Reactions of 1-Methyl-2-phenylindole with Malondialdehyde and 4

Sep 30, 1998 - Role of the Electrophilic Lipid Peroxidation Product 4-Hydroxynonenal in the Development and Maintenance of Obesity in Mice ...... Toxi...
0 downloads 11 Views 129KB Size
Download PDF
1176

Chem. Res. Toxicol. 1998, 11, 1176-1183

Reactions of 1-Methyl-2-phenylindole with Malondialdehyde and 4-Hydroxyalkenals. Analytical Applications to a Colorimetric Assay of Lipid Peroxidation Dominique Ge´rard-Monnier,*,† Irene Erdelmeier, Kheira Re´gnard, Nathalie Moze-Henry, Jean-Claude Yadan, and Jean Chaudie`re‡ Centre de Recherche Oxis International S.A., 94385 Bonneuil sur Marne, France Received October 6, 1997

Under acidic and mild-temperature conditions, 1-methyl-2-phenylindole was found to react with malondialdehyde (MDA) and 4-hydroxyalkenals to yield a stable chromophore with intense maximal absorbance at 586 nm. The use of methanesulfonic acid results in optimal yields of chromophore produced from MDA as well as from 4-hydroxynonenal. By contrast, the use of hydrochloric acid results in an optimal yield of chromophore produced from MDA and a negligible reaction of 4-hydroxynonenal. Taking advantage of such chromogenic reactions, we developed a new colorimetric assay of lipid peroxidation. Using a methanesulfonic acidbased medium, MDA and 4-hydroxyalkenals can be measured at the 586 nm wavelength. However, the presence of endogenous inhibitors of the reaction with 4-hydroxyalkenals is common, and this means that the latter may be underestimated in some biological samples. The assay performed in a hydrochloric acid-based medium enables the specific measurement of MDA in the presence of 4-hydroxyalkenals. Upon hydrolysis of Schiff bases in hydrochloric acid (pH 1.5), either assay can be used to specifically measure the amount of total MDA in biological samples because 4-hydroxyalkenals undergo an irreversible cyclization reaction under the hydrochloric acid-based conditions of hydrolysis. The two assays were applied to the determination of the amount of MDA alone and of MDA and 4-hydroxyalkenals in an in vitro model of lipid peroxidation. This methodology was also used to clarify complex patterns of tissue-specific MDA production in vivo, following hydrolysis of Schiff bases, in rodents treated with doxorubicin.

Introduction MDA1

and 4-hydroxyalkenals are important toxic byproducts of lipid peroxidation. The measurement of the amounts of such aldehydes has been widely used as an index of lipid peroxidation in vitro and in vivo (1, 2). When heated at acidic pH, MDA reacts with a number of nucleophiles to yield a variety of condensation/ dehydration products (3). 2-Thiobarbituric acid (TBA) has been widely used as a reagent for the colorimetric measurement of MDA amounts, due to the stability and to the high molar extinction coefficient of the resulting adduct at 532 nm. The TBA method is fast and easy to use, but it suffers from several limitations and drawbacks (3-5). Much interference is due to side reactions of TBA and endogenous compounds at high temperatures and low pH. More reliable assays of specific products of lipid peroxidation are usually based on chromatographic separations and are typically time-consuming. In other * To whom correspondence should be addressed. † Present address: INSERM, U 479, Faculte ´ Xavier Bichat, 16 rue Henri Huchard, 75018 Paris, France. Phone: (33) 1 44 85 62 11. Fax: (33) 1 44 85 62 07. ‡ Present address: UFR de Biologie, University of Paris VII, 2 place Jussieu, 75251 Paris Cedex 05, France. 1 Abbreviations: MDA, malondialdehyde; TBA, 2-thiobarbituric acid; TMP, 1,1,3,3-tetramethoxypropane; 4-HNE, 4-hydroxy-2(E)-nonenal; BSA, bovine serum albumin; 13(S)-HPOTE, 9,11-octadecatrienoic acid, 13-hydroperoxy; 15(S)-HPETE, 5,8,11,13-eicosatetraenoic acid, 15hydroperoxy.

spectrophotometric methods for the determination of MDA amounts, 1-methylpyrrole and indole are the only chromogenic reagents which do not require prolonged heating (6). Reactions between indoles and aldehydes are ordinarily initiated by acid-catalyzed attack at the 3-position of the indole ring, as illustrated by the reaction of two molecules of the indole reagent with one molecule of saturated aldehyde, to give a diindolylalkane. A trimethine dye was postulated as the structure of the chromophores obtained from the reactions of MDA with 1-methylpyrrole and indole (6). The products of reaction of MDA with either 1-methylpyrrole (6), indole (6), or 2-methylindole (7) are not stable and do not exhibit strong visible absorbance (5). However, we have found that 2-aryl-1-methylindoles react with MDA to yield a stable chromophore whose maximal absorbance is in the 580-620 nm region (5). 1-Methyl-2-phenylindole was selected as the optimal reagent, because of its fast reaction with MDA, even in the presence of water,2 yielding a chromophore with a high molar extinction coefficient at its maximal absorption wavelength of 586 nm. Here we show that this reaction is MDA-specific in the presence of hydrochloric acid, whereas the same chromophoric product is obtained from both MDA and 4-hydroxyalkenals in the presence of methanesulfonic 2 D. Ge ´ rard-Monnier, I. Erdelmeier, K. Re´gnard, J.-C. Yadan, and J. Chaudie`re, unpublished results.

10.1021/tx9701790 CCC: $15.00 © 1998 American Chemical Society Published on Web 09/30/1998

Colorimetric Assay of Lipid Peroxidation

acid. Accordingly, we describe the use of 1-methyl-2phenylindole for the colorimetric measurement of amounts of MDA and 4-hydroxyalkenal and for the specific determination of amounts of MDA alone. A detailed mechanistic study of the reactions on which our assays are based can be found in the following paper (46).

Experimental Section Caution: 4-HNE is cytotoxic and mutagenic. Alkanals and alkenals are irritants; hydrochloric and methanesulfonic acid are strong corrosive liquids. All these chemicals should be handled with care to minimize exposure. Chemicals. 1,1,3,3-Tetramethoxypropane (TMP) was used as a source of MDA. When required, free MDA was obtained upon hydrolysis of TMP in 1% sulfuric acid (1). The diethylacetal derivatives of 4-hydroxy-2(E)-nonenal (4-HNE) and 4-hydroxy-2(E)-hexenal were synthesized as described by Esterbauer and Weger (8). Free 4-HNE was either synthesized according to Gardner et al. (9), and the concentration checked in water using a molar extinction coefficient of 13 750 M-1 cm-1 at 223 nm (10), or prepared extemporaneously by hydrolysis of the diethylacetal. A 100 µM solution of the free aldehyde was obtained immediately by diluting a stock solution of the diethylacetal (10 mM in acetonitrile) in a mixture of acetonitrile/0.1 N methanesulfonic acid (2:1, v/v). Concentrated methanesulfonic acid (>99% purity) and 37% hydrochloric acid (99.999% purity) were purchased from SAF (St. Quentin-Fallavier, France). All the reagents were of the highest grade commercially available. An iodometric cholesterol color reagent (Merck, Nogent sur Marne, France) was used to measure the concentration of fatty acid hydroperoxides in ethanolic solutions (11). 1-Methyl-2phenylindole was synthesized as described previously (12). The seleno-organic compounds 2-phenyl-1,2-benzisoselenazol-3(2H)one (Ebselen) and 3,4-dihydro-4,4-dimethyl-2H-1,2-benzoselenazine (BXT 51072) were synthesized as previously described (13, 14). Materials. Disposable test tubes made of borosilicated glass were used for the reactions. Prior to use, other glassware was washed exhaustively with acetic acid and rinsed with deionized water. Biological Samples. Sprague-Dawley male rats, Balb C female mice, and hepatic microsomes from phenobarbitaltreated rats were obtained from Iffa-Credo (Les Oncins, France). Plasma samples from rats or human volunteers were used. Venous blood was obtained from healthy volunteers after obtaining informed consent. Three milliliters of blood was collected in a test tube containing 48 µL of 0.17 M potassium EDTA and then the mixture centrifuged at 2000g and 4 °C for 10 min. Colorimetric Assays of Lipid Peroxidation by Means of 1-Methyl-2-phenylindole. Two procedures have been used. In the hydrochloric acid-based assay (procedure 1), which is MDA-specific, 200 µL of the aqueous sample was added to 650 µL of a solution of 1-methyl-2-phenylindole in a mixture of acetonitrile/methanol (3:1). The final concentration of the reagent was 10 mM. The reaction was then started by adding 150 µL of 37% hydrochloric acid. The 586 nm absorbance (A) was measured upon incubation of the reaction mixture at 45 °C for 40 min. For each series of assays, a blank (aldehyde replaced by water) was included, and its postreaction absorbance (A0) was further subtracted from the A value. The level of this background (A0) was 0.003 ( 0.002 (mean ( SE of 10 measurements). For each assay in plasma or homogenate, a sample blank in which the reagent was replaced by acetonitrile/ methanol (3:1, v/v) was included. In this HCl-based medium, the final 586 nm absorbance was a linear function of MDA concentration, and an apparent molar extinction coefficient of 110 000 M-1 cm-1 was obtained. The 586 nm chromophore was stable for at least 2 h. In the methanesulfonic acid-based procedure 2, hydrochloric acid was replaced by the same volume of concentrated methanesulfonic acid containing 34 µM Fe(III).

Chem. Res. Toxicol., Vol. 11, No. 10, 1998 1177 The 586 nm absorbance (A) was measured upon incubation of the reaction mixture at 45 °C for 30 min. The blank value (aldehyde replaced by water) (A0) was 0.007 ( 0.002 (mean ( SE of 10 measurements). The final 586 nm absorbance was a linear function of the concentrations of both MDA and 4-hydroxyalkenals. An apparent molar extinction coefficient of 110 000 M-1 cm-1 was obtained with MDA as well as with the diethylacetals of either 4-HNE or 4-hydroxyhexenal (4-HHE). On the basis of 10 measurements of the reagent blank (15), the detection limit of the assays was less than 0.1 µM in the final reaction mixture. The within-series precision was determined by performing 20 measurements with the same batches of reagents (15), using concentrations of aldehydes ranging from 1 to 5 µM. The reproducibility of the methods was determined by performing two series of assays every day over 10 days under the same experimental conditions (15), using concentrations of aldehyde standards ranging from 1 to 5 µM and obtained from 10 mM stock solutions of the corresponding acetals stored at 4 °C. Hydrolysis of Schiff Bases from MDA and Proteins. Tissues or cells were homogenized at 4 °C in 20 mM Tris buffer (pH 7.4). The homogenates were centrifuged for 10 min at 2000g and 4 °C. The supernatants were adjusted to pH 1.5 with diluted HCl, and then 200 µL aliquots were hydrolyzed for 80 min at 60 °C. Diluted 1-methyl-2-phenylindole (650 µL) in acetonitrile/methanol (3:1) or acetonitrile/methanol alone (sample blanks) was directly added to the hydrolyzed samples. The reaction was started by the addition of 150 µL of 37% hydrochloric acid or methanesulfonic acid. The 586 nm absorbance of the reaction mixture was read upon incubation at 45 °C and subsequent centrifugation at 9000g for 10 min. Plasma samples were adjusted to pH 1.5 with diluted hydrochloric acid, and then 200 µL aliquots were hydrolyzed for 80 min at 60 °C and subsequently used for the assays, as described above. Alternatively, tissue homogenization or cell lysis can be directly performed in diluted hydrochloric acid, at pH 1.5. The resulting sample was hydrolyzed for 80 min at 60 °C, and then centrifuged (low-speed, 10 min, 4 °C). The assay was performed as described above, using procedure 1 or 2, on 200 µL aliquots of the supernatant. 1H NMR studies have shown that 4-hydroxyalkenals undergo a fast and irreversible cyclization reaction to 2-pentylfuran in the presence of hydrochloric acid, if the chromogenic reagent is absent (46). Hence, upon hydrolysis in diluted hydrochloric acid, 4-hydroxyalkenals were converted into furan compounds that were unable to react with 1-methyl-2-phenylindole. Therefore, the hydrochloric acid-based assay (procedure 1) as well as the methanesulfonic acid-based assay (procedure 2) allows the specific measurement of the amount of total MDA upon prior hydrolysis of Schiff bases in hydrochloric acid. Preparation of Schiff Bases from MDA and Bovine Serum Albumin. Schiff bases were prepared by incubating bovine serum albumin (BSA) with MDA at pH 4.3 and 37 °C (16) for 4 h. The formation of the resulting BSA adducts was followed by direct spectrophotometry and spectrofluorimetry, and also by the colorimetric TBA test (17). The modified BSA was dialyzed against water for 72 h at 4 °C with three changes of water and then lyophilized. In one preparation, covalent modification of BSA was achieved upon reaction with TMP instead of MDA. The resulting modified BSA did not give any fluorescence but gave an increased absorption at 280 nm, the maximal absorbance wavelength of enaminals (18). This preparation was assumed to be Schiff bases 1/1. The modified BSA which was obtained upon reaction with free MDA exhibited fluorescence (excitation at 350 or 390 nm and emission at 460 nm) characteristic of 1,1′-disubstituted 1-amino-3-iminopropenes (19, 20). Accordingly, it was assumed to be conjugated Schiff bases MDA 1/BSA 2. In Vitro Lipid Peroxidation. Microsomes (0.5 mg of proteins/mL) were incubated at 37 °C in 20 mM Tris buffer (pH 7.4) in the presence of 6 µM FeSO4 and 0.5 mM ascorbate.

1178 Chem. Res. Toxicol., Vol. 11, No. 10, 1998

Figure 1. Time courses of the reactions of malonaldehyde and 4-hydroxynonenal with 1-methyl-2-phenylindole at 45 °C. The reactions were performed with either aldehyde at 20 µM, in the presence of 10 mM 1-methyl-2-phenylindole and 15% (v/v) hydrochloric acid or methanesulfonic acid. The reaction of MDA was slower in the presence of hydrochloric acid (0) than in the presence of methanesulfonic acid (9). The reaction of 4-HNE is negligible in the presence of hydrochloric acid (4). Complete reaction of 4-HNE was achieved within 10 min in the presence of methanesulfonic acid (2).

Figure 2. Structure of the chromophore obtained from the reaction of MDA and/or 4-HNE with 1-methyl-2-phenylindole under acidic conditions. This chromophore exhibits maximal absorbance at 586 nm. Aliquots were taken at time intervals of 10 min, up to 120 min. Our spectrophotometric procedure 1 was compared with an HPLC procedure for the determination of MDA as the TBA derivative (21). In Vivo Lipid Peroxidation. Six-week-old mice were injected with doxorubicin (ip) at 20 mg/kg. The liver and the heart were homogenized in dilute hydrochloric acid at pH 1.5. The resulting homogenates were hydrolyzed for 80 min at 60 °C. The assay of MDA was performed on 200 µL aliquots of the supernatant using the hydrochloric acid-based procedure. Either BXT 51072 (14) or ebselen (13, 22) (73 µmol/kg), two synthetic mimics of glutathione peroxidase, was administered per os, 1 h before administration of doxorubicin. Statistics. Results are expressed as the mean ( SE. For statistical evaluation of the data, comparisons were made using the paired two-tailed Student’s t test.

Results and Discussion Acid-Induced Reactions of MDA and 4-HNE with 1-Methyl-2-phenylindole. In the presence of methanesulfonic acid, both MDA and 4-HNE yielded a stable chromophore with the same maximal absorbance at 586 nm, whereas in the presence of hydrochloric acid, this 586 nm absorbance was selectively obtained with MDA (Figure 1). On the basis of 1H NMR and mass spectrometry studies, the product of the reaction of 4-HNE was found to be the same as the MDA product (46) (Figure 2).

Ge´ rard-Monnier et al.

Figure 3. Example of standard curves obtained in the HCland CH3SO3H-based acidic media. Enaldehyde concentrations are expressed as micromolar in the final reaction medium. The data which are plotted are those obtained in the HCl-based medium. The standard curves of MDA and 4-HNE in the methanesulfonic acid-based medium were not significantly distinct from that of MDA in the HCl-based medium. Using HCl instead of methanesulfonic acid, the amount of MDA can be specifically measured in the presence of 4-hydroxyalkenals (see the discussion in the text): (A) 1,1,3,3-tetramethoxypropane and (B) 4-hydroxynonenal diethylacetal.

However, the reproducibility of the 4-HNE reaction was not satisfactory, and it was found to vary markedly for distinct batches of methanesulfonic acid. Moreover, when the reactions were monitored in stoppered cuvettes, that of 4-HNE was inhibited, unlike that of MDA (data not shown). Thus, oxygen was specifically required for the chromogenic reaction of 4-HNE, and we hypothesized that trace contaminants of methanesulfonic acid were involved in essential oxidation steps. This reasoning led us to discover that the addition of 5 µM Fe(III) to the reaction medium did eliminate batch-to-batch variations and resulted in a complete and reproducible reaction of 4-HNE, independent of its own concentration. This concentration of Fe(III) did not interfere with the MDA reaction, which enabled us to use a stock solution of methanesulfonic acid containing 34 µM Fe(III) throughout this work. In agreement with our initial hypothesis, Cu(II) could be substituted for Fe(III), unlike redoxinactive metals, such as Zn(II) or Mg(II) (data not shown). The underlying mechanisms are described and discussed in the following paper (46). Optimization of the Reaction Conditions. We optimized the reaction conditions to obtain optimal yields of both MDA and 4-HNE products of 1-methyl-2-phenylindole in the presence of methanesulfonic acid, and to obtain optimal yields of the MDA product and a minimal reaction of 4-HNE in the presence of hydrochloric acid. We studied the influence of different parameters on the reactivity of MDA and 4-HNE in methanesulfonic acid and hydrochloric acid: nature and percentage of organic solvents, percentage of water, percentage of concentrated acid, and reagent concentration (data not shown). The resulting optimal conditions for the two procedures are described in the Experimental Section. The optimized kinetics are shown in Figure 1. Figure 3 gives an example of the standard curves that were obtained. Precision of the Assays of MDA and 4-Hydroxyalkenals. The within-series precision of the assay of MDA ranged from 0.9 to 3.2% (standard errors of the mean) in procedure 1 and from 1.2 to 2.4% in procedure 2. The within-series precision of the assay of 4-HNE ranged from 1.3 to 3.2%.

Colorimetric Assay of Lipid Peroxidation

Chem. Res. Toxicol., Vol. 11, No. 10, 1998 1179

Figure 4. Chromogenic interference of various aldehydes at 586 nm. The reactivities of acrolein, trans,trans-2,4-heptadienal (90%), and methylglyoxal dimethylacetal were measured in the methanesulfonic acid-based medium, as described in the text. At a final aldehyde concentration of 5 µM, the chromophoric yields were 10.3 ( 2.1, 6.9 ( 1.5, and 1.5 ( 0.2% of that of 4-HNE diethylacetal, respectively. Under the same conditions, the chromophoric yields from the reactions of trans-2-nonenal, trans-2-hexenal, and hexanal were 1.3 ( 0.9, 0.2 ( 0.7, and 0.5 ( 1.3% of that of 4-HNE diethylacetal, respectively.

The reproducibility of the MDA assay ranged from 1.3 to 3.0% (standard errors of the mean) in procedure 1 and from 1.2 to 3.6% in procedure 2. The reproducibility of the 4-HNE assay ranged from 2.8 to 4.9%. Specificity of the Assays. In the TBA test, it is known that lipid-derived monoaldehydes form adducts with TBA and interfere with the spectrophotometric measurement of the MDA-TBA adduct (3). The 586 nm absorbance was highly specific for MDA and 4-hydroxyalkenals. Other aldehydes did not give interfering absorption spectra. In the HCl-based method, the chromogenic reactivity of aldehydes other than MDA was very weak. The yields of 586 nm absorbance obtained with acrolein, trans,trans-2,4-heptadienal, methylglyoxal dimethylacetal, 4-HNE, or hexanal (final concentrations of 5 µM) were 3.6, 0.4, 1.5, 3.8, and 1.6% of that of MDA, respectively. In the presence of HCl, the chromogenic reaction of 4-HNE became significant if the samples contained 100 µM Fe(III). The role of Fe(III) is discussed in the following paper (46). Under such conditions, we found that the selectivity for MDA could be restored by the addition of probucol at a final concentration of 1 mM. Adding probucol to the reaction medium is therefore one way to perform a specific assay of MDA on samples containing exogenous Fe(III). Dienals, alkenals, alkanals, acrolein, and methylglyoxal dimethylacetal did not give any significant 586 nm absorption with the methanesulfonic acid-based method (Figure 4). Interestingly, in addition to the 586 nm chromophore, 4-hydroxyalkenals gave a second chromophore with maximal absorbance at 505 nm (Figure 5) and alkanals produced a single chromophore with maximal absorbance at 505 nm (Figure 5). This is consistent with the fact that 1 equiv of hexanal is produced in addition to the 586 nm chromophore, in the iron(III)-catalyzed reaction of 1-methyl-2-phenylindole with 4-HNE [see the following paper (46)]. The mechanism of transformation of alkanals to the 505 nm chromophore was further investigated and described (46). This distinct chromophore did not

Figure 5. Absorption spectra resulting from the reactions of 20 µM aldehydes with 5 mM 1-methyl-2-phenylindole in the methanesulfonic acid-based medium. Absorbance scans were performed after incubation at 45 °C. See the discussion in the text: (A) MDA, (B) 4-HNE, and (C) hexanal.

interfere with the measurement of MDA and 4-hydroxyalkenals at 586 nm. In the two assays, oxyhemoglobin gave a slight absorbance at 577 nm. A concentration of 0.20 g per liter of plasma enhanced the optical density at 586 nm by approximatively 0.010. Therefore, sample blanks should be run to eliminate interference due to 586 nm-absorbing substances such as oxyhemoglobin in hemolyzed samples. Sodium and lithium heparinate induced a significant absorbance at 586 nm. Up to 2.5 mM EDTA did not interfere with the two assays. Hence, the use of EDTA to prevent blood clotting is recommended with the two methodologies for ex vivo measurements. Lipid peroxides are known to decompose at high temperatures and low pH in the presence of transition metals. It has been reported that even a temperature of 37 °C and trace iron were sufficient for lipid hydroperoxide decomposition to MDA (23). This may be a complication of the TBA test since it is usually performed in the pH range of 1-3. To evaluate the interference of lipid hydroperoxides, we monitored the decomposition of 13(S)-HPOTE and 15(S)-HPETE at a final concentration of 20 µM. It is interesting to note that in our methanesulfonic-based conditions, polyunsaturated fatty acid hydroperoxides gave only a minor 586 nm absorbance, probably because of the very acidic reaction conditions; the endogenous production of MDA reached a plateau in 30 min and accounted for 5.5% of the 13(S)-HPOTE and 5.7% of the 15(S)-HPETE. The partial decomposition of

1180 Chem. Res. Toxicol., Vol. 11, No. 10, 1998

various polyunsaturated fatty acid hydroperoxides, including those derived from linoleate, linolenate (HPOTE), arachidonate (HPETE), and docosahexaenoic acid standards, was also observed by Armstrong and Brown using our methodology.3 On the basis of the highest concentrations of endogenous lipid hydroperoxides that could be envisaged in biological samples, a 6% yield of decomposition should not alter the specificity of the methanesulfonic-based assay. However, such side reactions may significantly interfere with the assay in the presence of exogenous hydroperoxides. Under the hydrochloric acid conditions, we did not observe any decomposition of polyunsaturated fatty acid hydroperoxides into MDA. Inhibitors of the Reaction of 4-HNE. Low concentrations of dienals were shown to inhibit the reaction of 4-HNE with 1-methylphenylindole. At a final concentration of 5 µM, heptadienal and hexadienal inhibited the reactivity of 4-HNE (20 µM) by 16.6 and 10.8%, respectively. At a final concentration of 10 µM, the inhibition reached 89.6 and 85.8%, respectively. This may be an important interference, as major TBA-reactive substances in the liver homogenate were reported to be alkadienals (24). Other aldehydes tested (hexenal, nonenal, acrolein, and methylglyoxal) did not interfere with the reaction. The mild oxidation conditions which are necessary for the complete reaction of 4-hydroxyalkenals suggested that some antioxidants might have an inhibitory effect. Phenolic antioxidants such as probucol and vitamin E, at a final concentration of 10 µM, were found to induce a 5% decrease in the reactivity of 4-hydroxyalkenals. BHT had no inhibitory effect. Ethoxyquine decreased the reactivity of 4-HNE by 18% at a final concentration of 100 µM. Final concentrations of 100 µM ascorbate and 200 µM GSH did not interfere with the assay. Therefore, the concentrations of natural antioxidants which may be found in tissues should not interfere with the assay. However, if higher concentrations are imposed through cell culture media or in the course of in vitro experiments, they should first be tested for possible inhibitory effects in procedure 2. Intrinsic Limitations of the Measurement of Aldehydes in Biological Samples. Intrinsic limitations of the measurement of aldehydes include the timedependent loss of free aldehydes in biological samples, due to degradation or to reaction with nucleophilic groups of biological or exogenous compounds. This is illustrated by a slow but significant reaction of diphenylamine with free MDA; at a final concentration of 500 µM, diphenylamine decreased the reaction yield of the MDA product by 9.3% in the methanesulfonic acidbased assay and by 5.6% in the presence of HCl. H2O2 was shown to interfere with the TBA test (25), and it was further demonstrated that this interference was due to its reaction with MDA (26). H2O2 was tested at final concentrations ranging from 5 to 100 µM in our two assays. The amount of MDA was underestimated by 13% in the presence of 50 µM H2O2. The inhibition was dose-dependent and reached 27% at 100 µM H2O2. Our results confirm that the MDA concentration is lowered in the presence of high concentrations of H2O2. The accuracy of the methods applied to biological samples was checked by adding known amounts of free 3

D. Armstrong and R. Browne, personal communication.

Ge´ rard-Monnier et al.

Figure 6. pH dependence of the hydrolysis of Schiff bases prepared from MDA and BSA. The experiments were performed in diluted HCl and in diluted NaOH at 60 °C for 30 min. The hydrolysis was maximal between pH 1 and 2. Subsequent kinetic experiments performed at pH 1, 1.5, and 2 showed that the hydrolysis of either BSA adduct was complete in 80 min at pH 1.5 (data not shown).

aldehydes (1 and 3 µM in the final reaction mixture) to rat brain homogenates and to human plasma. Using procedure 1, the percentages of recovery of MDA were 80.7 and 88.9% for rat brain samples and 80.2 and 85.5% for human plasma samples, respectively. Using procedure 2, the percentages of recovery of MDA were 79.8 and 90.1% for rat brain samples and 78.5 and 85% for human plasma samples, respectively. In experiments dealing with the recovery of MDA from rabbit plasma, it was found that the yield of the 586 nm product was time-dependent and decreased from 80%, assayed for 1 min upon addition of MDA, to 56%, 2 h upon addition. In biological samples, most of the MDA may quickly bind to proteins in the form of Schiff bases. To measure the hydrolysis of Schiff bases under the assay conditions, we prepared Schiff bases from the reaction of BSA with MDA. Under the methanesulfonic acid- as well as under the HCl-based conditions, the hydrolysis of such BSA adducts was very weak, i.e., 8.2 ( 2.0 and 3.0 ( 1.9%, respectively. The significance of free MDA as a marker of ex vivo lipid peroxidation is more questionable (2, 27, 28) than that of total MDA. Due to the lack of hydrolysis of Schiff bases under the mild conditions of our assay, the assay of total MDA requires prior hydrolysis of protein-containing samples. Quantitative hydrolysis of Schiff bases produced from BSA and MDA was only achieved at pH 1.5 and 60 °C in 80 min (Figure 6). Hydrolysis was drastically inhibited below pH 1. In the literature, various pH conditions were described to hydrolyze Schiff bases before measurement of the amount of MDA, or in the TBA-containing reaction

Colorimetric Assay of Lipid Peroxidation

medium. However, very few attempts were made to characterize optimal conditions of protein-bound MDA hydrolysis. As an exception, using the TBA reaction carried on tissue homogenates and protein precipitates, it was reported that a pH of 1.8 was necessary for maximal release of bound MDA, and that only a part of bound MDA was released at pH