Kinetic studies on the removal of extracellular hydrogen peroxide by cultured fibroblasts.

To investigate the function of antioxidant enzymes in intact cells, we examined the removal of extracellular H2O2 by cultured fibroblasts (IMR-90). H2O2 concentration dependence of the reaction rate was interpreted as that the process involves two kinetically different reactions (referred to as reactions 1 and 2). Reaction 1 was characterized by a relatively low Km value (about 40 microM), and reaction 2 by linear dependence of the rate up to 500 microM H2O2. The magnitude of reaction 1 was reduced by treatment of the cells with diethyl maleate or 6-amino-nicotinamide, while reaction 2 was inhibited by 3-amino-1,2,4-triazole treatment. It was concluded that reactions 1 and 2 are principally due to GSH peroxidase and catalase, respectively. The values of kinetic parameters were estimated by curve-fitting, and it was inferred that 80 to 90% of H2O2 is decomposed by GSH peroxidase at H2O2 concentrations lower than 10 microM. The contribution of catalase increases with the increase in H2O2 concentration. The intact cells showed a low catalase activity (about 15%), as compared with the activity found in the solubilized cells. The low catalase activity was ascribed to the latency of the enzyme caused by localization in peroxisomes. Fibroblasts also removed intracellular H2O2 generated by menadione. Treatment with diethyl maleate greatly impaired the H2O2-removing capability and caused H2O2 efflux into the medium.


Kinetic Studies on the Removal of Extracellular Hydrogen Peroxide by Cultured Fibroblasts*
(Received for publication, June 2, 1993, and in revised form, August 9, 1993) Zbaraki 305, Japan I , To investigate the function of antioxidant enzymes in intact cells, we examined the removal of extracellular Hz02 by cultured fibroblasts . H202 concentration dependence of the reaction rate was interpreted as that the process involves two kinetically different reactions (referred to as reactions 1 and 2). Reaction 1 was characterized by a relatively low K, value (about 40 w), and reaction 2 by linear dependence of the rate up to 500 I.~M HzO2. The magnitude of reaction 1 was reduced by treatment of the cells with diethyl maleate or &aminonicotinamide, while reaction 2 was inhibited by 3-amino-l,2,4-triazole treatment. It was concluded that reactions 1 and 2 are principally due to GSH peroxidase and catalase, respectively.
The values of kinetic parameters were estimated by curve-fitting, and it was inferred that 80 to 90% of HzOz is decomposed by GSH peroxidase at H202 concentrations lower than 10 w. The contribution of catalase increases with the increase in H202 concentration. The intact cells showed a low catalase activity (about lS%), as compared with the activity found in the solubilized cells. The low catalase activity was ascribed to the latency of the enzyme caused by localization in peroxisomes.
Fibroblasts also removed intracellular Hz02 generated by menadione. Treatment with diethyl maleate greatly impaired the Hz02-removing capability and caused Hz02 efflux into the medium.
In living cells, H202 is mainly generated from mitochondria, microsomes, and peroxisomes (1, 2). Peroxisomes contain H20z-producing oxidases, while mitochondria and microsomes produce the superoxide anion as a by-product in the O2 reduction, and the anion quickly dismutates to H202 and 0 2 spontaneously or by the action of superoxide dismutase. H20z itself is not very reactive with cellular constituents, but in the presence of transition metal ions and appropriate reductants it is converted to the hydroxyl radical which is highly reactive with organic compounds and is hazardous to living cells. In this relevance H202 is regarded as a key substance in the oxygen toxicity, and its elimination is important for the cell.
GSH peroxidase and catalase are involved in the decomposition of Hz02, but they are differently localized in the cell. GSH peroxidase activity is found in cytoplasm and mitochondria (3), whereas catalase is localized in peroxisomes and lower * The costs of publication of this article were defrayed in part by the "advertisement" in accordance with 18  density particles, and very little enzyme is free in the cytoplasm (4). The two enzymes are also different in kinetic behavior. Catalase shows a very high K,,, value for H202, and its reaction apparently follows the first-order kinetics, but GSH peroxidase shows a relatively low K, value, obeying the Michaelis-Menten kinetics. On the basis of their kinetics, it is believed that in intact cells GSH peroxidase is more effective at relatively low H20z concentrations, while catalase is more effective at high concentrations, as exemplified by the studies on erythrocytes (5-7) and hepatocytes (8).
The reduction of H202 by GSH peroxidase is accompanied by the oxidation of GSH, which is restored by GSSG reductase using NADPH as a reductant. The supply of NADPH is generally rate-limiting in the reaction sequence, and it is mainly produced by the pentose phosphate pathway and cytosolic isocitrate dehydrogenase. KaufFman et al. (9) have shown that the latter enzyme supplies the majority of NADPH in hepatocytes.
Although plenty of qualitative knowledge has been accumulated on the H20z-eliminating enzymes, we still do not understand quantitatively their behavior in the living cells, principally because of the difficulty in determining the intracellular H202. In addition, hepatocytes have been used frequently in studies on the oxidative stress and the related enzyme systems, but the properties of other cells are not well understood.
In this study we evaluated the activity of cultured human fibroblasts (IMR-90) in removing extracellular H202 and analyzed the reaction in terms of the enzyme kinetics. To distinguish the actions of the H202-eliminating enzymes, the effects of several inhibitors (AT,' DEM, and 6-AN) were examined. AT inactivates catalase irreversibly, only when the enzyme is functioning (10). DEM reacts with GSH irreversibly, depleting it from the cells. 6-AN, a nicotinamide derivative, is metabolized to the 6-amino derivative of NADP, which is a strong inhibitor of 6-phosphogluconate dehydrogenase and blocks the pentose phosphate pathway (11).
In this study we also used menadione to generate H202. Quinone derivatives including menadione are reduced by NADH-or NADPH-linked oxidoreductases and form H20z via superoxide anion (12-14). It can be used to estimate the capability of the cells to decompose intracellularly generated HzOz.

H202 Removal
Cell Culture-Human fibroblasts, strain IMR-90, were grown in Eagle's basal medium (Life Technologies Inc.) supplemented with 10% fetal bovine serum (Whittaker). They were subcultivated every 3 or 4 days with the split ratio of 1:2. For the experiments, the cells between the 30th and 50th passages were used. They were plated in dishes with 60-mm diameter (2-4 x IO5 cells per dish) and were incubated for 2 days before the experiments.
When the effects of inhibitors were examined, the cells were treated with 10 m~ AT (for 6 h), 1 m~ DEM (for 3 h), or 0.5 m~ 6-AN (for 5 h) before the experiments. The cell viability was tested by exclusion of nigrosin (0.05% in PBS). No change in the viability was observed after the treatment with inhibitors.
Measurement of HzOz Removal by Fibroblasts-The cells grown in a 60-mm dish were rinsed three times with warmed PBS (140 m~ NaCl, 2.7 m~ KCI, 0.9 m~ CaClZ, 0.5 m~ M&12, and 10 m~ phosphate at pH 7.4) plus 5.6 m~ glucose. The dish was placed in a small plastic box, of which temperature was maintained at 37 "C in a water bath. For the measurement of removal rate, 5 ml of PBS + glucose containing an appropriate concentration of HzOz (2-500 p~, warmed at 37 "C) was added to the dish, and the dish was shaken at 100 cycledmin. At appropriate time intervals (1 to 3 min), 50-pl portions of the medium were withdrawn and were subjected to the H,Oz determination, as described below. No change in the cell viability was observed after incubation with 500 p HzOz for 10 min, as tested by the nigrosin exclusion.
Determination of HzO,-The method of Williams et al. (15) using the peroxyoxalate luminescence was applied with some modifications. A vial (17 mm in diameter) contained 60 pl of water, 90 pl of PBS, and 200 pl of 0.005% (v/v) triethylamine solution in methanol, to which 50 pl of a test solution was added. The vial was placed in a chemiluminescence detector (Aloka BLR-BOl), and the reaction was started by addition of 200 pl of ethyl acetate containing 0.06% bis(2,4,6-trichlorophenyl) oxalate and 0.005% perylene. The intensity of luminescence was integrated between 10 and 20 s after the reaction was started. HzOz was quantT1ed by comparison with the results for standard HzOz solutions. The method gave linear response between 0.1 nmol and 0.1 p o l of HzO2. The concentration of standard HzOz solution was determined from the absorbance at 230 nm and the extinction coefficient of 61 cm-'. Calculation ofHzOz Removal Rate-Time courses of HzOz removal by fibroblasts were measured as described above, and natural logarithms of HzOZ concentrations were plotted against the reaction time. The pseudo first-order rate constant k' (/mid was obtained from the slope of the linear part of the plots by the least-square method and was converted to the reaction rate u (nmol/min/mg of protein) as: where [HzOzl, is the HzOZ concentration at the midpoint of the measurement. In the absence of inhibitors, the plot gave a straight line after 1 min of reaction (cf. Fig. 2), and k' was calculated from the data obtained between 1 and 10 min. When using the cells treated with the inhibitors, one had to wait for 3 4 min before the reaction reached a steady-state. Usually, 6-10 data points obtained in the initial part of the reaction (less than 30% of the total) were used for the calculation. Assay for GSH Peroxidase Activity ana' Total Glutathione "Fibroblasts cultured in 60-mm dishes were solubilized with 0.5% Triton X-100 and 4 m~ EDTA in 0.1 M phosphate buffer (pH 7.4). Samples from two dishes were combined and were subjected to assays for GSH peroxidase activity and total glutathione content. The GSH peroxidase activity was measured by the method of Awasthi et al. (16) with some modifications. The reaction mixture (600 pl in a final volume) contained 0.2 m~ NADPH, 4 m~ GSH, 1 unit of glutathione reductase, 4 m~ EDTA, and 4 m~ sodium azide in 0.1 M phosphate buffer at pH 7.4. After the addition of solubilized cell (200 pl, about 0.2 mg of protein), the reaction was started by the addition of 2.4 m~ H202 (50 pl), and the absorbance change of NADPH at 340 nm was followed at 37 "C. Under the assay conditions, the rate of non-enzymatic reaction was considerably high (about 10 nmoymin), and it was subtracted from the observed rate. The activity was expressed as nanomoles of HzOz consumed min/mg of protein, which was equivalent to the NADPH consumption rate. For the determination of total glutathione content, 1 ml of cold 6.5% trichloroacetic acid solution was added to 300 pl of the solubilized cell, and the precipitate was removed by centrifugation at 10,000 rpm for 10 min at 0 "C. Glutathione content of the supernatant was determined by the method of Tietze (17).
Assay for Catalase Actiuity-Fibroblasts were solubilized, as described above, and 0.3 ml 0; the solution was added to 2.7 ml of 50 m~ potassium phosphate buffer (pH 7.0) containing HzOz (10 m~ in final concentration). The absorbance change at 240 nm was recorded for 10 where u, and u', are the observed and calculated values, since the standard error for vi was found to be roughly proportional to vi itself. For the curve-fitting, 39 data points were used covering the HzOz concentration 2 to 500 p~. It was confirmed that varying the starting parameter values does not influence the final values. The confidence intervals were determined by the F test (19), although the values are only approximate because of the nonlinearity of the model.

RESULTS
Permeability and Difiusibility of H202 in the Cells--In this work, the H202 removal activity of fibroblasts was measured by determination of extracellular H202, and it is important whether H202 diffuses into the cell rapidly enough. The fibroblasts attach to the surface of a culture plate and spread out forming a thin monolayer, of which thickness is estimated as small as 1 pm or less. Unfortunately, the diffusion coefficient of H202 in the cell membrane is unknown, but its diffusion behavior is thought to be similar to that of O2 in cellular environments. The diffusion coefficients of O2 in the cell membrane and matrix have been reported to be 4-15 x cm2 (20, 21). The mean diffusion distance ( d ) in time t for particles with a diffusion coefficient of D is given as d = 2 f i , and the mean time required for O2 (or HzOz) to diffuse for 1 p~ is calculated to be less than 0.4 ms. Thus, it is thought that Hz02 diffuses into the cell quickly enough in the time span of the experiments.
Removal of Extracellular H20z by Fibroblasts-Fibroblasts decomposed extracellular H202 considerably rapidly (Fig. 1). the plot deviated from the straight line, indicating a deviation from the first-order kinetics.
As described under "Experimental Procedures," the removal rate was calculated from the initial part of the reaction (less than 30% of the total) and was plotted against the H20z concentration (Fig. 3). The rate of Hz02 removal was proportional to the cell density, showing that the activity is apparently independent of the stage of proliferation. Fig. 4 shows the concentration dependence of the Hz02 removal reaction. The plot was biphasic with respect to HzOz concentration; the slope at lower concentrations was definitely higher than that observed at higher concentrations.
The Effects of Inhibitors- Fig. 5 shows the effect of pretreatment of the cells with a catalase inhibitor, AT. The slope of the plot at higher H202 concentrations was considerably decreased, whereas the reaction at lower concentrations was much less influenced. Fig. 6 shows the effect of DEM treatment. In contrast to the results for the AT-treated cells (Fig. 5), the slope at lower HzOz concentration was definitely reduced, whereas the reaction at higher concentrations was not appreciably affected. Tables I and I1 show the results of enzyme and glutathione assays of solubilized cells. As shown in Table I, AT reduced the catalase activity to Ut3 of the original value, but DEM showed no appreciable effects. As shown in Table 11, DEM treatment decreased the glutathione content to 5% of the control value, but AT showed no effect. GSH peroxidase activity was not affected by either treatment. Fig. 7 shows the effect of 6-AN on the H20z removal reaction. It reduced the reaction rate at lower H202 concentrations, but little affected the slope of the plot at higher concentrations. The effect of 6-AN was similar to that of DEM, although the extent was significantly less.
Curve-fitting Study-From the results shown in Figs. 4-7, it is presumed that two different types of kinetics are involved in the Hz02 removal reaction. One is of the Michaelis-Menten type with a relatively low K,,, value, and its rate is close to the saturation at 100 1.1~ Hz02. The other is characterized by a large K,,, value, and the rate increases linearly up to 500 p~. Equation 3 (together with Equations 4 and 5) was fitted to the Hz02 removal data by the nonlinear least-square method.
As seen in Fig. 4, the fitting was satisfactory. Table I11 shows the estimated parameter values, together with the 70% confidence limits, which roughly correspond to the range for * 1 S.D.
in the normal distribution.
From the kinetic properties and the effects of inhibitors, it is  (Fig. 4). The dotted lines were calculated from Equations control experiments (Fig. 4). The dotted lines were calculated from presumed that reactions 1 and 2 are attributable to GSH peroxidase and catalase, respectively. On this assumption, we tried to calculate the Hz02 removal rate by the AT-and DEMtreated cells from Equations 3-5 as a function of HzOz concentration. "he values of kinetic parameters were estimated from those for the untreated cells (Table 111) and from the analytical data (Tables I and 11) on the solubilized cells, as follows. It has been reported that the apparent V , , and K,,, values of GSH peroxidase for H202 are proportional to the GSH concentration (22). The GSH concentration can be approximated to that of the total glutathione (Table II), because it is mostly reduced in the   (Fig. 4). The dotted lines were calculated from Equations 3-5 and with V , , = 11.7 nmol/midmg of protein and K, = 11.55 p~ (30% of the control values shown in Table 111). The k value was assumed to be the same as the control value (0.201 nmol/midmg of p r o t e i d p concentration).

Fitted ualues and confidence limits of kinetic parameters for the H202
removal reaction by untreated fibroblasts Reactions 1 and 2 are those following the Michaelis-Menten and the first order kinetics, respectively, as defined in Equations 3-5 in the text. cell. The V , , value is also proportional to the enzyme concentration, which can be represented by the enzyme activity of the solubilized cells (Table 11). The K value was taken as proportional to the total catalase activity ( Table I). The results of the calculation were found to be consistent with the observation, as shown in Figs. 5 and 6. For the 6-AN-treated cells, as shown in Fig. 7, the observed data could be explained as that both V , , and Km values are reduced to approximately 30% of the original values. This is attributable to a decrease in GSH concentration resulting from the lowered NADPH supply, as described above. It is possible to calculate the rates of reactions 1 and 2 from Equations 4 and 5 and the values of the kinetic parameters for the untreated cells (Table 111). Fig. 8 shows the contribution of reaction 1 to the Hz02 removal, expressed as vl/(uz + u2), as a function of Hz02 concentration. When the HzOz concentration is lower than 10 p~, the fraction of H202 removed by GSH  (Table 111). The dotted and dashed lines were calculated with the k values fixed at the lower and higher 70% confidence limits (Table 111) Removal of Hz02 Generated by Menadione- Fig. 9A shows the H202 generation observed on addition of menadione to the cells. In the control cells, the Hz02 concentration reached a plateau in 20 to 30 min. Presumably, a steady-state was reached between the H202 generation and decomposition. The extracellular H20z concentration was less than 1 p~ under the experimental conditions. As shown in Fig. 9A, pretreatment of the cells with AT or 6-AN exerted no appreciable effects. This is interpreted as that catalase does not contribute much to the Hz02 removal at the low Hz02 concentration and that 6-AN does not completely inhibit the NADPH supply. In the DEMtreated cells, on the other hand, a considerable increase in HzOz generation was observed (Fig. 9B), indicating that DEM greatly impaired the HzOz-removing capability of the cell. The rate of H202 generation increased with time and, after 20 min, reached 1.5 and 3.5 nmollmidmg of protein at, respectively, 0.1 and 0.2 m M menadione. The results confirm that the GSHdependent system is more important in removing H20z at a low generation rate. DISCUSSION The Kinetics of H2OZ Removal by Fibroblasts-In the present study we analyzed quantitatively the H202-removing activity of fibroblasts by determining the rate of H2O2 removal in the extracellular medium. The kinetic measurement was made possible by the chemiluminescence method, which allowed quick and sensitive determination of H20z. In principle, the present technique is applicable to any other adhesive cells, but in the case of thicker cells the rate of H202 diffusion may have to be considered in kinetic analyses.
Contributions of the Enzymes to H202 Elimination-The present results indicate that the removal of extracellular H20z by cultured fibroblasts involves two reactions showing different H202 concentration dependences. From the effect of DEM ( Fig.  6 and Table 111), it is thought that reaction 1 is mostly due to GSH peroxidase. The nonenzymatic reaction between GSH and H202 is negligibly slow at physiological pH (5). Reaction 2 was effectively inhibited by AT ( Fig. 5 and Table III), and it is mostly attributable to catalase. Nonenzymatic catalysts such as transition metal ions may participate in the reaction, but their contribution is estimated to be small, because their concentrations are low, and because the specific activity of catalase is extremely high. It is generally believed that at low H202 concentration GSH peroxidase is more effective in removing HzOz, but the contribution of catalase increases as the H202 concentration increases. The idea was confirmed quantitatively by the present study, as shown in Fig. 8. It was calculated that, at H20z concentrations lower than 10 p~, 80-90% of H202 is decomposed by GSH peroxidase (or reaction l), and its contribution is lower at higher H202 concentrations. Under normal conditions, the intracellular H202 concentration is low, and it is thought that GSH peroxidase is the primary enzyme in removing Hz02.
Latency of Catalase Activity-The k value of reaction 2 for the untreated cells was about 0.2 nmollmidmg of proteidm concentration (Table 1111, where the unit (nmollmidmg of proteidw concentration) is equivalent to mllmidmg of protein. On the other hand, the catalytic activity of the solubilized cell was 0.459/midmg at 10 m M H20z (Table I). Because the volume of the test solution was 3 ml, the activity is equivalent to the k value of 1.38 mllmidmg. On the assumption that the rate of the reaction is proportional to H202 concentrations up to 10 mM, it is calculated that only 15% (&20% for 70% confidence range) of the total catalytic activity was effective in removing the extracellular H202. In intact cells, catalase is confined in peroxisomes and smaller structures (4), and isolated peroxisomes show only 10-15% of the total activities (23,24). It is thought that the low enzyme activity is due to the localization of the enzyme.
The Michaelis Constant of GSH Peroxidase-According to The above coefficient values have been obtained for bovine GSH peroxidase, but they would give a good approximation for human enzyme. The total glutathione content of fibroblasts was about 39 nmoVmg of protein (as GSH), as shown in The Rate of NADPH Supply-It is generally thought that the NADPH supply is rate-limiting in the H20z-removing system using GSH. This seems to be true also in fibroblasts, because the estimated V,, value (Table 111) was smaller than the GSH peroxidase activity in the solubilized cells (Table 11).
The NADPH supply rate in hepatocytes has been estimated to be 5-8 nmoVmid1O6 cells determined in the present study (about 39 nmollmidmg) was relatively close to that obtained by Sies and Summer (28) but considerably greater than the other values. The discrepancy may have arisen from the difference in the experimental methods. It has been reported that glucose-6-phosphate dehydrogenase is inhibited by NADPH, and the inhibition is counteracted by GSSG (29). As a result, the pentose phosphate pathway is stimulated under the oxidative stress (30). It is thus probable that, in the experiments using high concentrations of organic hydroperoxide or H202, a higher NADPH supply rate is observed as a result of the NADPH and GSH oxidation. Areported value for the activity of GSSG reductase in fibroblasts is high enough (36 nmoVmidmg of protein) (31) to support the HzOz removal rate observed here.
The 6-AN treatment caused a decrease in GSH peroxidase activity to 30% (Fig. 7). It has been reported for fibroblasts that 6-AN decreases the activity of the pentose phosphate pathway to 14% of the original value (111, but the Hz02 elimination was inhibited to a significantly less extent. The results suggest that the pentose phosphate pathway supplies the majority of NADPH in fibroblasts and also suggest the presence of another NADPH source, such as isocitrate dehydrogenase (9, 32).

Removal of Intracellularly Generated
HzOzIn this study we also examined the H202 removal activity of fibroblasts by using menadione. The method is rather qualitative but allows us to estimate the action of the HzOz-removing enzymes on intracellularly generated HzO2. Menadione is reduced by intracellular enzymes to the semiquinone radical or hydroquinone (12-14). Since the semiquinone reacts very rapidly with 0 2 (331, it would form the superoxide anion (and H202) mostly before leaking out of the cell, although the possibility remains that a part of HzOZ is generated in the cell membrane, as recently reported for yeast (34). The present results conform to those obtained from the experiments with extracellularly added H20z, showing that H202 is eliminated principally by GSH peroxidase at a low HzOz generation rate. It is suggested that the intracellular HZO2 production and its metabolism can be studied by the present method.