. side angle light scatter, Fig. 2A).

Figure 9.

Reversible NAC-evoked mitochondrial depolarization in luteal cells. Luteal cells from midphase corpora lutea (n=4) were exposed to NAC and to effectors of free calcium, and then loaded with JC-1. The orange fluorescence of JC-1 aggregates which correspond to the mitochondrial potential difference (Δψ) was measured in single cells by flow cytometry. The bars stand for the mean ± SEM of the fluorescence intensity (MFI) expressed as difference between MFI (NAC) and MFI (NAC + additives) relative to MFI (NAC). The asterisks symbolize significant responses (P < 0.05 vs. respective NAC doses). EGTA denotes a Ca2+-chelator (0.5 mM), A23187 a Ca2+-ionophore (1 μM), CCCP a protonophore (1.6 μM). The MFI values ± SEM (without NAC) were 18.0 ± 1.75 (basal), 17.5 ± 1.35 (EGTA), 19.2 ± 1.7 (A23187), 18.2 ± 1.7 (CCCP), 12.2 ± 0.65 (valinomycin). Depolarization by K+-ionophore valinomycin (0.5 μM) used as positive control of polarization was not significantly change by NAC (data not shown).

NAC-Induced Decrease in the Δψ

To directly test whether mitochondrial function is a target of NAC, the mitochondrial fraction from luteal cells was assessed to form JC-1 aggregates. They are generated by the Δψ-dependent partitioning of the stain into the mitochondrial inner space. A strong NAC-elicited decrease in JC-1 aggregates was observed using luteal cells from the midphase corpus luteum (Fig. 10, filled bars) in contrast to luteal cells from an early developmental stage (Fig. 10, empty bars). Mitochondria pretreated with NAC responded to a protonophore with an excessive increase in Δψ (Fig. 10), indicating interference with proton transport. This response returned to the value subsequent treatment with rotenone. On the contrary, rotenone, an inhibitor of respiratory chain complex I, one of the proton pumps and generator of ROS, failed to restore the transmembrane potential of the mitochondrial fraction from early corpora lutea (Fig. 10, empty bars). These results indicate an impact of NAC on mitochondrial function in dependency on the glandular stage, exhibiting different activity of proton pumping into the mitochondrial matrix at a sufficient level of ATP.

Figure 10.

The developmental stage of corpus luteum specifies the NAC-inducible decrease in the transmembrane potential of mitochondria. The mitochondrial fraction of luteal cells from midphase (filled bars, n = 5) and early (empty bars, n = 3) corpus luteum were exposed (30 min) to N-acetylcysteine (NAC, 1 mM), to NAC then to the protonophore, carbonyl cyanide 3-chlorophenylhydrazone, CCCP (2 min) or to NAC, then to CCCP and rotenone (2 min) in concentrations indicated. The mitochondrial transmembrane potential was assessed by the JC-1 technique. The red fluorescence of JC-1 aggregates was expressed relative to the control (incubation with the vehicle, dimethyl sulphoxide < 0.1 %). Bars stand for mean ± SEM, asterisks for P < 0.05 vs. control.

DISCUSSION

NAC has been widely studied in relation to its role in modifying the production of reactive oxygen species by a variety of cell types under physiological and pathophysiological conditions (12, 15, 16, 18). In contrast, the response of the ROS production in luteal cells to NAC remained unknown. Moreover, the source of ROS in the corpus luteum has been attributed to nonsteroidal cells (32). Our data, however, directly demonstrate a significant production of ROS in luteal cells with the oxidation of DHR as a measure. The resulting R 123 fluorescence was visualized by microscopy and quantified by flow cytometry to analyze small and large luteal cells. The results agree with the suggestions of recent studies (33, 34). Furthermore, our data show a dependency of the ROS level on the developmental stage of the corpus luteum and on the luteal cell size. While the level of ROS formation was lower in small luteal cells, the response of the ROS concentration to NAC in small luteal cells resembled those in large luteal cells. These findings are in line with reports (3, 35) that the steroidogenesis is linked to the formation of ROS. Consequently, an elevated level of reactive oxygen species in large luteal cells correlates with higher progesterone synthesis since the steroidogenesis of small luteal cells is known to be lower (17). In contrast to the well established increase in total progesterone production, the steroidogenic rate (i.e., steroidogenesis per cell and hour) did not significantly differ between early and midphase luteal cells, a result that is similar to data from a recent study (36). Therefore, regulative factors additional to the steroidogenic regulation have to be considered. The responses to NAC and to the drug in combination with oxy-radical scavengers revealed some of them. The drug behaved as an antioxidant in cultured luteal cells in a relatively low dose. Surprisingly, the drug lost its antioxidant property with raising the dose. Furthermore, short-time cultured luteal cells from the midphase gland responded to a moderate dose of NAC (1 mM) both with an increase in the portion of dead cells and in the intracellular oxidation of DHR.

That NAC can elicit cell killing is consistent with data from experiments using smooth muscle cells (16) and an ovarian cell line (37). In Chinese hamster ovary cells, a higher intracellular glutathione level following treatment with a moderate dose of NAC (1 mM) led to stimulation of hyperthermia-induced cell death rather than to a protective effect (37). These and our observations are in line with reports showing that thiols, including NAC and cysteine, can form S-nitrosylated intermediates (38, 39). They function as endogenous donors of nitric oxide. Nitric oxide may inhibit steroidogenesis (36). In turn, progesterone exerts antiproliferative effects. Thus, the thiol-free drug, aurintricarboxylic acid was tested to reverse NAC-provoked damage of luteal cell viability. In agreement with our observations, the drug inhibits cell death pathway that does not associate with oligonucleosome formation (40, 41).

In these experiments, we focused on responses of luteal cells from the midphase corpus luteum, since at this stage (preceding glandular regression) the expression of PGHS (42, 43), nitric oxide synthases (44), and the activity of lipoxygenases (45, 46) have been reported to be intensified. During the midphase, the supply of substrates for the oxygenases is augmented by a substantial increase in the lipoprotein lipase activity (47). It mediates luteal uptake of fatty acids, including arachidonate and linoleate, from plasma lipoproteins, relying the oxygenase activities increasingly to exogenous substrate.

To obtain specific information about the reactive oxygen species that oxidize DHR in living luteal cells from the glandular midphase, the cells were exposed to inhibitors of activities known to play an important role in the maintenance of functional corpus luteum, then to NAC.

Both indomethacin and ebselen are inhibitors of the activity of prostaglandin H synthases (PGHS). However, ebselen acts also as a powerful scavenger of peroxynitrite and mimics gluthathione peroxidase activity (48, 49). We found that, as opposed to indomethacin, ebselen strongly decreased the oxidation of DHR. This effect was independent of exposing the cells to NAC. In agreement with our findings that luteal cells produce nitric oxide and contain nitrotyrosine, these results are indicative of peroxynitrite as a major component of the oxidation of DHR. The activity of PGHS itself, however, is likely to associate with prevention from oxidation of DHR, since NAC subsequent indomethacin can increase this oxidation.

As expected by the activities of NDGA, a polyaromatic compound that blocks the activity of lipoxygenases and quenches the hydroxyl radical (50, 51); this drug decreased the oxidation of DHR. Subsequently to a low dose of NAC, the oxidation returned to the control value, indicating NAC-evoked oxidative activity which is not inhibited by NDGA. A higher dose of NAC subsequent NDGA rather supported than abolished the decrease in the oxidation of DHR. This response of the cells is consistent with a hydroxyl scavenger activity of NAC (13) only evident at higher level of the drug.

The action of aurintricarboxylic acid (ATA), a polyaromatic and polyanionic drug known as a stimulator of tyrosine phosphorylation of MAP kinases and a potent inhibitor of DNA topoisomerase II (41, 52), and NAC subsequent ATA resembles that of NDGA. In aqueous solution, ATA generates a stable radical (41, 53) which is likely to be hydroxyl-reactive. However, only ATA was capable of preventing luteal cells from NAC-induced cell death. Moreover, NAC used in a moderate dose of 1 mM abolished the ATA-elicited decrease in the oxidation of DHR on the contrary to NDGA- and ebselen-provoked responses. They were not found to relate to a change in the portion of hypodiploid cells. This uncoupling of the prevention from cell death and the oxidation of the fluorogenic probe renders NAC-induced reactive oxygen species themselves unlikely to be causative for an induction of luteal cell death. Thus, in agreement with data from other reports (2, 7, 36, 46), the activities of prostaglandin H synthases and lipoxygenases are likely to be essential for survival of luteal cells from midphase corpus luteum. During the catalytic cycle, these enzymes consume nitric oxide and hydroperoxide including peroxynitrite (6, 54). These activities may be the rationale of our observations mentioned above. The strong dose dependency of the effect of NAC alone and in combination with the used inhibitors on oxidative attacks against DHR indicates a high sensitivity of the equilibrium between pro-oxidants and anti-oxidants in luteal cells from the midphase corpus luteum. NAC seems to cause an imbalance in this equilibrium. The data, indicating dose-dependent interactions, may be arising from different reactivity of its thiol moiety with hydroxyl and nitric oxide radicals. This difference may most likely result in disulfide and nitroso derivates, exerting divergent effects on the cells.

Because NAC-induced reactive oxygen species per se are unlikely to cause luteal cell death, both cells and mitochondrial fractions from luteal cells were used to assess changes in mitochondrial function evoked by NAC. In luteal cells from the mature midphase corpus luteum, NAC induced a reduction of Δψ in a dose-dependent fashion as measured by the JC-1 aggregation. These orange fluorescent aggregates have been validated to specifically correspond with changes in Δψ (19, 31, 55–57). The NAC-induced dissipation of Δψ was also attenuated by the calcium chelator, EGTA, but not by a calcium ionophore. These treatments were used because free calcium may facilitate oxidation of NAC to a disulfide derivative (38, 39). However, an increase in the dose of NAC in presence of EGTA did not prevent NAC-induced depolarization of luteal cells. In contrast, a protonophore reversed the NAC-caused collapse of Δψ. This observation is consistent with reports (19, 31, 55) that the fluorescence emission of JC-1 changes reversibly in living cells (independent of changes in the pH) from green to orange as the mitochondrial membranes become more polarized. Thus, our data from luteal cells strongly suggest that NAC specifically interferes with mitochondrial proton transport.

Despite the substantial mitochondrial depolarization, Δψ can be assumed to remain sufficiently high to accumulate R 123 (as observed by imaging the R 123 fluorescence) in line with data from other cell types (58, 59). In turn, R 123 binds also to sites freely accessible in deenergized mitochondria (60). Therefore, these sites may contribute to the fluorescence structure arising from the oxidation of DHR to R 123.

As we did not find any published information about assessment of Δψ in luteal cells by the JC-1 technique, we attempted to get further evidence for the findings using a mitochondrial fraction from luteal cells. This approach allowed us to examine the response to both NAC and rotenone, a plasma membrane-impermeable inhibitor of the complex I of the respiratory chain. These tests were performed because the complex I has been shown to be one of the main cellular sources of ROS and to largely determine the mitochondrial activity (21, 22).

Our results demonstrate that in presence of a protonophore the mitochondrial formation of JC-1 aggregates in response to NAC exceeded the control value. Subsequent inhibition of the complex I of the respiratory chain by rotenone, excessive JC-1 aggregation disappeared, indicating an interference of NAC with the reentry of protons into the mitochondrial matrix inner space while the respiratory complex remains intact. In luteal cells from early phase corpus luteum, excessive JC-1 aggregation after exposure to NAC and the protonophore was not observed and rotenone decreased Δψ as expected from results of other cell types (21, 22, 59, 61). Therefore, considering the NAC-provoked reduction of the viability of luteal cells from midphase corpus luteum, our data suggest that proton leakage seems to be beneficial for the mitochondrial function.

In sum, our observations allow proposing a detrimental effect of N-acetylcysteine on luteal cells from the mature corpus luteum in contrast to luteal cells from developing gland. An increase in reactive oxygen species seems to be rather an accompanying affect of NAC than causative for NAC-inducible luteal cell death. The marked dependency of the observed effects on the corpus luteum stage may be important for a therapeutic timing using N-acetylcysteine as a drug.

Acknowledgements

The authors thank Renate Hantel and Ulrike Wiedemuth for excellent technical assistance.

Ancillary