Effect of Hb-Encapsulation with Vesicles on H2O2 Reaction and

of red blood cell substitutes are their infusion in emergency situations without ..... Artificial Blood Project) of the Ministry of Health, Labor ...
1 downloads 0 Views 204KB Size
Download PDF
1302

Bioconjugate Chem. 2002, 13, 1302−1308

Effect of Hb-Encapsulation with Vesicles on H2O2 Reaction and Lipid Peroxidation Shinji Takeoka, Yuji Teramura, Tomoyasu Atoji, and Eishun Tsuchida* Advanced Research Institute for Science and Engineering, Waseda University, Tokyo 169-8555, Japan.. Received May 1, 2002; Revised Manuscript Received September 26, 2002

We encapsulated a purified and concentrated hemoglobin (Hb) solution with a phospholipid bilayer membrane to form Hb vesicles (particle diameter, ca. 250 nm) for the development of artificial oxygen carriers. Reaction of Hb inside the vesicle with hydrogen peroxide (H2O2) is one of the important safety issues to be clarified and compared with a free Hb solution. During the reaction of the Hb solution with H2O2, metHb (FeIII) and ferrylHb (FeIV)O) are produced, and H2O2 is decomposed by the catalase-like reaction of Hb. The aggregation of discolored Hb products due to heme degradation is accompanied by the release of iron (ferric ion). On the other hand, the concentrated Hb within the Hb vesicle reacts with H2O2 that permeated through the bilayer membrane, and the same products as the Hb solution are formed inside the vesicle. However, there is no turbidity change, no particle diameter change of the Hb vesicles, and no peroxidation of lipids comprising the vesicles after the reaction with H2O2. Furthermore, no free iron is detected outside the vesicle, though ferric ion is released from the denatured Hb inside the vesicle, indicating the barrier effect of the bilayer membrane against the permeation of ferric ion. When vesicles composed of egg york lecithin (EYL) as unsaturated lipids are added to the mixture of Hb and H2O2, the lipid peroxidation is caused by ferrylHb and hydroxyl radical generated from reaction of the ferric iron with H2O2, whereas no lipid peroxidation is observed in the case of the Hb vesicle dispersion because the saturated lipid membrane of the Hb vesicle should prevent the interaction of the ferrylHb or ferric iron with the EYL.

INTRODUCTION

Enormous efforts have been made to develop red blood cell substitutes for clinical applications in recent years (1-3). Potential benefits of red blood cell substitutes are their infusion in emergency situations without concern for blood type, virus infections, and long-term storage. Hemoglobin (Hb)-based oxygen carriers (HBOCs), which have been developed so far, are generally classified into two types: one is the acellular-type modified Hb molecules such as intramolecularly cross-linked Hb (4), recombinant cross-linked Hb (5), polymerized Hb (6), and intramolecularly polymer-conjugated Hb (7). Some of these Hb modifications are currently in the final stage of clinical trials. However, clinical trials have revealed some side effects such as hypertension relating to vasoconstriction, which have been extensively reported in vivo and in vitro (8, 9). Another is the cellular-type Hb such as Hb vesicles which are being developed by our group (10) or liposome-encapsulating Hb (11). They have a vesicular structure in which concentrated Hb molecules are encapsulated, like red blood cells. Though the Hb vesicles have not yet been clinically studied, their safety and efficacy have been confirmed in comparison with Hb modifications in order to insist on the benefits of the cellular structure (12-17). From safety tests involving microvascular responses (12, 13) and heme detoxification in liver (17), we have clarified the sufficient oxygentransporting ability comparable with red blood cells in 40% hemorrhage shock (14) and 90% exchange transfusion (16). * To whom correspondence should be addressed. E-mail: [email protected].

Recently, after the administration of the Hb modifications, it has been noted that heme-mediated reactions such as ligand coordinations and redox reactions could cause organ dysfunction and/or tissue damage (18, 19). Especially, redox reactions may affect the physiological protection against reactive oxygen species (20). The oxidation of oxyHb by H2O2 is known to generate ferrylHb and metHb accompanied by heme degradation and the release of free iron (21). Furthermore, during the autoxidation of oxyHb to metHb, reactive oxygen species such as superoxide, hydrogen peroxide, and the hydroxyl radical are generated to damage not only the remaining oxyHb but also living cells and organs (22-24). Especially, ferrylHb is known to be a potent oxidant which catalyzes the peroxidation of lipids comprising the biomembrane and other biomaterials (25, 26). In normal human plasma, the concentration of H2O2 is 4-5 µM (27) and elevates to 100-600 µM under inflammatory (28) or ischemia/reperfusion conditions (29), in fact, ferrylHb can be found both in the red blood cells (30) and in the endothelial cells model after hypoxia reoxygenation (31, 32). Several in vitro studies suggest that free radicals or degradation products catalyzed by ferrylHb could damage the endothelial cells in the presence of acellular-type Hb modifications. Hb-mediated cytotoxicity via ferrylHb is one of the important safety issues of HBOCs (20, 3336). On the other hand, in the cellular-type Hb vesicles (37), reactive oxygen species generated within the Hb vesicles during metHb formation were completely consumed by Hb (unpublished data). Though such a reaction leads to Hb oxidation, no reactive oxygen species have been detected outside the vesicles.

10.1021/bc025546k CCC: $22.00 © 2002 American Chemical Society Published on Web 10/19/2002

Encapsulation Effect of Hb on Peroxidation

Here, we describe the chemical reactions between H2O2 and Hb (acellular type), or H2O2 and Hb vesicle (cellular type) and demonstrate the influence of the resulting ferrylHb on the peroxidation of unsaturated lipids, which were added to the system as egg york lecithin vesicles to clarify the advantage of Hb vesicle over acellular-type Hb concerning heme-mediated cytotoxicity via ferrylHb. MATERIALS AND METHODS

(1) Purification of an Hb Solution from Red Blood Cells (38). Hb was purified from outdated human red blood cells provided by Japanese Red Cross. The red blood cells were washed three times with saline by centrifugation (2000g, 10 min) and concentrated by the removal of the supernatant. They were hemolyzed by the addition of the equal volume of the water for injection, and stromata were removed by ultrafiltration (cutoff Mw 1000 kDa, Biomax-1000V, Millipore Co., Ltd., Bedford). The ligand exchange from O2 to CO was carried out for the stroma-free Hb solution by CO gas flowing. The proteins other than HbCO were denatured by heat treatment at 60 °C for 12 h and removed as precipitates. The HbCO solution was fractionated using ultrafiltration filters of which cutoff molecular weight were between 1000 kDa and 8 kDa (Biomax-8V, Millipore) followed by concentration with the same 8 kDa ultrafilter. (2) Preparation of an Hb Vesicle Dispersion (37). Pyridoxal 5′-phosphate (PLP, Merck, Whitehouse Station) as an allosteric effector was added to the HbCO solution (40 g/dL) at a 3/1 molar ratio of PLP to Hb. Presome PPG-I [1,2-dipalmitoyl-sn-glycero-3-phosphatidylchorine (DPPC)/cholesterol/1,2-dipalmitoyl-sn-glycero-3-phosphatidylglycerol (DPPG), Nippon Fine Chemical, Osaka] powder was mixed with the HbCO solution, and the mixture was stirred at 4 °C for 12 h. The resulting dispersion of multilamellar vesicles was extruded through membrane filters (Fuji Film Co., Tokyo) with a Remolino (Millipore). The Hb vesicles with an average diameter of 250 ( 20 nm were obtained after extrusion through the membrane filter with 0.22 µm pore size. After the separation of unencapsulated Hb by ultracentrifugation (10000g, 60 min), the precipitate of the Hb vesicles was redispersed into saline. 1,2-Distearoyl-sn-glycero-3-phosphatidylethanolamine-N-PEG (PEG molecular weight was 5000, Sunbright DSPE-50H, H-form, NOF Co., Tokyo) was added to the dispersion of the Hb vesicles until its concentration to the membrane components became 0.3 mol %. The mixture was incubated at 37 °C for 2 h for the surface modification with PEG-DSPE. Finally, unincorporated DSPE-50H was removed by ultracentrifugation (10000g, 60 min) as a supernatant. (3) Analyses of Hb Samples (Hb Solution or Hb Vesicle Dispersion). After the decarbonylation of HbCO (HbCOfHbO2) by the irradiation of visible light to the liquid film, the oxyHb solution and the oxyHb vesicle dispersion containing 1.6 mM deferoxamine mesylate (DFO, Sigma, St. Louis, MO) were used as Hb samples. Reaction of the Hb samples ([heme] ) 20 µM) with H2O2 at various ratios in phosphate buffer (pH7.4, at 37 °C) was assessed by repetitive scanning of a visible region from 450 to 700 nm at 2 min intervals by using an UVvis spectrophotometer (V-570, Jasco, Tokyo). Under a nitrogen atmosphere, deoxyHb or metHb was rapidly mixed with the H2O2 solution using a rapid scan apparatus (model TSP-601, UNISOKU, Osaka), and the repetitive scanned spectra were recorded to calculate the reaction rate. After the reaction of the Hb samples with H2O2, the bottom of the vials was carefully observed to confirm

Bioconjugate Chem., Vol. 13, No. 6, 2002 1303

whether precipitates formed. The turbidity change was measured as ∆OD at 700 nm. The particle size of the Hb vesicles was measured by a dynamic light-scattering method (Coulter particle analyzer N4SD, Coulter Co., FL). (4) Quantitative Analyses of H2O2 and Ferric Ion. The concentration of H2O2 was measured spectrofluorometrically by the formation of 6,6′-dihydroxy-[1,1′-biphenyl]-3,3′-diacetic acid (DBDA, Ex: 317 nm, Em: 405 nm) during horseradish peroxidase (HRP)-catalyzed reaction of p-hydroxyphenylacetic acid (HPA) with H2O2 (39). To the reaction mixtures of H2O2 and the Hb samples were added HRP and HPA to be 4 µM and 6mM, respectively, to determine the remaining amount of H2O2 during the reaction of H2O2 with the Hb samples. The amount of produced DBDA was calculated after separating the Hb samples by centrifugal filtration (cutoff 10 kDa, Ultrafree-MC, Millipore) and measured with a fluorescence spectrophotometer (F-4500. Hitachi, Tokyo). After the reaction of the Hb samples with H2O2 in different periods at 37 °C, the aqueous phase was separated by centrifugal filtration. Ferric ion in the filtrate was measured by an ICP emission spectroscopy (SPS-7000, Seiko Instruments Inc., Chiba). (5) Lipid Analysis of Hb Vesicle by Liquid Chromatography-Mass Spectroscopy. The Hb vesicles were lyophilized after reaction with H2O2 and dissolved in chloroform and filtered through PTFE membrane filters with the pore size of 0.22 µm. The mixed lipid components were analyzed with a liquid chromatography-mass spectrometry system (LCQ-DECA, Thermo Quest, Tokyo), equipped with an apparatus of atmospheric pressure chemical ionization (ACPI) and a gel permeation chromatographic column (TSK gel Silica 60, 7.6 mm o.d. × 300 mm h, TOSOH, Tokyo). Two micro liter of the sample solutions was injected and developed with a mixed solvent (chloroform/acetone/methanol/acetic acid/water ) 10/4/2/2/1, by vol ratio). The flow rate was 1.0 mL/min, and the temperature was maintained at 40 °C with a column oven. In the mass spectroscopic system, the gas flow rate, the applied voltage, and the CDL temperature were 0.5 mL/min, 2.5 kV, and 250 °C, respectively. (6) Lipid Peroxidation of Egg York Lecithin Vesicles. The powder of egg york lecithin (EYL) was dispersed into pure water (40 mM) and hydrated at 4 °C for 2 h under an argon atmosphere. The resulting multilamellar vesicles were extruded through the membrane filters with the final pore size of 0.22 µm to prepare the vesicles of which average diameter is 270 ( 50 nm. The Hb samples were mixed with the EYL vesicles (37 mM) and reacted with H2O2 at 37 °C for 30 min, and trichloroacetic acid (42 mM) was added to precipitate the Hb samples. The supernatant separated from the Hb samples by centrifugation (12000g, 30 min) was mixed with thiobarbituric acid (TBA, 28 mM) and heated for 30 min at 100 °C. Resulting thiobarbituric acid-reacted substance in the solution was measured with a fluorescence spectrophotometer (Ex: 532 nm, Em: 553 nm). RESULTS AND DISCUSSION

Figure 1 shows the spectral change in the Hb solution (A) and the Hb vesicle dispersion (B), both of which have a heme concentration of 20 µM, after the addition of H2O2 of three different concentrations: (a) 20 µM, (b) 200 µM, and (c) 2 mM. In Figure 1A (a), the two peaks (540 and 575 nm) in the Q-band region of oxyHb gradually decreased after the addition of 20 µM H2O2 accompanied

1304 Bioconjugate Chem., Vol. 13, No. 6, 2002

Takeoka et al.

Figure 1. Spectral changes of (A) the Hb solution and (B) the Hb vesicle dispersion ([heme] ) 20 µM) during the reaction with H2O2 at 37 °C ((a) 20 µM, (b) 200 µM, and (c) 2 mM). The repeatative sacanning was started at a 2 min interval, immediately after the addition of a solution of H2O2 to Hb sample solutions.

by the appearance of a small peak at 630 nm showing an isosbestic point at 586 nm. It is considered that oxyHb was gradually oxidized to metHb by H2O2 and stopped when H2O2 was consumed before the completion of the metHb formation. In the dispersion of Hb vesicles in which a purified Hb solution was encapsulated with a phospholipid bilayer membrane (particle diameter, 250 nm, Figure 1B (a)), the large decline in the baseline from the lower wavelength typically shows the turbidity of the vesicles. The oxyHb in the Hb vesicle was converted to metHb after reaction with 20 µM H2O2, and the degree of metHb conversion was greater than that of the oxyHb solution. When the added amount of H2O2 was increased to 200 µM, a completely different feature in the spectral change was observed as shown in Figure 1AB (b). Namely, in the Hb solution, the decrease in the two peaks attributed to oxyHb was accompanied by the total increase in the absorbance from 478 nm to over 700 nm, which can be identified as ferrylHb. The following growth of the peak at 630 nm then had an isosbestic point at 612 nm, indicating the conversion of ferrylHb to metHb. A similar trend was observed for the Hb vesicles except for the baseline change as shown in Figure 1B(b). It is noted that the ratio of metHb in both Hb samples was significantly small after the reaction with 200 µM H2O2, in comparison with the metHb spectra of the Hb samples ([heme] ) 20

µM) completely converted by the reaction with excess NaNO2, suggesting the existence of discolored Hb products (degraded heme) (21). When 2 mM of H2O2 was added to the Hb solution, the transient formation of ferrylHb with the subsequent conversion to the discolored products was confirmed from the baseline increase with no isosbestic points as shown in Figure 1A (c). We also confirmed the white aggregates at the bottom of the flask. A similar trend was also seen in the Hb vesicles but with no aggregation (Figure 1B (c)). The composition changes of oxyHb, metHb, ferrylHb, and discolored products after the addition of H2O2 are summarized in Figure 2. These curves were obtained by a curve fitting technique using computer-simulated spectra from those Hb elements measured one by one. Figure 2A (a) shows the conversion of the oxyHb solution, of which the heme concentration was 20 µM, to the other Hb products after the reaction with 20 µM H2O2. The oxyHb was reduced to 66% within 8 min accompanied by the formation of 29% metHb. The release of ferric ion and the generation of discolored products were slightly detected. The amount of H2O2 was below the detectable level (ca. 1 µM) within 7 min, thus showing the catalytic decomposition of H2O2 in this system. The direct conversion of oxyHb to metHb induced by a stoichiometric amount of H2O2 was also reported in

Encapsulation Effect of Hb on Peroxidation

Bioconjugate Chem., Vol. 13, No. 6, 2002 1305

Figure 2. Changes of Hb components and the concentrations of H2O2 and Fe3+ in (A) the Hb solution and (B) the Hb vesicle dispersion during the reaction with H2O2 at 37 °C. Hb or Hb vesicle ([heme] ) 20 µM) were reacted with H2O2 ((a) 20 µM, (b) 200 µM and (c) 2 mM, respectively). Each Hb component in the vesicle ratio was calculated from the repetative UV-vis spectral changes. b: oxyHb, O: metHb, 4: ferrylHb, 0: discolorated products. black line: Released Fe3+ measured by an ICP emission spectroscopy, grey line: H2O2 concentration determined with HRP and HPA.

some papers (40, 41) and explained in terms of “comproportionation” between ferrylHb and oxyHb to two metHbs by Giulivi and Davies (41). On the other hand, to the dispersion of oxyHb vesicles having the heme concentration of 20 µM, the same concentration of H2O2 was added as shown in Figure 2A (b). The oxyHb was reduced to 5% within 13 min accompanied by the formation of 85% metHb and 10% discolored products. No ferric ion was detected during our observation period (30 min). The main difference between these two samples is that the conversion of metHb from the oxyHb solution remained at 29%, whereas that from the oxyHb vesicle dispersion was over 85%. This can be explained as follows. Initially, oxyHb would react with H2O2 to convert metHb via comproportionation of ferrylHb with a large amount of oxyHb. The resulting metHb would show more effective catalase-like reactivity than oxyHb (21, 30). In the Hb solution, 20 µM H2O2 was completely decomposed by only 6 µM heme. In the case of the Hb vesicle dispersion, the rate-determining stage would be the reaction between oxyHb molecules that are located the vicinity of the bilayer membrane inside the vesicle and H2O2 that had just permeated through the bilayer membrane. The resulting metHb would diffuse toward the core of the vesicle and exchange with unreacted oxyHb, and then the oxyHb would react with H2O2 to convert metHb. This would result in the higher metHb conversion and ristriction of the catalase-like reaction of metHb. In fact, as shown in Figure 2 AB (a), the half-lives of the H2O2 decomposition in the Hb solution and in the Hb vesicle dispersion were 12 s and 115 s, respectively.

The reaction between 10-fold (200 µM) H2O2 and oxyHb (20 µM) in the solution state or within a vesicle as a dispersion is shown in Figure 2AB (b) as the conversion of each Hb product. The concentrations of H2O2 and ferric ion are also included in these figures. OxyHb completely disappeared within 7 min after the addition of H2O2. It was transiently converted to ferrylHb with a maximum at 4 min, followed by the conversion of metHb from the ferrylHb. Furthermore, discolored products were gradually generated from the beginning of the reaction. It is noted that the release of ferric ion was accompanied by the generation of the discolored products in the oxyHb solution. However, no ferric ion was observed in the outer aqueous phase of the oxyHb vesicle dispersion. The halflives of 200 µM H2O2 in the 20 µM oxyHb solution and the vesicle dispersion were 21 s and 200 s, respectively. When compared with those in Figure 2AB (a), only twice the time was needed to decompose the 10-fold H2O2. It is indicated that the catalase-like activity that cycles between metHb and ferrylHb (globin) radical, which is one oxidizing equivalent above ferrylHb, remained during the reaction. As described above, once metHb formed by the reaction of oxyHb with H2O2 via ferrylHb, the resulting metHb should be rapidly oxidized by H2O2 to the ferrylHb radical. The ferryl Hb radical is considered to act like the so-called compound I of catalase, which converts H2O2 to H2O and O2, and back to metHb (30). Since the latter reaction would be slow in comparison with the ferrylHb formation, the ferrylHb from oxyHb and ferrylHb radical from metHb are apparently the main products during the reaction with H2O2. These

1306 Bioconjugate Chem., Vol. 13, No. 6, 2002

ferrylHbs gradually returned to metHb after the extinction of H2O2 because of its instability. The discolored products should be degraded to heme fragments and apo-protein during the redox cylcle between oxyHb and ferrylHb, because we confirmed that metHb showed a more stable equilibrium state between metHb and the ferrylHb radical during reaction with H2O2 (data not shown). A recent report described that heme degradation was caused by superoxide generated from the reaction of ferrylHb with H2O2 (21). It is easy to understand that heme degradation triggered the release of ferric ion, and the resulting apo-proteins were so unstable that they aggregated as white precipitates. The generation of such discolored products indicates the nonperfect catalase-like activity of Hb against H2O2. When 1000-fold H2O2 was added to the Hb sample solutions, oxyHb in both cases instantly disappeared to convert ferrylHb as an intermediate, and interestingly, the final product was not metHb but discolored products. This indicates that the comproportionation of ferrylHb with oxyHb did not occur because almost all of the oxyHb was instantly converted to ferrylHb, and heme degradation should occur due to the reaction of ferrylHb with the large amount of H2O2. The decomposition of H2O2 could also be recognized even if no ferrylHb and metHb existed. The half-lives of H2O2 in the oxyHb solution and the oxyHb vesicle dispersion were 390 s and 760 s, respectively. H2O2 was hardly decomposed if H2O2 was added to the discolored Hb products in the presence of DFO, indicating that the discolored products should have no catalase activity, but H2O2 should be decomposed by a Fenton reaction caused by ferric ion released from the degraded heme. For the Hb vesicles, we did not observe the ferric ion in the outer aqueous phase; therefore, H2O2 should be decomposed by ferric ion in the inner aqueous phase of the Hb vesicle. Figure 3 shows the turbidity changes monitored at 700 nm for the six samples depicted in Figure 2 and two samples, i.e., the Hb solution and the Hb vesicle dispersion without the addition of H2O2, as references. The turbidity of the Hb solution gradually increases during the reaction with H2O2 (Figure 3a). The slope of the turbidity change increases with the increasing amount of H2O2. This clearly indicates that the turbidity increase is caused by the reaction of Hb with H2O2. We observed the white precipitates at the bottom of the vial soon after finishing the experiments, suggesting the formation of apo-protein, from which degraded heme molecules are lost. On the other hand, there were no changes in the turbidity and the diameter for all the Hb vesicle samples after the reaction with H2O2 as shown in Figure 3b. Of course, no white precipitate was observed. We could confirm the encapsulation of all the Hb products such as ferrylHb, metHb, and discolored products within the vesicle after the separation of the vesicles by ultracentrifugation. Next, the free iron (ferric ion) was determined by ICP emission spectroscopy in the reaction system between H2O2 and the Hb samples after 24 h (Figure 4). With the addition of 20 µM, 200 µM, and 2 mM H2O2 to the Hb solution, 5.1 µM, 7.2 µM, and 19.8 µM ferric ion were released, respectively. These values respectively correspond to 26, 36, and 99 mol % Hb denaturation. On the other hand, for the Hb vesicle despersion, even with the addition of 2 mM H2O2, only 2.0 µM ferric ion was detected in the outer aqueous phase of the vesicles after 24 h. However, when the Hb vesicles after the reaction with H2O2 for 24 h was solubilized by 5 mM polyoxyethylene 10 lauryl ether (C12E10), which is a good surfactant

Takeoka et al.

Figure 3. Changes in (a) the solution state of Hb and in (b) the dispersion state of Hb vesicles during the reaction with H2O2 at 37 °C. Hb or Hb vesicle ([heme] ) 20 µM) were reacted with H2O2 (20 µM (0), 200 µM (b), and 2 mM (O), respectively). (9): the Hb solution or the Hb vesicle dispersion without the addition of H2O2. The turbidities of the Hb solution and the Hb vesicle dispersion were monitored as ∆OD at 700 nm. The particle diameter was determined by a dynamic light scattering method.

Figure 4. Released Fe3+ from Hb (O), Hb vesicle (b), or Hb vesicle solubilized with lauryl ether (0) after the reaction with H2O2 and incubation at 37 °C for 24 h. All samples have the same heme concentration of 20 µM.

for the solubilization of the vesicles and does not cause Hb denaturation, the amount of ferric ion similar to that in the Hb solution system was detected. The amount of ferric ion increased corresponding to the added amount of H2O2 as shown in Figure 4. Therefore, it was confirmed that ferric ion was also released from the denatured Hb inside the vesicle, but the ferric ion was hard to permeate through the bilayer membrane because of the high resistance of the membrane to ferric ion and the high integrity of the membrane. Besides the physical integrity of the Hb vesicles, we analyzed the chemical stability of the bilayer membrane

Encapsulation Effect of Hb on Peroxidation

Figure 5. The LC-MS elution profiles of the lipid components of Hb vesicles ([heme] ) 20 µM) (a) before and (b) after the reaction with 2 mM H2O2. The peaks at around 2.2, 2.5, and 7.9 min correspond to DPPG, cholesterol, and DPPC, respectively.

Figure 6. Lipid peroxidation of the EYL vesicles catalyzed by the Hb samples during the reaction with H2O2 in the presence or absence of DFO. The 37 mM EYL vesicle was dispersed in the Hb solution or the Hb vesicle dispersion ([heme] ) 20 µM) and reacted with H2O2 at 37 °C for 30 min. O: Hb, b: Hb + DFO, 0: Hb vesicle, 9: Hb vesicle + DFO

by liquid chromatography-mass spectroscopy (LC-MS). The membrane components were extracted from the freeze-dried powder of the Hb vesicles after reaction with 2 mM H2O2. In the elution profiles, the three peaks were detected at 2.2, 2.5, and 7.9 min, corresponding to DPPG,

Bioconjugate Chem., Vol. 13, No. 6, 2002 1307

cholesterol, and DPPC, respectively, as shown in Figure 5. From the mass spectrum of each elution peak and the peak position, we could not confirm any new products and degraded fragments of the lipids after the reaction with even 100-fold H2O2 to Hb in the vesicles. Therefore, it was clarified that the saturated phospholipids and cholesterol were stable against H2O2, the active products from the reaction of H2O2 with Hb, such as ferrylHb and reactive oxygen species. In the reaction with egg york lecithin (EYL), we could not confirm that either Hb or H2O2 had peroxidized the EYL comprising vesicles. However, EYL was surely peroxidized when H2O2 was added to the EYL vesicle dispersion in the presence of 20 µM ferric ion, indicating that the hydroxyl radical produced from the Fenton reaction between ferric ion and H2O2 should cause lipid peroxidation. Such lipid peroxidation was inhibited by the addition of 1.6 mM DFO as a chelator of ferric ion. To the mixture of 37 mM EYL vesicles, 20 µM Hb, and 1.6 mM DFO were added various concentrations of H2O2, and the lipid peroxidation value was measured by a TBA method after incubation for 30 min at 37 °C. The closed circle plots in Figure 6 show evidence of the lipid peroxidation in the triadic reaction system of EYL, Hb, and H2O2. No concentration dependence of H2O2 was seen above the H2O2 concentration of more than 1.0 mM. This suggests that the rate-determining process would be the reaction of Hb with EYL because both concentrations were constant. Therefore, we looked at the relationship between the concentration of ferrylHb and the concentration of malondialdehyde. The linear relationship proves that the ferrylHb converts EYL to the peroxidized lipid (data not shown here). To study the influence of the ferric ion released from denatured Hb, the same experiment was carried out in the absence of DFO. The open circle plots in Figure 6 clearly demonstrate that the ferric ion significantly contributes to the lipid peroxidation through the Fenton reaction. The proportional relationship between the ferric ion concentration from FeCl3 artificially added to the solution instead of Hb and the concentration of the resulting malondialdehyde suggests that the reaction of H2O2 and ferric ion released from the Hb denatured by H2O2 should produce the hydroxyl radical which results in the lipid peroxidation. On the other hand, in the Hb vesicle dispersion, the small increase in malondialdehyde despite the absence of DFO is due to the encapsulation

Figure 7. Interaction of acellular-type Hb and cellular-type Hb vesicle with H2O2, and their effects on lipid peroxidation of EYL.

1308 Bioconjugate Chem., Vol. 13, No. 6, 2002

effect of Hb and ferric ion with the saturated phospholipid bilayer membrane. This suggests the low cytotoxicity of the endotherial cells in the Hb vesicles in comparison with Hb modifications under oxidative stress (33). We are now studying the effect of the Hb samples reacted with H2O2 on the endotherial cells. Figure 7 concludes our experimental results. H2O2 reacts with the acellular-type Hb solution to produce metHb and ferrylHb. The ferrylHb and ferric ion from those unstable Hb products facilitates the lipid peroxidation of the EYL vesicles, which represents the cytotoxicity of the reaction series. On the other hand, for the celullar Hb vesicle dispersion, those products are also generated by the reaction of Hb with H2O2 permeated through the bilayer membrane of the vesicle. However, they are stably encapsulated within the vesicle and do not cause the lipid peroxidation due to the separation from the EYL vesicles. These results indicate the high safety of the Hb vesicles which enclose the reactive Hb products in the reaction with H2O2. ACKNOWLEDGMENT

The authors thank Prof. Takashi Yonetani and Dr. Antonio Tsuneshige (Medical Center, Univerisity of Pennsylvania) and Dr. Hiromi Sakai and Dr. Keitaro Sou (ARISE, Waseda University) for their advice and useful discussions. This work was supported in part by Health and Labor Sciences Research Grants (Research on Pharmaceutical and Medical Safety, Artificial Blood Project) of the Ministry of Health, Labor and Welfare, Japan, and Grants in Aid for Scientific Research (B) from the Ministry of Education, Science, Sports, and Culture, Japan (12558112). One of the authors (Y.T.) thanks Research Fellowships of the Japan Society for the Promotion of Science for Young Scientists grant. LITERATURE CITED (1) Tsuchida, E. (1995) Artificial Red Cells, John Wiley and Sons, New York. (2) Winslow, R. M. (1995) Nature Med. 1, 1212-1215. (3) Chang, T. M. S. (1991) Biomater. Artif. Cell Immobil. Biotechnol. 20, 159-182. (4) Przybelski, R. J., Daily, E. K., and Birnbaum, M. L. (1997) Advances in Blood Sbstitutes: Industrial Opportunities and Medical Challenges, Birkhauser, Boston. (5) Murray, J. A., Ledlow, A., Launspach, J., Evans, D., Loveday, M., and Conklin, J. L. (1995) Gastroenterology 109, 1241-1248. (6) Gould, S. A., Moore, E. E., Hoyt, D. B., Burch, J. M., Haenel, J. B., Garcia, J., DeWoskin, R., and Moss, G. S. (1998) J. Am. Coll. Surg. 187, 113-122. (7) Caron, A., Menu, P., Labrude, P., and Vigneron, C. (1998) Artif. Cells, Blood Substitutes, Immobilization Biotechnol. 26, 293-308. (8) Doherty, D. H., Doyle, M. P., Curry, S. R., Vali, R. J., Fattor, T. J., Olsen, J. S., and Lemon, D. D. (1998) Nat. Biotechnol. 1, 672-676. (9) Saxena, R., Wijnhoud, A. D., Carton, H., Hacke, W., Kaste, M., Przybelski, R. J., Stern, K. N., and Koudstaal, P. J. (1999) Stroke 30, 993-996. (10) Tsuchida, E., Ed. (1998) Blood substitutes: present and future perspectives, Elsevier Science, Amsterdam.

Takeoka et al. (11) Rudolph, A. S., Sulpizio, A., Hieble, P., MacDonald, V., Chavez, M., and Feuerstein, G. (1997) J. Appl. Physiol. 82, 1826-1835. (12) Sakai, H., Yuasa, M., Onuma, H., Takeoka, S., and Tsuchida, E. (2000) Bioconjugate Chem. 11, 56-64. (13) Sakai, H., Hara, H., Yuasa, M., Tsai, A. G., Takeoka, S., Tsuchida, E., and Intaglietta, M. (2000) Am. J. Physiol. Heart Circ. Physiol. 279, H908-H915. (14) Izumi, Y., Sakai, H., Hamada, K., Takeoka, S., Yamahata, T., Kato, R., Nishide, H., Tsuchida, E., and Kobayashi, K. (1996) Circ. Care Med. 24, 1869-1873. (15) Sakai, H., Tomiyama, K., Sou, K., Takeoka, S., and Tsuchida, E. (2000) Bioconjugate Chem. 11, 425-432. (16) Sakai, H., Takeoka, S., Park, S.-I., Kose, T., Izumi, Y., Yoshizu, A., Nishide, H., Kobayashi, K., and Tsuchida, E. (1997) Bioconjugate Chem. 8, 15-22. (17) Kyokane T., Norizumi, S., Taniai, H., Yamaguchi, T., Takeoka, S., Tsuchida, E., Naito, M., Ishimura, Y., and Suematsu, M. (2001) Gastroenterology 120, 1227-1240. (18) Alayash, A. I., and Cashon, R. E. (1995) Mol. Med. Today 1 122-127. (19) Sharma, V., Traylor, T., and Gardiner R. (1987) Biochemistry 26, 3837-3843. (20) Alayash, A. I. (1999) Nat. Biotechnol. 17, 545-549. (21) Giulivi, C., and Cadenas, E. (1998) Free Radic. Biol. Med. 24, 269-279. (22) Gutteridge, J. M. C. (1986) FEBS Lett. 201, 291-295.23. Macdonald, V. W. (1994) Methods Enzymol. 231, 480-490. (23) , Everse, J., Johnson, M. C., and Marini, M. A. (1994) Methods Enzymol. 231, 547-561. (24) Grisham, M. B., and Everse, J. (1991) Peroxidases in Chemistry and Biology, pp 335-344, CRC Press, Boca Raton, FL. (25) Yamamoto, Y., Brodsky, M. H., Baker, J. C., and Ames, B. N. (1987) Anal. Biochem. 160, 7-13. (26) Grisham, M. B., Gaginell, T. S., Von Ritter, C., Tamai, H., Be, R. M., and Granger, D. N. (1990) Inflammation 14, 531542. (27) Nagababu, E., and Rifkind, J. M. (2000) Biochemistry 39, 12503-12511. (28) Gunther, M. R., Sampath, V., and Caughey, W. S. (1999) Free Radic. Biol. Med. 26, 1388-1395. (29) Svistunenko, D. A., Patel, R. P., Voloshchenko, S. V., and Wilson, M. T. (1997) J. Biol. Chem. 272, 7114-7121. (30) Clark, M. R. (1988) Physiol. Rev. 68, 503-554. (31) Mcleod, L. L., and Alayash, A. I. (1999) Am. J. Physiol. Heart Circ. Physiol. 277, H92-H99. (32) Goldman, D. W., Breyer, R. J., III, Yeh, D., Brockner-Ryan, B. A., and Alayash, A. I. (1998) Am. J. Physiol. Heart Circ. Physiol. 275, H1046-H1053. (33) D’Agnillo, F., and Alayash, A. I. (2000) Am. J. Physiol. Heart Circ. Physiol. 279, H1880-H1889. (34) Darley-Usmar, V., and White, R. (1997) Exp. Physiol. 82, 305-316. (35) Natarajan, V. (1995) J. Lab. Clin. Med. 123, 461-463. (36) Takeoka, S., Ohgushi, T., Terase, K., Ohmori, T., and Tsuchida, E. (1996) Langmuir 12, 1755-1759. (37) Sakai, H., Hamada, K., Takeoka, S., Nishide, H., and Tsuchida, E. (1996) Biotechnol. Prog. 12, 119-125. (38) Hinsberg, W. D., III, Milby, K. H., and Zare, R. N. (1981) Anal. Chem. 53, 1509-1512. (39) Yusa, K., and Shikama, K. (1987) Biochemistry 26, 66846688. (40) Giulivi, C., and Davies, K. J. (1990) J. Biol. Chem. 265, 19453-19460.

BC025546K