JOURNAL OF CELLULAR PHYSIOLOGY 147:427433 (1991

Replacement of Media in Cell Culture Alters Oxygen Toxicity: Possible Role of Lipid Aldehydes and Glutathione Transferases in Oxygen Toxicity S.J. SULLIVAN,* R.J. ROBERTS, AND D.R. SPlTZ Department of Pediatrics, University of Virginia Hospital, Charlottesville, Virginia 22908 Replacement of media in cell cultures during exposure to hyperoxia was found to alter oxygen toxicity. Following 100 hr of exposure to 95% or 80% 0,, the surviving fraction (SF) of Chinese hamster fibroblasts, as assayed by clonogenicity, was less than 1 x 10- when the culture media was replaced only at the onset of the 0, exposure. Media replacement every 24 hr throughout the hyperoxic exposure resulted in SFs of 1.7 X lo-' (95% 0,) and 1.9 x lo-' (80% 0,j at 95 hr. Cellular resistance to and metabolism of 4-hydroxy-2-nonenal (4HNEj, a cytotoxic byproduct of lipid peroxidation, was examined in cells 24 hr following exposure to 80% 0, for 144 hr with media replacement. These 0,-exposed cells were resistant to 4HNE, requiring 2.6 times as long in 80 p,M 4HNE to reach 30% survival as compared to density-matched normoxia control. Furthermore, during 40 and 60 min of exposure to 4HNE, the 0,-preexposed cells metabolized greater quantities of 4HNE (fmoleicell) relative to control. The activity of glutathione S-transferase (GST), an enzyme believed to be involved with the detoxification of 4HNE, was significantly increased in the 0,-preexposed cells compared with controls. Catalase activity was significantly increased, but no change was found in total glutathione content, glutathione peroxidase, manganese superoxide dismutase, and copper-zinc superoxide dismutase activities at the time of 4HNE treatment in the 0,-preexposed cells relative to density-matched control. The results demonstrate that in vitro tolerance to the cytotoxic effects of hyperoxia can be achieved through media replacement during 0,exposure. Tolerance to oxygen toxicity conferred resistance to the cytotoxic effects of 4HNE, possibly through GST-catalyzed detoxification. These results provide further support for the hypothesis that toxic aldehydic byproducts of lipid peroxidation contribute to hy peroxic injury .

The pathologic, morphologic, and biochemical changes that occur following exposure to hyperoxia have been described in humans, animals, and cultured cells (Frank and Massaro, 1980; Crapo et al., 1980; Bucher and Roberts; 1981, Spitz et al., 1990a). Although a complete understanding of the mechanics of oxygen toxicity is lacking, it has been hypothesized that the toxicity arises from the production of reactive oxygen species (02-,.OH, H202)at a rate exceeding the protective capacity of the cellular detoxification systems (Freeman et al., 1982; Yusa et al., 1984). These highly reactive oxygen species can exert various cytotoxic effects, including inactivation of sulfhydryl proteins, DNA damage, and lipid peroxidation of cellular membranes (Jamieson et al., 1986; Joenje, 1989). Lipid eroxidation of cell membranes can give rise to lipid treakdown products, such as hydroperoxides and aldehydes, which are also hi hly reactive and capable of causing cell injury (Ester auer et al., 1986; Benedetti et al., 1981; Poot et al., 1988; Spitz et al., 1990b). Nutritional modifications that could alter the pro-

fl

0 1991 WILEY-LISS, INC

duction andlor detoxification of lipid aldehydes or hydroperoxides have been shown to affect the severity of oxygen injury. Glutathione (GSH) depletion enhances the cytotoxicity of hyperoxia in vivo, possibly by limiting the availability of GSH for glutathione transferase (GST)catalyzed conjugation with lipid aldehydes or by limiting the detoxification of hydroperoxides by glutathione peroxidase (GPx) (Yam and Roberts, 1979; Deneke et al., 1985).Vitamin E, which is believed to act as an antioxidant by inhibiting lipid peroxidation chain reactions, has been shown to exert a protective effect against O2 toxicity but only in vitamin E-deficient animals or cells in culture (Wispe et al., 1986; Wender et al., 1981;Jacobson et al., 1990; Michiels et al., 19901. The activity of the selenium-dependent GPx, which catalyzes the reduction of toxic lipid hydroperoxides (as well as hydrogen peroxide), can be altered by dietary Received November 28, 1990; accepted February 25, 1991. *To whom reprint requestsicorrespondence should be addressed.

428

SULLIVAN ET AL.

selenium manipulation. Oxygen injury is enhanced in the selenium-deficient state and attenuated in selenium-supplemented animals (Jenkinson et al., 1983, 1984). If lipid aldehydes contribute to the cytotoxicity associated with hyperoxia, tolerance to oxygen-induced injury facilitated by nutritional manipulation should be accompanied by resistance to the cytotoxicity of an aldehydic byproduct of lipid peroxidation. The purpose of this investigation was twofold: 1) to explore the effects of nutritional modification on the cytotoxicity of hyperoxia in a Chinese hamster fibroblast cell line and 2) to examine the cytotoxicity of the lipid peroxidation derived aldehyde, 4-hydroxy-2-nonenal (4HNE), in cells tolerant to the cytotoxicity of hyperoxia as a result of media replacement. The overall goal of this research is to provide a theoretical basis for pharmacologic and/or dietary manipulations that could attenuate the toxic effects of oxygen therapy by altering the production and/or detoxification of the cytotoxic byproducts of lipid peroxidation.

MATERIALS AND METHODS Cells and culture conditions Chinese hamster fibroblasts, designated HA-1, were maintained in Eagle's minimal essential medium (MEM) supplemented with 10% fetal calf serum (FCS; Hyclone). Cells used for experimentation were plated into 60 mm culture dishes containing MEM supplemented with 10% FCS and penicillin-streptomycin (100 Uiml-0.1 mgiml; Sigma). Cell cultures were grown at 37°C in a humidified tri-gas incubator (NUAIRE). Normoxic conditions were 5% C 0 2 and air. Hyperoxic conditions were 80% or 95% O2 and 5% COz. Ambient gas concentrations were varied using a timed injection system which mixed 100% C02, 100% O2 and air (Spitz et al., 1990a). O2 and C 0 2 concentrations were monitored with Beckman OM-11 and LB-2 medical gas analyzers, respectively, calibrated to a certified gas standard (95% 02,5% C02 _t 0.1%). Following disturbance of the internal environment of the hyperoxic incubator, the system was purged with an external source of premixed gas (95% 02,5% COz) to reinstate the desired O2 concentration as quickly as possible. Cell cultures were routinely screened for mycoplasma contamination (Clinical Laboratories, Division of Microbiology, University of Virginia). 4HNE preparation and storage 4HNE was provided by Dr. Hermann Esterbauer, (University of Graz, Graz, Austria). 4HNE was stored in methylene chloride at -20°C (10 mgiml). Aliquots for experiments were prepared under sterile conditions by eva orating the methylene chloride under N2 followed y dissolving the residue in sterile water. The concentration of 4HNE was determined spectrophotometrically at 223 nm (E = 13,750 M-lcm-l). Immediately prior to cell treatment, the aqueous solution of 4HNE was added to MEM without antibiotics or FCS t o minimize the disappearance of 4HNE due to the presence of serum protein (Kaneko et al., 1987).

B

Survival experiments For oxygen survival experiments, cells were plated (5.5-6.5 x lo3 cells/cm2)into 60 mm dishes with 4 ml of complete media and Town in 21% O2 until a cell density of 5.0-6.5 x 10 cells/cmz,was reached. At this time all cultures received fresh media, and a portion of the exponentially growing cell cultures were placed in hyperoxia, the remainder being maintained in 21% 02. Cell density of treatment groups in each experiment was rigorously controlled to avoid the effect of cell density on oxygen toxicity (Spitz et al., 1990a). Cultures undergoing media replacement were removed from incubation every 24 hr, and the media was replaced with 4 ml of complete media. At specific time intervals, cultures were removed from incubation and trypsinized, and the resulting single cell suspension was counted (Coulter Counter) and appropriately diluted such that 100-300 cells were plated for survival analysis in the most diluted replicate, 1,000 to 3,000 in the next dilution, 10,000 to 30,000 in the next dilution, and 100,000 to 300,000 in the least diluted replicate. Each replicate dilution was plated into two or three separate cloning dishes. At each time point, three individually treated dishes were processed as described above. After a cloning period of 8-10 days at 37"C, the colonies were fixed (70% ethanol), stained with Coomassie blue, and counted under a dissecting microscope. The dilution replicates that yielded 50-250 surviving colonies were used for survival analysis. A surviving colony was defined as one containing at least 50 cells. Plating efficiencies for cultures in 21% O2 were 60%-90%. Cell survival in cultures exposed to O2 were normalized to the appropriate normoxia control. After 60 hr of O2 exposure in cultures that did not receive media replacement (and 70 hr of O2exposure in cultures receiving media replacement), some cells detached from the culture dish and floated into the media as evidenced by decreased cell numberidish and visual observation of disorganized cell monolayers using a phase-contrast microscope. When plated for survival, these detached cells had surviving fractions at least 100 times lower than the surviving fraction obtained from those cells that remained attached t o the culture dish (Spitz et al., 1990a). Since the detached cells did not contribute significantly to the survival data for the total cells in the dish, they were considered reproductively inactivated for the purpose of analysis. For 4HNE survival experiments, cultures that had received media replacement every 24 hr throughout the hyperoxic exposure (144 hr of 80% 02),were removed from O2 and allowed to recover for 24 hr in 21% 02,at which time the cell density was 2.0 x lo5 cells/cm2. Cultures were incubated (21% 02,37°C) with aqueous 4HNE in MEM without serum or antibiotics for varying time intervals. Following treatment, each culture dish was washed twice with sterile saline, trypsinized, counted, diluted, and plated into two or three cloning dishes as described above to determine cell survival. Controls for 4HNE survival consisted of densitymatched normoxia-exposed cells (media supplementation every 24 hr for 72 hr). At each time interval, for both groups, three separate culture dishes were treated and processed for cell survival as described above.

MEDIA REPLACEMENT ALTERS OXYGEN TOXICITY

Control plating efficiencies for untreated cells were 60%-90%. Survival data were normalized t o appropriate control plating efficiencies (see Results). 4HNE removal experiments At the onset and at the conclusion of the 4HNE treatment interval, a 0.5 ml aliquot of MEM was removed from each culture and mixed with 0.5 ml of acetonitrileiacetic acid (24:1, viv) to stabilize the sample 4HNE content (Spitz et al., 1990b). The fmoles of 4HNE consumed during the exposure interval was calculated from the difference between time zero and the conclusion of the experiment. Identically treated media samples, not exposed to cells, were collected in triplicate at each time interval for determination of 4HNE consumption by noncellular-mediated processes. This was subtracted from the total amount removed by the cell cultures to determine that amount of 4HNE removed by the cells during the treatment interval. The moles of 4HNE removed by the cells was divided by the number of cells in the culture t o obtain the moles removedicell.

Assay of 4HNE in MEM 4HNE concentration in MEM (no serum or antibiotics), mixed 1:l with 24:l acetonitrileiacetic acid, was assayed by high-performance liquid chromato raphy (HPLC) using an isocratic Beckman System Go d (UV detector model 166) with an ALTEX Ultrasphere ODS (particle size 5 pm; internal diameter 4.6 mm; length 4.5 cm) precolumn and an ALTEX Ultrasphere ODS (particle size 5 pm; 4.6 mm x 15 cm) column. Standard solutions containing a known molarity of 4HNE were prepared in 100% methanol and stored at -20°C. A VARIAN 4270 integrator was used to determine the areas under the HPLC peaks. The HPLC conditions for the assay of 4HNE were mobile phase, acetonitrilei water, (40:60, v:v); flow rate, 0.7 mlimin; wavelength, 223 nm. Under these conditions, the retention time of 4HNE was 9.8-10.0 min. A standard curve for 4HNE was constructed by plotting the area under the peaks obtained from 20 pl injections of each standard solution vs. the concentration of 4HNE in each standard. A linear regression line was then fitted to the standard curve, The correlation coefficient (CC) of the standard curve (from 1to 50 pM 4HNE) was always greater than 0.995. Unknown 4HNE concentrations in 20 pl injections of sample were then determined by comparisons of the areas under the peaks with the standard curve.

f

Total GSH and enzyme assays Total GSH and antioxidant enzyme activities were measured in four treatment groups: 1) 21% 02,with media replacement every 24 hr for 144 hr; 2) 21% O2 density-matched (for 4HNE survival, removal experiments) controls, with media replacement every 24 hr for 72 hr; 3) 80% 02,with media replacement every 24 hr for 144 hr; and 4) 80% 02,with media replacement every 24 hr for 144 hr and allowed to recover for 24 hr in 21% 02.Cell pellets were prepared as previously described (Spitz et al., 1990a). The cell pellets were subjected to one freeze-thaw and homogenized in hypotonic 50 mM phosphate buffer, pH 7.8, with 1.34 mM

429

diethylenetriaminepentaacetic acid (DETAPAC), and an aliquot, was assayed for protein content (Lowry et al., 1951). At least three individual samples were collected and assayed for total GSH, GST activity, GPx activity, catalase (CAT) activity, and superoxide dismutase (SOD) activity. Total GSH content was determined by the method of Anderson (1985) and expressed as pgimg protein. Sulfosalicylic acid precipitation of the Sam le was done immediately after addition of DETAPIC buffer. GST activity using 1 mM l-chloro2,4-dinitrobenzene (CDNB) as substrate was determined by the method of Simons and Vander Jagt (1977) and expressed as mU/mg protein, where 1U of activity is defined as that amount of protein that catalyzes the formation of 1 pmoleimin of thioether. Total GPx activity, using 1.5 mM cumene hydroperoxide as substrate, was assayed by the method of Lawrence and Burk (1976) and expressed as mUimg protein. One Unit of enzymatic activity was defined as that amount of protein that oxidized 1 kmoleimin of NADPH. CAT activity was determined spectrophotometrically by the method of Beers and Sizer (1952) and expressed in kUimg protein (Aebi, 1984).This method measures the disappearance of H202 (10 mM) at 240 nm in the presence of cellular homogenates. Total SOD activity was assayed by the method of Spitz and Oberley (1989) and expressed as Uimg protein, where 1U is defined as that amount of protein that inhibits the nitroblue tetrazolium reduction by xanthine-xanthine oxidase generated 02; to 50% of maximum. This assay utilizes the inhibition of CuZnSOD by cyanide to differentiate between CuZnSOD and MnSOD activities. CuZnSOD activity was determined by subtracting MnSOD activity from total SOD activity, and the errors associated with CuZnSOD activity were calculated using propagation of error theory. Statistical analysis Statistical differences between two population means were examined using Student’s t test. Differences between three or more population means were determined with analysis of variance (ANOVA) and comparisons of individual means were accomplished with Duncan’s new multiple range test. Significance was accepted when P < 0.05.

RESULTS Oxygen toxicity in HA-1 cells with and without media replacement Previous reports have established that the toxic effects of hyperoxia in cell culture are the consequence of a direct insult of O2 on cells rather than O2 interacting in some indirect fashion with cell media to produce a cytotoxic event (Spitz et al., 1990a). Figure 1 shows the clonogenic cell survival curves obtained following exposure to 95% O2 (top panel) and t o 80% O2 (bottom panel) with and without media replacement ever 24 hr. At both O2 concentrations without media rep acement, loss of reproductive integrity is manifest after 48 hr of O2exposure. Disruption of the cell monolayers, as viewed with phase-contrast microscopy, accompanied the loss in clonogenicity. Once clonogenic inactivation began, it proceeded in an exponential fashion at both O2

P

430

SULLIVAN ET AL.

1E-2

t

\

1E-61

40

0

0

z

120

80

160

200

TIME in 95% 02 (hrs)

7

1E-3

1E-5

.\

0-0

control

0-0

media replaced every 24 hr

0

1E-64

0

20

40

60

80

100

120

140

1

TIME in 80% 02 (hrs)

Fig. 1, Oxygen toxicity in control and media replaced cultures. At time 0, cultures (5.0-6.5 x lo4 cells/cm2) received fresh media and were placed into 95% O2 (top) or 80% O2 (bottom).A t s ecific time intervals, three separately treated culture dishes from e a c l treatment group were processed for clonogenic cell survival. At both 0, concentrations, without further media replacement, loss of clonogenicity began after 48 h r and proceeded in a n exponential fashion. When the cultures received media replacement every 24 hr throughout the hyperoxic exposure, cell survival plateaued at 5%and 25% by 144 h r at 95% and 80% 0, concentrations, respectively. The points on the graph represent mean 1 S.D.

*

concentrations. Without media replacement, surviving fractions of less than 1 x lop5were obtained at 94 hr of 95% 0, and 3.5 X at 98 hr of 80% 0,. Plating efficiencies for cultures grown in 21% 0, for 97.5 hr without media replacement fell from 0.67 (time zero plating efficiency) to 0.22. No reduction in plating efficiency was observed prior to 90 hr in 21% 0, without media replacement. Survival data for cultures exposed to 80% 0, or 95% 0, for 94 and 98 hr without media replacement have been normalized to the 0.22 control surviving fraction. In cultures receiving media replacement every 24 hr (Fig. 11,lag periods of 48 and 62 hr are seen prior to the onset of cell death in 80% and 95% 02,respectively. Exponential loss of clonogenicity, as well as disruption of the cell monolayers, proceeded in a fashion similar to that seen in cultures without media replacement for approximately 1 6 2 0 hr. However, in striking contrast to cultures that did not receive media replacement, surviving fractions in the cultures that received media replacement plateaued at 5% and 25% by 144 hr for 95% and 80% 0, exposure conditions, respectively.

Further 95% O2 exposure up to 192 hr did not lead to any reduction in surviving fraction. Plating efficiencies for 21% 0,-exposed cultures receiving media replacement were not different from time zero control plating efficiencies (0.67) throughout the duration of the experiment. 4HNE survival experiments Aldehydic byproducts formed following the peroxidation of lipid membranes have been implicated in the cytotoxicity of hyperoxia (Joenje, 1989; Spitz et al., 1990b). If aldehyde-induced cytotoxicity is involved in the toxicity associated with hyperoxia, tolerance to the cytotoxicity of hyperoxia should be accompanied by resistance to the toxicity of an aldehydic byproduct of lipid peroxidation, such as 4HNE. The following experiments were designed to examine the cytotoxicity of 4HNE in cells that tolerated a preexposure to hyperoxia. Figure 2 illustrates the cell survival following exposure to 4HNE in density-matched media-replaced (21% O,, 72 hr) control cultures and cultures that were pretreated with 144 hr of 80% O2 with media replacement every 24 hr followed by a recovery period of 24 hr in 21% 02.02-preexposed and control cultures were treated with 80 pM 4HNE for varying time intervals up to 120 min. Survival was significantly greater at 40,60, and 120 min of ex osure in the 0,-preexposed cultures relative to contro 'I( < 0.01). The dose modification factor (DMF), which is defined as

Y

time to reach 30% survival in 02-preexposed cells DMF = , time to reach 30% survival in control cells

was 2.6, indicating that, in 80 pM 4HNE, it required 2.6 times longer to effect a similar reduction in clonogenicity in the 02-tolerant cells as compared to control cultures.

4HNE removal experiments Since increased resistance to the cytotoxicity of 4HNE seen in the 0,-preexposed cells could be the result of increased rates of metabolic detoxification of 4HNE from the culture medium, experiments were designed to quantitate 4HNE removal in 02-preexposed cells. Figure 3 illustrates the fmoles of 4HNE removed per cell in density-matched media-replaced controls (21% O,, 72 hr) and 0,-preexposed cells at 20, 40, and 60 min of exposure to 50 pM 4HNE. In previous work it has been shown that heat inactivation of cells reduced the quantity of 4HNE removed to an amount similar to that removed by noncellular-mediated processes, indicating that cellular metabolism was required for 4HNE removal (Spitz et al., 1990b). The fmoles of 4HNE removed by noncellular mediated processes ranged from 0% to 20% of 4HNE added to the media. The quantity of 4HNE removed is expressed on a per cell basis. The fmoles of 4HNE removed per cell by the 0,-preexposed cells at 40 and 60 min of exposure was significantly greater than control (P< 0.05).

43 1

MEDIA REPLACEMENT ALTERS OXYGEN TOXICITY 100

,-

1

0control 02preexposed

'?

1E-3 0

20

40

60

80

100

120

20

Fig. 2. 4HNE survival in 0,-tolerant and control cultures. Following exposure to 144 hr of 80% 0, with media replacement every 24 hr and recovery for 24 hr in 21% 0,, cultures (2.0 X lo5 cellsicm') were incubated (37PC, 21% 0,) for 20, 40, 60, and 120 min with 80 p,M 4HNE. Density-matched media-replaced normoxia controls were treated in an identical fashion. At the conclusion of the treatment intervals, clonogenic cell survival was assayed in both treatment groups. Three separately treated dishes from each treatment group were processed for cell survival at each time interval. Cell survival was significantly greater in the OZyexposedcultures at 40,60, and 120 min of exposure (P< 0.01). The points on the graph represent mean 1 S.D.

*

GST and GPx activity and total GSH content It has been shown in vitro that GST-catalyzed conjugation of 4HNE with GSH provides for the efficient detoxification of 4HNE (Spitz et al., 1991). Therefore, alterations in GST activity may contribute to the resistance to hyperoxic injury and the cytotoxicity of 4HNE seen in the media replaced 02-preexposed cells. Table 1 presents GST and GPx activity and total GSH content in 0,-preexposed and normoxia-exposed cultures. Values are reported on a per milligram protein basis. GST activity in the 80% O2 144 hr media replaced and recovered cells (group 4) was significantly increased (P < 0.05) relative to both their densitymatched media-replaced controls (group 2) and cells exposed to 80% O2 for 144 hr with media replacement but not recovered in 21% oxygen (group 3). Cells cultured for 144 hr in 21% O2 with media replacement (group 1)also had increased GST activity relative to the density-matched media-replaced controls (group 2j. GPx activity was significantly decreased (P < 0.05) in cultures exposed to 80% O2 for 144 hr with media replacement which were collected prior to recovery (group 3) relative to both normoxia-exposed (groups 1 and 2) and 02-exposed following recovery in 21% O2 (group 4). We found no significant differences in total GSH among any of the experimental groups examined. Catalase, CuZn, and Mn SOD enzyme activities Other antioxidant enzyme activities are also given in Table 1. Values are reported on a per milligram protein basis. CAT activity was significantly increased in the 02-exposed and recovered cells ( oup 4) relative to all other groups examined. Total S D activity in cultures

8-

40

TIME in 50 pM

TIME in 80 p M 4HNE (min)

4HNE (min)

60

Fig. 3. 4HNE removal in 0,-tolerant and control cultures. Following exposure to 144 hr of 80% 0, with media replacement every 24 hr and recovery for 24 hr in 21% 0,, cultures (1.0 x lo5 cellskm? were incubated (37"C, 21% 0,) for 20, 40, and 60 min with 50 )I.M 4HNE. MEM was assayed by HPLC for 4HNE content prior to and at the conclusion of the treatment intervals. Samples collected prior to treatment were assayed for determination of the quantity of 4HNE available to the cells a t time zero. Samples collected at the conclusion of the exposure were assayed for determination of the quantity of 4HNE consumed during the exposure interval (corrected for amount lost by noncellular processes). Samples from each of three separately treated dishes were analyzed. The number of moles of 4HNE removed per cell was significantly increased, relative to control, at 40 and 60 rnin of exposure (P < 0.05). The points on the graph represent mean 2 1 S.D.

exposed to 80% O2 with media replacement collected prior to recovery (group 3) was significantly less relative to all other groups. This reduction in total SOD activity was accounted for by a significant decrease in CuZnSOD activity in the 02-exposed cultures prior to recovery (group 3). There were no significant differences in the MnSOD activity among the experimental groups examined.

DISCUSSION Lipid aldehydes and hydroperoxides, produced as a consequence of free radical-initiated peroxidation of polyunsaturated fatty acids in cell membranes, have been implicated in mechanisms of oxygen injury (Joenje, 1989; Spitz et al., 1990b).These lipid aldehydes and hydroperoxides can diffuse from the original site of production and induce cytopathological effects that include cell lysis, inhibition of protein and DNA synthesis, as well as DNA fragmentation and thus may contribute to oxygen injury (Benedetti et al., 1977, 1981; Poot et al., 1988). 4HNE has been identified as one of the most cytotoxic aldehydes produced in significant abundance following the peroxidation of liver microsomes and tumor cells (Winkler et al., 1984; Esterbauer et al., 1986). The attenuation of hyperoxic injury following cell culture media replacement during the hyperoxic exposure (Fig. 1) may result from the removal of lipid aldehydes, or other toxic species, produced as a consequence of the hyperoxic exposure when the old media is discarded. Media replacement may also provide additional exogenous antioxidant capacity, including the

432

SULLIVAN ET AL

TABLE 1. Total GSH content and antioxidant enzyme activities

Group

1. 21% 144 hr 2. 21% 72 hr 3. 80% 144 hr

4. 80%144 hr plus recovery (24 hr 21%)

GPx (mU/md

GST (mU/mg)

GSH (pdmg)

17 f 1 (n = 3) 16 !c 2 (n = 3) 7 1* (n = 3) 14 f 3 (n = 3)

270 i 21 (n = 6) 170 i 12* (n = 12) 250 k 24 (n = 6) 320 f 28** (n = 9)

2.3 i 0.2 (n = 3) 3.6 5 0.2 (n = 9) 2.6 f 0.3 (n = 3) 2.4 k 0.5 (n = 3)

Mn SOD (U/mg)

CuZn SOD

16+1 (n = 3) 20 2 (n = 4) 23 f 1 (n = 3) 24 4 (n = 3 )

25 2 (n = 3) 22 f 3 (n = 4) 9 f 2* (n = 3) 17 f 4 (n = 3)

*

Group 1. 21% 144 hi2. 21% 72 hr

3. 80% 144 hr

4. 80% 144 hr plus

recovery (24 hr 21%)

* *

41 2 (n = 3) 42 2 (n = 4)

+

32 2* (n = 3) 41 f 1 (n = 3)

* *

CAT (kU/mg)

(x lo3)

*

37 3 (n = 3) 31 f 4 (n = 7) 43 !c 6 (n = 3) 56 i 5* (n = 5)

(U/mg)

*

Total GSH content and antioxidant enzyme activities in four treatment groups: 1) 21%02mediareplacedevery24hrfor144 hr; Z)density-matched(forlHNEsurvival and removal experiments), media replaced in 21%Ozevery 24 hr for 72 hr; 3) 80%0 2 , media replaced every 24 hr for 144 hr; and 4) 80% 0 2 , media replaced every 24 hr for 144 hr and allowed tarecover for 24 hrin 21%Oa.n=numberofindividuallycollected samples. Values represent mean k SEM. *Significantly different from all other groups. **Significantly different from groups 2 and 3.

provision of necessary substrates for the synthesis of critical biomolecules required for repair and/or detoxification reactions. We have demonstrated that cultures that are tolerant to hyperoxic injury, as a result of media replacement every 24 hr during the O2 exposure, are also resistant to the cytotoxicity of the peroxidation-derived aldehyde 4HNE (Fig. 2). Furthermore, the O,-tolerant cells, which resist 4HNE cytotoxicity, also metabolize, on a per cell basis, more 4HNE than their densitymatched media-replaced normoxia controls (Fig. 3). These data suggest that media replacement during O,-exposure may have provided the necessary substrates (e.g., sulfnydryl groups) allowing for the more efficient metabolic detoxification of 4HNE. GSTs, a group of multifunctional isoenzymes with remarkably high catalytic activity toward 4HNE and related aldehydes, have been suggested to provide rotection against the cytotoxic effects of lipid aldeydes (Jensson et al., 1986; Danielson et al., 1987). It is believed these enzymes catalyze the conjugation of 4HNE and related aldehydes with glutathione for subsequent transport of the less toxic conjugate across the cell membrane (Ishikawa et al., 1986). H202-resistant cells, derived from the HA-1, are resistant to the cytotoxic effects of 4HNE and have been shown to have increased GST activity as well as increased rates of 4HNE metabolism (Spitz et al., 1990b). H202-resistant cells have also been shown to be resistant to the cytotoxic effects of hyperoxia (Spitz et al., 1990a).It has been hypothesized that the resistance to the cytotoxic effects of 4HNE and hyperoxia in these cells could be in

E

part due t o elevated GST activity and total GSH, allowing for increased metabolic detoxification of lipid peroxidation-derived aldehydes (Spitz et al., 1990bj. Furthermore, GSH depletion with buthionine sulfoximine (BSO) enhances the cytotoxicity and reduces the metabolism of 4HNE in HA-1 and H202-resistantvariants (Spitz et al., 1991). To examine the involvement of GST and GSH in the tolerance to O2 injury seen with media supplementation, GST activity and total GSH content were measured in the four treatment groups (Table l). GST activity was increased in the O,-pretreated cells following 24 hr recovery relative to both the density-matched normoxia-exposed cultures and the 02-pretreated cells prior to recovery. If GST activity is inhibited during 02-exposure by reactive O2 species or lipid breakdown products, removal from the hyperoxic environment could result in an increase in enzyme activity (Pigeolet et al., 1990). We also observed increased GST activity in cultures supplemented for 144 hr in 21% 02.These cultures had 20-30 x lo7 cells per dish by 144 hr of growth in 21% 0,. Despite media supplementation every 24 hr, these cultures may have been nutrient deprived, as evidenced by the acidic nature of the old media prior to replacement, due to the large number of cells per dish. Enhanced GST activity in these cultures may represent changes in response to stresses imposed upon cultures in plateau phase. For this reason, density-matched controls, 80% O2 144 hr-supplemented and 0,-supplemented plus recovery cultures, were all examined with respect to enzyme activities, at cell densities similar to those found in exponentially growing cultures. GSH is believed to react with 4HNE, resulting in the formation of a 4HNE-SG conjugate. GSH is presumed to be consumed in the process (Ishikawa et al., 1986; Jensson et al., 1986). Increased GSH synthesis from amino acid precursors as well as increased GSH regeneration from GSSG could provide additional GSH for GST-catalyzed detoxification reactions in media-supplemented cultures. Once synthesized, GSH may react quickly with 4HNE or related aldehydes, thus removing it from any measurable 001 of GSH (Ishikawa et al., 1986). Although we foundp no changes in total GSH content among the four culture conditions examined, such measurements may not accurately reflect the critical balance between enhanced rates of GSH synthesis, regeneration from GSSG, or consumption as a result of conjugation with 4HNE or related aldehydes in the 02-exposed cells. Catalase, GPx, and SOD have been shown to be susceptible to inactivation by reactive oxygen species (Pigeolet et al., 1990). Decreased activities of GPx and CuZnSOD in cultures exposed to hyperoxia (prior to recovery) could be the result of inactivation of the enzyme by reactive 0, species, possibly H202,produced during O,-exposure (Pigeolet et al., 1990). Support for this hypothesis comes from the fact that when the cultures were removed from hyperoxic stress and recovered in normoxia, the activity of GPx and CuZnSOD returned to that, observed in normoxia-exposed cultures. CAT activity was not changed during 0, exposure but increased significantly following 24 hr recovery in 21% 0,. These data also suggest possible

433

MEDIA REPLACEMENT ALTERS OXYGEN TOXICITY

inhibition of CAT activity during 0,-exposure under our experimental conditions. The elucidation of protective cellular responses to hyperoxia and 4HNE following media replacement during 02-exposure may provide further insights into the mechanism of hyperoxic injury. The results presented here provide further support for the hypotheses that 1)aldehydic byproducts of lipid peroxidation may contribute to hyperoxic injury and 2) tolerance t o oxygen toxicity may involve detoxification of lipid derived aldehydes, such as 4HNE, possibly through GST catalyzed conjugation with GSH. The identification of an accumulated toxic lipid breakdown product or its detoxified metabolite in cultures following hyperoxic exposure would also provide further insight into the involvement of lipid peroxidation in the cytotoxic processes associated with hyperoxic injury in this system.

ACKNOWLEDGMENTS We thank Dr. Hermann Esterbauer for maciouslv providing 4HNE. This work supported by NIH grants F32-HL08277, R01-HL42057, R01-HL33964. and DK38942. Technical assistance was provided bv D.T. Adams and C.M. Sherman. LITERATURE CITED Aebi, H. (1984) Catalase in vitro. Methods Enzymol., 105:121-126. Anderson, M.E. (1985) Tissue glutathione. In: Handbook of Methods for Oxygen Radical Research, Creenwald, R.A., ed. CRC Press, Inc., Boca Raton, FL, pp. 317-323. Beers, R.F., and Sizer, I.W. (1952) A spectrophotometric method for measuring the breakdown of H,O, by catalase. J . Biol. Chem., 195: 133-1 40. Benedetti, A,, Barbieri. L., Ferrali, M., Casini, A.F., Fulceri, R. and Comporti, M. (1981) Inhibition of protein synthesis by carbonyl compounds (4-hydroxyalkenals) originating from the peroxidation of liver microsomal lipids. Chem.-Biol. Interact., 35:331-340. Benedetti, A., Casini, A.F., and Ferrali, M. (1977) Red cell lysis coualed to the aeroxidation of liver microsomal lioids. ComuartmentalLation of the hemolytic system. Res. C o m m k Chem. Pathol. Pharmacol., 17519-528. Bucher. J.R.. and Roberts. R.J. (1981)The develoament ofthe newborn rat lung in hyperoxia: A dose-response stidy of lung growth, maturation, and changes in ant,ioxidant enzyme activities. Pediatr. Res., 15999-1008. Crapo, J.D., Barry, B.E., Foscue, H.A., and Shelburne, J. (1980) Structural and biochemical changes in rat lungs occurring during exposures to lethal and adaptive doses of oxygen. Am. Rev. Respir. Dis., 122:123-143. Danielson, H., Esterbauer, H. and Mannervik, B. (1987) Structureactivity relationships of 4-hydroxyalkenals in the conjugation catalyzed by mammalian glutathione transferases. Biochem J., 247:707-713. Deneke, S.M., Lynch, B.A. and Fanburg B.L. (1985) Transient depletion of lung glutathione by diethylmaleate enhances oxygen toxicity. J. Appl. Physiol., 58:571-574. Esterbauer, H., Benedetti, A., Lang, J., Fulceri, R., F a d e r , G., and Comporti, M. (1986) Studies on the mechanism of formation of 4-hydroxynonenal during microsomal lipid peroxidation. Biochim. Biophys. Acta, 876:154-166. Frank, L., and Massaro, D. (1980) Oxygen toxicity. Am J. Med., 69tl17-126. .. __. Freeman, B.A., Topolosky, M.K., and Crapo, J.D. (1982) Hyperoxia increases oxygen radical production in rat lung homogenates. Arch. Biochem. Bioahvs.. 216:477-4!34 Ishikawa, T., Gsterbauer, H., and Sies, H. (1986) Role of cardiac ~~

d ~ > - - -

glutathione transferase and of the glutathione S-conjugate export system in biotransformation of 4-hydroxynonenal in the heart. J. Biol. Chem., 261:1576-1581. Jacobson, J.M., Michael, J.R., Mokarram, H.J., and Gurtner G.H. 11990) Antioxidants and antioxidant enzymes protect against pulmonary oxygen toxicity in the rabbit. J. Appl. Physiol., 68:12521259. Jamieson, D., Chance, B., Cadenas, E., and Boveris, A. (1986) The relation of' free radical production to hyperoxia. Annu. Rev. Physiol., 48:703-719. Jenkinson, S.G., Lawrence, R.A., Burk, R.F., and Gregory, P.E. (1983) Non-selenium-dependent glutathione peroxidase activity in rat lung: Association with lung glutathione S-transferase activity and the effects of hyperoxia. Toxicol. Appl. Pharmacol., 68r399404. Jenkinson, S.G.,, Long, ,R.J.,and Lawrence, R.A. (1984) Endotoxin protects selenium-deficient rats from hyperoxia. J. Lab. Clin. Med., 103:143-1 5 1. Jensson, H., Guthenberg, C., Alin, P., and Mannervik, B. (1986) Rat glutathione transferase 8-8, a n enzyme efficiently detoxifying 4hydroxyalk-2-enals. FEBS Lett., 203:207-209. Joenje, H. (1989) Genetic toxicology of oxygen. Mutat. Res., 219:193208. Kaneko, T., Honda, S., Nakano, S., and Matsuo, M. 11987) Lethal effects of a linoleic acid hydroperoxide and its autoxidation products, unsaturated aliphatic aldehydes, on human diploid fibroblasts. Chem.-Biol. Interact., 63r127-137. Lawrence, R.A., and Burk, R.F. 11976) Glutathione peroxidase activity in selenium-deficient rat liver. Biochim. Biophys. Res. Commun., 71:952-958. Lowry, O.H., Rosebrough, N.J., Farr, A.L., and Randall, R.J. (1951) Protein measurement with folin phenol reagent. J. Biol. Chem., 193:265-275. Michiels. C.. Toussaint. 0..and Remzicele. J. (1990! Comoarative studv of oxygen toxicity in human fibroblasts and endotfielial cells. j. Cell. Physiol., I44:295-302. Pigeolet, E., Corbisier, P.. Houbion. A., Lambert, D., Michiels, C., Raes. M.. Zacharv. M.. and Remacle. J. (1990) Glutathione aeroxidase,' superoxide dlsmutase, and catalase inactivation by peioxides and oxygen derived free radicals. Mech. Ageing Dev., 51:283-297. Poot, M., Verkerk, A,, Koster, J.F., Esterbauer, H., and Jongkind, J.F. (1988) Reversible inhibition of DNA and protein synthesis by cumene hydroperoxide and 4-hydroxy-nonenal. Mech. Ageing Dev., 43: 1-9. Simons, P.C., and Vander Jagt, D.L. (1977)Purification of glutathione S-transferases from human liver by glutathione-affinity chromatography. Anal. Biochem., 82:334-341. Spitz, D.R., Elwell, J.H., Sun, Y.. Oberley, L.W., Oberley T.D., Sullivan, S.J. and Roberts, R.J. (1990aJ Oxygen toxicity in control and H,O,-resistant Chinese hamster fibroblast cell lines. Arch. Biochem. Biophys., 279:249-260. Spitz, D.R., Malcolm, R.R., and Roberts, R.J. (1990b) Cytotoxicity and metabolism of 4-hydroxy-2-nonenal and 2-nonenal in H,O,-resistant cell lines. Biochem. J., 267:453-459. Spitz, D.R., and Oberley, L.W. (1989) An assay for superoxide dismutase activity in mammalian tissue homogenates. Anal. Biochem., 7 74.8-1 A

_ . Y . V

_Y.

Spitz, D.R., Sullivan, S.J., Malcolm, R.R., and Roberts, R.J. 11991) Glutathione dependent metabolism and detoxification of 4-hydroxy2-nonenal. (submitted). Wender, D.F., Thulin, G.E., Smith, G.J.W. and Warshaw, J.B. (1981) Vitamin E affects lung biochemical and morphologic response to hyperoxia in the newborn rabbit. Pediatr. Res., 15:262-268. Winkler, P., Lindner, W., Esterbauer, H., Schauenstein, E., Schaur, R.J., and Khoschsorur, G.A. 11984) Detection of 4-hvdroxvnonenal as a aroduct of haid Deroxidation in native Ehrlich" ascit"es ~ ~tumor . . . ~ ~~ ~ ~ .~_ . cells.sBiochim. Bi'ophis. Acta, 796:232-237. Wispe, J.R., Knight, M., and Roberts, R.J. 119861h i d ueroxidation in newborn rabbcts: Effects of oxygen, lipid emulsion, and vitamin E. Pediatr. Res.. 20505-510 Yam, J , and Roberts, R J. (1979) Pharmacological alteration of oxygen-induced lung toxicity Toxicol Appl. Pharmacol , 4 7 367375 Yusa, T , Crapo, J D , and Freeman, B A (1984) Hvperoxia enhances lung and liver nuclear superoxide generation. Biochim. Biophys. Acta, 798:167-174. ~

~~~

~~~

~~

~

~~~~~~~~~