Characterization of Acatalasemic Erythrocytes Treated with Low and High Dose Hydrogen Peroxide

The effects of hydrogen peroxide on normal and acatalasemic erythrocytes were examined. Severe hemolysis of acatalasemic erythrocytes and a small tyrosine radical signal (g = 2.005) associated with the formation of ferryl hemoglobin were observed upon the addition of less than 0.25 mm hydrogen peroxide. However, when the concentration of hydrogen peroxide was increased to 0.5 mm, acatalasemic erythrocytes became insoluble in water and increased the tyrosine radical signal. Polymerization of hemoglobin and aggregation of the erythrocytes were observed. On the other hand, normal erythrocytes exhibited only mild hemolysis by the addition of hydrogen peroxide under similar conditions. From these results, the scavenging of hydrogen peroxide by hemoglobin generates the ferryl hemoglobin species (H-Hb-Fe(IV)=O) plus protein-based radicals (·Hb-Fe(IV)=O). These species induce hemolysis of erythrocytes, polymerization of hemoglobin, and aggregation of the acatalasemic erythrocytes. A mechanism for the onset of Takarara disease is proposed.

In 1952, Takahara reported that Japanese acatalasemia (catalase deficiency) patients suffered from progressive oral gangrene (Takahara disease) as a result of being infected with hydrogen peroxide-generating bacteria (1). Subsequently, Swiss and Hungarian acatalasemia-afflicted people were reported, but they did not evidently suffer from the symptomatic features of the disease (2,3). The residual catalase (EC 1.11.1.6) activity in Swiss and Hungarian acatalasemic erythrocytes was found to be higher than in the erythrocytes from the Japanese. Ogata et al. (4) suggested the catalase deficiency in the blood to be the etiological cause of the disease. We became interested in the scavenging of hydrogen peroxide in erythro-cytes to understand the metabolism of hydrogen peroxide in acatalasemic erythrocytes. The hydrogen peroxide scavenging rates in normal and acatalasemic mouse and human erythrocytes were examined (5)(6)(7)(8). It was found that hydrogen peroxide was dominantly scavenged by hemoglobin in hemolysates of acatalasemic erythrocytes. The scavenging rate was enhanced in the presence of reduced pyridine nucleotides (NAD(P)H) or ascorbic acid (AsA). 2 A cyclic process has been proposed for the scavenging mechanism; hydrogen peroxide oxidizes hemoglobin, which is then reduced by NAD(P)H or AsA (8).
We suspected that during the process, hemoglobin was oxidized to ferryl hemoglobin (H-Hb-Fe(IV)ϭO) with hydrogen peroxide, and then a comproportionation reaction of Hb-Fe(IV)ϭO and ferrous hemoglobin generated methemoglobin (H-Hb-Fe(III) ϩ ) (9). Methemoglobin was then further oxidized to radicals of ferryl hemoglobin ( ⅐ Hb-Fe(IV)ϭO) with hydrogen peroxide (10). In this study, we report that the oxidized hemoglobin species in the scavenging process were identified and that a novel property of acatalasemic erythrocytes was subjected to hydrogen peroxide treatment; erythrocytes aggregated to be insoluble in water at a high concentration (Ն0.5 mM) of hydrogen peroxide, whereas the severe hemolysis of acatalasemic erythrocytes resulted at low concentrations of hydrogen peroxide (Ͻ0.25 mM).

EXPERIMENTAL PROCEDURES
Materials-Male mice (8 weeks old) of the C3H/AnLCs a Cs a (normal) and C3H/AnLCs b Cs b (acatalasemia) strains established by Feinstein et al. (11) were maintained on a laboratory diet (MF diet, Oriental Yeast Co. Ltd., Tokyo, Japan) and water ad libitium until experiments. Chemicals of analytical grade were purchased from Sigma and Wako Pure Chemical Ind. (Osaka, Japan).
General Procedures-Hemoglobin contents were determined by the method of Drabkin and Austin (12), and the concentrations of hemoglobin indicated were calculated as tet-ramer. Catalase activity was measured according to a previous method (5). SDS-PAGE was performed according to the method of Laemmli (13). The concentrations of the materials indicated were those of the final ones used in samples.
Preparation of Erythrocyte Suspension and Hemolysates-Mouse blood was collected in heparinized tubes. Erythrocytes were separated by centrifugation and washed three times with 10 mM sodium phosphate buffer containing 152 mM sodium chloride (PBS, pH 7.4). After the last washing, packed cells (100% of erythrocytes) were diluted with PBS, and the suspension was used for experiments. Hemolysates were prepared by dilution of the packed erythrocytes with 9 volumes of water.
Treatment of Normal and Acatalasemic Hemolysates with Hydrogen Peroxide-Hemolysates (0.1%) in PBS were treated with 0.1 mM hydrogen peroxide, and the spectra (380 -680 nm) were recorded.
Treatment of Normal and Acatalasemic Erythrocytes with Hydrogen Peroxide-A 2% erythrocyte suspension was treated with 0.0 -5.0 mM hydrogen peroxide at 37°C. For the measurement of the protein leakage from the erythrocytes treated with hydrogen peroxide, 0.06 ml of the mixture was immediately diluted with 0.36 ml of PBS at 5 and 30 min after the addition of hydrogen peroxide. Each mixture was centrifuged at 1,100 ϫ g for 5 min, and the absorbance of the supernatant was measured at 280 nm. For the determination of the water-soluble protein contents in the erythrocytes treated with hydrogen peroxide, 0.06 ml of the incubation mixture was put into a tube containing 0.36 ml of water. The absorbance of the supernatant was measured at 280 nm. For measurement of ESR spectra, a 2% erythrocyte suspension was charged in an ESR tube (inner diameter 5 mm) and incubated with hydrogen peroxide at 37°C for 5 min. The mixture was immediately frozen in liquid nitrogen. ESR spectra were measured at 120 K using a JES-FE1XG (JOEL,  Tokyo, Japan). Manganese dioxide was used as the internal standard.
Hemolysis of Erythrocytes Treated with a Low Concentration (0.1 mM) of Hydrogen Peroxide-A 2% erythrocyte suspension was treated at 37°C with 0.1 mM hydrogen peroxide. Two samples (0.06 ml ϫ 2) were taken from the mixture. One sample was diluted with 0.36 ml of PBS, and the other was diluted with 0.36 ml of water. Both mixtures were centrifuged at 1,100 ϫ g for 5 min, and the absorbance of each supernatant at 540 nm was measured. Hemolysis of erythrocytes in the reaction mixture was calculated according to the method of Miki et al. (14). To examine the effect of glucose or glucose-AsA on hemolysis, an acatalasemic erythrocyte suspension containing 5 mM glucose or 5 mM glucose and 0.2 mM AsA was preincubated at 37°C for 30 min. The hemolysis of erythrocytes by hydrogen peroxide was determined as described above. Hemolysis of erythrocytes with 2,2Ј-azobis-(2-amidinopropane) dihydrochloride (AAPH) was examined by replacing hydrogen peroxide with 50 mM AAPH.

SDS-PAGE Analysis of Erythrocytes Treated with a High Concentration (Ն1 mM) of Hydrogen
Peroxide-A 2% erythrocyte suspension (0.25 ml) was incubated with 1-10 mM hydrogen peroxide at 37°C for 5 min. PBS was then immediately added to the mixture to dilute the hydrogen peroxide. After the supernatant was removed, 0.4 ml of sample buffer containing 2-mercaptoethanol (13) was added to the residue, and a sample for SDS-PAGE analysis was prepared. For Western blot analysis, the gel was blotted onto a nitrocellulose membrane. Western blot was carried out with a rabbit polyclonal antibody for mouse hemoglobin (1:2,500, ICN Pharmaceuticals Inc.) and the ECL Western blotting analysis system (Amersham Biosciences, Buckinghamshire, UK).
Microscopic Studies of Erythrocytes-Two percent erythrocyte suspensions were mixed with hydrogen peroxide (at a final concentration of 1 mM). Erythrocytes were placed on glass slides and observed with an Olympus light microscope (Olympus, Tokyo, Japan) fitted with a high resolution digital camera system (Penguin 600CL; Pixera Co., Los Gatos, CA).  . ESR analyses of erythrocytes treated with hydrogen peroxide. A, overall ESR spectrum. A 2% acatalasemic erythrocyte suspension was treated with 1 mM hydrogen peroxide, and the spectrum was recorded at 120 K. The inset indicates the spectrum around g ϭ 2. B, relation between the intensity of the signal (g ϭ 2.005) and the concentration of hydrogen peroxide. Hydrogen peroxide was added to the erythrocyte suspensions, and the spectrum was measured at 120 K. Signal intensity (g ϭ 2.005) is expressed as the ratio of the ferryl radical signal to manganese oxide. nm) suggests oxidation of ferrous hemoglobin to a mixture of methemoglobin and ferryl hemoglobin species. The formation of methemoglobin is indicated by the shoulder band at 630 nm and the increase of absorbance at 500 nm (15,16). The formation of ferryl hemoglobin species is suggested by the weak bands at 538 and 574 nm and confirmed by the spectra of the sulfoderivative (Fig. 1, spectrum C, max 618 nm) with the addition of 1 mM sodium sulfide (9,16). However, the spectra of normal hemolysates did not change with the addition of hydrogen peroxide (data not shown).

Catalase Activity in Mouse Erythrocytes
Treatment of Erythrocytes with Various Concentrations of Hydrogen Peroxide-Two percent of the erythrocyte suspension contained 75.3 Ϯ 10.8 M hemoglobin. The protein leakage from erythrocytes treated with hydrogen peroxide is shown in Fig. 2A. The protein leakage was not observed after 5 min but was observed 30 min after the addition of 0.1 mM hydrogen peroxide. The water-soluble protein contents in the erythrocytes treated with hydrogen peroxide are indicated in Fig. 2B. The figure indicates that the acatalasemic erythrocytes treated with more than 0.5 mM hydrogen peroxide became insoluble in water 5 min after the addition. Precipitated cells subsequently never became soluble in water again.
Radical ESR signals (g ϭ 2.005 and 2.03) were detected when the acatalasemic erythrocyte suspension was treated with 1 mM hydrogen peroxide (concentration of hemoglobin was 69.8 M, Fig. 3A). These signals were assigned as tyrosine (g ϭ 2.005) and tryptophan peroxyl (g ϭ 2.03) radicals, respectively (10,17). A small signal of g ϭ 6 was also observable and was assigned as methemoglobin (high spin). The relation between the intensity of the tyrosine radical signal and the concentration of hydrogen peroxide is indicated in Fig. 3B. The small signal was observed in the acatalasemic erythrocytes, but the signal did not rapidly increase at levels up to 0.25 mM hydrogen peroxide. The signal intensity increased and reached a peak at 2.0 mM hydrogen peroxide.   Fig.  4A. After 30 min had elapsed from the addition of hydrogen peroxide, most (Ն60%) of the acatalasemic erythrocytes had become hemolyzed, but normal erythrocytes had not. The hemolysis of acatalasemic erythrocytes by hydrogen peroxide was prevented by the addition of glucose and more effec-tively by the addition of glucose-AsA (Fig. 4C). When compared with the hemolysis caused by the addition of hydrogen peroxide, normal and acatalasemic erythrocytes were treated with AAPH (Fig. 4B). There was no observable difference in hemolysis between them.
Color Change of the Erythrocyte Suspension by Addition of Hydrogen Peroxide-When erythrocytes were treated with 1 mM hydrogen peroxide, the color of the acatalasemic erythrocytes dramatically changed from red to dark brown (1). The color change was attributable to the oxidation of hemoglobin and precipitation of the erythrocytes by hydrogen peroxide.
SDS-PAGE Analysis of Erythrocytes Treated with 1-10 mM Hydrogen Peroxide-The bands for hemoglobin (16 kDa) in the acatalasemic erythrocytes decreased with the treatment of hydrogen peroxide, and the protein bands of greater than 207 kDa increased (Fig. 5). Western blot indicated the latter bands to be polymerized adducts of hemoglobin. On the other hand, the SDS-PAGE pattern of normal erythrocytes was not significantly affected by 1-10 mM hydrogen peroxide treatment.
Microscopic Studies of Erythrocytes-Light microscopy revealed the acatalasemic erythrocytes to be deformed and significantly aggregated (Fig. 6B, arrows), whereas normal erythrocytes exhibited no discernable aggregation after treatment with 1 mM hydrogen peroxide (Fig. 6A).

DISCUSSION
First, the visible spectral changes in the hemolysates were examined subsequent to the addition of hydrogen peroxide. Although the spectra of the normal hemolysates did not change after the addition of hydrogen peroxide, the spectra of the acatalasemic hemolysates did (Fig. 1). The spectral changes indicated the formation of methemoglobin (H-Hb-Fe(III)) and the ferryl hemoglobin species (H-Hb-Fe(IV)ϭO and ⅐ Hb-Fe(IV)ϭO) (9).
Subsequently, we examined the effects of hydrogen peroxide on erythrocytes. Although the protein leakage from normal erythrocytes was hardly observed 30 min after the addition of 0.1 mM hydrogen peroxide, the protein leakage from the acatalasemic erythrocytes was severe. Surprisingly, with the addition of a higher concentration of hydrogen peroxide, protein leakage from the acatalasemic erythrocytes was reduced ( Fig. 2A), and the erythrocytes immediately became insoluble in water in the presence of more than 0.5 mM hydrogen peroxide (Fig. 2B). On the other hand, normal erythrocytes treated with hydrogen peroxide exhibited only mild hemolysis. These results indicate that acatalasemic erythrocytes are easily distinguishable from normal erythrocytes after the addition of hydrogen peroxide.
A close examination of the time course of hemolysis at a dose level of 0.1 mM hydrogen peroxide would provide a close characterization of hemolysis. It was expected there would be a big difference between normal and acatalasemic erythrocytes (Fig. 4A). However, hemolysis by AAPH, which decomposes with heat to form a carbon-centered radical, indicated that there was in fact no difference in hemolysis between them (Fig. 4B). These results suggested that hydrogen peroxide penetrated through the membrane and generated ferryl hemoglobin species in the acatalasemic but not normal erythrocytes because hydrogen peroxide was decomposed by catalase (8). On the other hand, AAPH cannot penetrate the membrane, and the radical from AAPH equally breaks down the cell membrane of both types of erythrocytes from the outside.
The scavenging reaction of hydrogen peroxide in erythrocytes was traced by ESR measurements. A signal of the tyrosine radical was always observable in the acatalasemic erythrocytes. The signal was strongly observed when treated with hydrogen peroxide (1 mM) (Fig. 3A). The relation of the signal intensity to the concentration of hydrogen peroxide indicated that a low concentration of hydrogen peroxide (Ͻ0.5 mM) generated only a low concentration of the radical in the acatalasemic erythrocytes. We speculated that the low con-centration of hydrogen peroxide oxidized hemoglobin to generate ferryl hemoglobin, and then a reaction of the ferryl hemoglobin with ferrous hemoglobin produced methemoglobin (9), which concomitantly reacts with hydrogen peroxide to generate the radicals of ferryl hemoglobin (10). This account explains the previous observation that the concentration of methemoglobin in mouse acatalasemic erythrocytes was higher than that found in normal erythrocytes (15). Furthermore, the formation of ferryl hemoglobin and the radicals in acatalasemic erythrocytes induces hemolysis. This is because ferryl hemoglobin species are strong oxidants and can promote lipid peroxidation of the membrane (9,16). Such hemolysis may explain the clinical finding that heme degradation products in the urine of acatalasemia patients are five times higher than in normal patients (18).
The prevention of acatalasemic erythrocyte hemolysis by the addition of glucose-AsA and/or glucose (Fig. 4C) was observed. As the scavenging reaction of hydrogen peroxide in the presence of hemoglobin was enhanced by NAD(P)H (8), the preventive effect by glucose was explained by the reduction with NAD(P)H since NADPH generation through the pentose phosphate pathway was enhanced by the generation of hydrogen peroxide and low catalase activity in erythrocytes (19). As the prevention effect of the hemolysis with glucose-AsA is consistent with that of the myocardinal cell damage caused by the ferryl myoglobin species (20,21), the effect is attributable to the direct reduction of oxidized hemoglobin by AsA (8).
To characterize the water-insoluble erythrocytes that resulted from the treatment with a high concentration of hydrogen peroxide (more than 1 mM), insoluble erythrocytes were examined by SDS-PAGE and Western blot analyses. The results indicate that the amounts of hemoglobin decreased and that hemoglobin polymerized to form adducts in the acatalasemic mouse erythrocytes. The ESR spectra revealed the tyrosine radical of ferryl hemoglobin to be generated and increased at more than 0.5 mM hydrogen peroxide. As it was expected that the amounts of the radical would be increased with more than 0.3 mM hydrogen peroxide, a high concentration of the radical would make hemoglobin polymerize to become adducts, and the erythrocytes thus would become insoluble in water. The abnormal reaction with hydrogen peroxide caused the shape change of the erythrocytes, which resembled that of cardiac myocytes treated with lipid hydroperoxide (22). Furthermore, aggregation was observed in the system (Fig. 6). This aggregation may be due to a change in the charge of the erythrocyte membrane induced by the high concentration of the reactive tyrosine radical. The aggregated erythrocytes might impede the normal flow of the blood stream and thus contribute to the oral gangrene observed in Japanese acatalasemia patients (Takahara disease) (1). Thus, we propose the catabolism of hydrogen peroxide in acatalasemic erythrocytes as depicted in Fig. 7.