- Email: [email protected]
KINETIC ANALYSIS OF LIPID-HYDROPEROXIDES IN PLASMA
Bioassays for Oxidative Stress Status (BOSS). Edited by W.A. Pryor © 2001 Elsevier Science B.V. All rights reserved.
157
KINETIC ANALYSIS OF LIPID-HYDROPEROXIDES IN PLASMA ANTONIO M. PASTORINO, MATILDE MAIORINO, and FULVIO URSINI
Department of Biological Chemistry, University of Padova, Padova, Italy (Received 24 January 2000; Revised 18 April 2000; Accepted 20 April 2000) Abstract—We increased the precision of chemiluminescent procedure for measuring lipid hydroperoxides in plasma or lipoproteins by (i) escaping from extraction and chromatography of lipids, (ii) using detergent dispersed lipids, and (iii) calculating the results by fitting the photon emission rate with the integrated equation, which describes the model of the series of reactions. The use of kinetics instead of the crude integration of cps increases precision because at each measurement the correct reaction pathway is tested. This was relevant for the optimization of the analytical procedure, contributing to the elimination of possible side reactions. The relationship between lipid hydroperoxide content in the sample and cps is not linear; thus, the calculation of results through internal calibration is carried out using an exponential equation. This is in agreement with the reaction mechanism and raises the point of the linear calibration previously reported in other chemiluminescent procedures. Although sensitive and precise, this procedure suffers for being time consuming, requiring approximately 30 min per sample. Moreover, since no chromatography is used, information about the hydroperoxides in different lipid classes is missing. Obviously this will be solved when a validated procedure for quantitatively extracting lipid hydroperoxides is available. © 2000 Elsevier Science Inc. Keywords—Free radical, Lipid hydroperoxide, Luminol, Chemiluminescence
LIPID HYDROPEROXIDES Lipid hydroperoxides (LOOH) are the main product of autoxidation of fatty acids and are present in virtually all foods containing fats, of which, above a given threshold, they account for spoilage and degradation [1]. Lipid hydroperoxides are also present, obviously at a much lower concentration, in living organisms where they have Antonio M. Pastorino has degrees in Chemistry (1965) and Pharmacy (1984) from the University of Padova, Italy. He served for 30 years as a research scientist at the Glaxo-Wellcome Laboratory for Biomedical Research in Verona. In the framework of studies on the anti-atherosclerotic activity of drugs, he had a long-term collaboration with the Department of Biological Chemistry of Padova, where he now serves as guest scientist. Matilde Maiorino is Associate Professor of Biochemistry at the School of Medicine of the University of Padova. Her scientific experience ranges from mechanisms of lipid peroxidation and antioxidant defense to enzymology and regulation of gene expression. Fulvio Ursini is Full Professor of Biochemistry at the School of Medicine of the University of Padova. He has experience on enzymology, lipid peroxidation, and antioxidant mechanism. His major fields are now the function of selenoenzymes and role of nutrition on oxidative resistance of plasma lipoproteins. Address correspondence to: Fulvio Ursini, M.D., Department of Biological Chemistry, University of Padova, viale G. Colombo 3, 1-35121, Padova, Italy; Fax: +39 (049) 8073310; E-Mail: ursini@civ. bio.unipd.it.
been suggested to play a role in the molecular mechanism of several pathological processes [2,3]. In fact, the concentration of LOOH and their decomposition products increases in diseases where oxidant free radicals are produced and, further, the level of lipid peroxidation has been adopted as a marker to support the claim about the free radical nature of a given pathophysiological condition [2-7]. The presence of lipid hydroperoxides in plasma lipoproteins is required for further oxidation of lipoprotein particles in the presence of both transition metals and lipoxygenase. Oxidatively modified lipoproteins seem to be involved in the activation of the series of cellular events ultimately resulting in atherosclerosis [8]. This brought to the focus the issue of understanding the source of lipid hydroperoxides in plasma lipoproteins, while the biological effect of oxidatively modified lipoproteins on vascular cells became a major issue in cardiovascular research. Hydroperoxidation of unsaturated fats is a thermodynamically favorable event, which is kinetically controlled in the foodstuff, as well as in vivo, by several elements which, because of this, are classified as antioxidants. Lipid hydroperoxides have a 2-fold relation-
Reprinted from: Free Radical Biology & Medicine, Vol. 29, No. 5, pp. 397-402, 2000
158
A. M. PASTORINO et al.
ship with antioxidants. Chain-breaking free radical scavengers limit the peroxidative chain reaction by reducing by one electron a lipid peroxy radical to a hydroperoxide, while peroxidolytic compounds, or enzymes, by reducing by two electrons hydroperoxides, prevent their decomposition leading to new initiations [9,10]. Thus, lipid hydroperoxides are the central molecules on which the major antioxidant systems operate and integrate. Following the discovery of lipoxygenases, a physiological role has been identified for lipid hydroperoxides and the concept of "peroxide tone" has been introduced, which implies that a steady state concentration of hydroperoxides is functionally relevant in maintaining the optimal cellular function, as far as activation of cycloxygenases and lipoxygenases is concerned [11,12]. Finally, the rapidly expanding field of redox regulation of elements involved in the control of gene expression and cellular signaling [13-16] highlighted the concept that, in some disease conditions, a poorly controlled balance between production and reduction of hydroperoxides could generate distorted cellular responses. REACTIONS OF LIPID HYDROPEROXIDES
The oxidizing potential of lipid hydroperoxides, through one-electron or two-electron redox transitions, drives reactions involved in (i) the damaging effect, (ii) the possible physiological effect, (iii) the antioxidant defense, and (iv) the analytic procedures. The easy molecule-assisted homolysis related to the low dissociation energy of the 0 - 0 bond (44 kcal/mol -1 ) or, more likely, the one-electron transfer in the presence of metal ions, produces radicals that cause further decomposition of fatty acid hydroperoxide, finally leading to a series of aldehydes [1]. These contribute to cellular damage and are measured in the popular thiobarbituric acid (TBA) test for lipid peroxidation. The oxidation of transition metal complexes in the presence of hydroperoxides has been used in some analytical procedures, where the chromogen is provided by the change of the redox status of the metal. The heterolytic displacement in the presence of a nucleophile—a two-electron redox transition—accounts for biological damage when thiols or methionine residues in proteins are oxidized. The same reaction provides antioxidant protection when the nucleophile belongs to the active center of an enzyme, which, in order to be fully active, has to react much faster than the molecule to be protected and the native form of which is rapidly regenerated by a suitable reductant [17]. It appears reasonable that also the possible physiological effect of hydroperoxides has to deal with oxidation of some targets via nucleophilic displacement, without free radical intermediates, possibly through an enzyme providing specificity to the reaction. A relevant example of
this mechanism has been recently described to take place in the late phase of spermatogenesis, when the mitochondrial capsule of spermatozoa is built up by oxidation of protein thiols catalyzed by the selenoenzyme phospholipid hydroperoxide glutathione peroxidase (PHGPx), which finally becomes a cross-linked component of the structure [18]. The nucleophilic displacement reaction of lipid hydroperoxides is the reaction involved in analytical procedures such as the enzymatic determination with selenium dependent peroxidases and the iodometric titration. ANALYSIS OF LIPID HYDROPEROXIDES
The analytical procedures based on one-electron redox transition have been widely used in both food chemistry and biomedical research. The oxidation of different ferrous iron complex [4,19,20] by lipid hydroperoxide generates an easily measurable colored complex, but the procedure suffers the limited specificity of the reaction. In fact, the oxidation potential of the ferrous iron complexes allows the electron transfer to other compounds, different from lipid hydroperoxides, possibly present in the reaction mixture, including oxygen. The kinetics of the reaction actually favors the reaction with LOOH, thus allowing their measurement, although in the presence of oxygen. Nevertheless, in our experience the approach provides limited precision when samples, such as plasma, containing a very low level of LOOH, are analyzed. This approach is further biased by the fact that, in the presence of ferrous iron complexes, free radicals are produced from lipid hydroperoxides, which could promote oxidation of lipids, thus leading to an overestimation of the peroxide title. Granted these radicals could be reduced by the excess of ferrous iron and that the reaction with hydroperoxides is rather fast, thus limiting the occurrence of side reaction, the intrinsic bias of the procedure is, in our opinion, not fully overcome. This is in agreement with the fact that plasma values obtained by these procedures are higher than those obtained by our chemiluminescent procedure (see below). The iodometric titration [21], taking place by a twoelectron nucleophilic displacement, is not biased by the production of new radicals. However, the problem of the easy oxidability of the donor substrate is even more relevant. Moreover, the iodine generated in the reaction could be variably adsorbed by proteins present in the sample, thus introducing a bias difficult to overcome also adopting an internal calibration. The enzymatic titration of lipid hydroperoxides with a selenium-dependent peroxidase active on lipids provides an excellent specificity, but the sensitivity is limited to that of detection of oxidized glutathione [22]. This, together with the absolute requirement of a solution where both the enzymes are active and the lipid substrate is
159
Kinetics of photon emission
fully soluble or homogeneously dispersed, limits the applicability of the procedure. Practically, using plasma or isolated lipoproteins dispersed with a detergent, the detection limit is, by our calculation, approximately 10 nmol per ml of plasma or per mg of LDL cholesterol, respectively.
NH 2
OH
NH 2
O
DETECTION OF LIPID HYDROPEROXIDES BY CHEMILUMINESCENCE
For all the above reasons, in the framework of our study on plasma lipid hydroperoxides we set up a sensitive procedure that is specific and precise as well [23-25]. The analysis of chemiluminescence emission of luminol oxidized in the presence of lipid hydroperoxides and hemin provides the required sensitivity. The specificity is substantially increased by the use of a ferric iron complex, which, practically, reacts only with hydroperoxides. This reaction generates both a hemin hydroxyl radical and a lipid alkoxy radical [1]. The latter rearranges to an epoxy ally lie radical that, on oxygen addition, generates a peroxy radical [1,25]. This is the final oxidant most likely involved in the formation of luminol radical. The limited reactivity of hemin is the first element of specificity of this test, in comparison with procedures involving ferrous iron. Moreover, in setting up and validating the analytical procedure we used PHGPx [24], a selenium peroxidase active on all lipid hydroperoxides [17]. The enzymatic reduction of LOOH completely prevented the photon emission, confirming the specificity for LOOH of the chemiluminescent reaction. We also have evidence that the chemiluminescent reaction takes places with the same kinetics for hydroperoxide derivatives of phospholipids, triglycerides, and cholesterol esters. It appears reasonable that endoperoxides must react as well, although we did not address this issue specifically. Hydrogen peroxide is at least two order of magnitude less efficient than LOOH in producing photon emission. A plausible reason for this is that the homolytic breakdown of the peroxide gives rise to an alkoxy radical and then a peroxy radical in the case of a lipid hydroperoxide, but to a hydroxyl radical on the case of hydrogen peroxide. The very high reactivity of the latter with several molecules could actually limit the oxidation of luminol. However, a different reaction pathway cannot be excluded. The oxidant generated by the interaction of hemin with the lipid hydroperoxide oxidized the luminol monoanion to the corresponding radical (Fig. 1). This can both reduce oxygen to Superoxide and add Superoxide in a concerted mechanism where two luminol radicals are required to generate the endoperoxide derivative of luminol [26]. The latter decomposes immediately giving rise to nitrogen and the excited form of aminopthalate. This excited species decays to the ground state by emitting a photon. The precision of the measurement was optimized by
+ No + hv
Fig. 1. Reactions of luminol oxidation: The dissociated enolic form of luminol is oxidized to the corresponding radical. This can both reduce oxygen to Superoxide and add Superoxide. The product of the concerted mechanism is an endoperoxide that decomposes, giving rise to nitrogen and the excited form of aminopthalate. The latter decays to ground state emitting a photon.
adopting a kinetic approach. In fact, the product of the reaction is not a given measurable compound but an electronically excited intermediate, the decay of which generates photons detected in the analytic procedure. Thus, an optimized analysis of LOOH has to take in account the kinetics of the reaction.
160
A. M. PASTORINO et al.
The photon emission rate is related to the hydroperoxide concentration and a nonlinear relationship is expected since more than one hydroperoxide is required to get the emission of one photon. In our procedure "peroxide free" Triton X-100 was used to obtain a homogeneous micellar dispersion of lipids and the hydrophobic milieu optimizes the superoxide-driven concerted reaction of luminol oxidation. Moreover, oxidative side reactions are minimized or prevented by the detergent, which operates a dilution of the peroxidic substrate in micelles [24]. A further advantage of the use of the detergent is the dissociation of hemin dimers, which are much less reactive than monomers in decomposing lipid hydroperoxides [27]. The detergent concentration has been optimized, using the kinetics of photon emission, which diverges from the model (see below) when side reactions take place. The detection hardware adopted includes a very sensitive, cooled, phototube capable of a highly efficient single photon counting phototube. The time course of the photon emission rate has the shape of a sharp peak with a rapid increase in few seconds, maximum after about 6 s and a smooth shoulder reaching background values after about 50 s. MATHEMATICAL MODEL OF PHOTON EMISSION RATE AND FITTING
To relate the photon counting profile to lipid hydroperoxide we adopted a mathematical fitting to a model [25] since a crude integration of counts not does permit the critical evaluation of the dynamics of the event. The photon emission is due to the decay of an unstable end product after a chain of reactions; therefore a simplified model of consecutive reaction can be worked out to account for the slowest reactions that represent the rate-limiting steps of the system:
where A = hydroperoxides, B = luminol-Superoxide intermediate, and C = aminopthalate excited state. The integration of the system of differential equations for two consecutive first-order reactions gives the time course of the transient intermediate B as B = A * (k!/(k2 - k t )) * (e~ klt - e~k2t)
(1)
where kj and k2 are complex constants actually accounting for more that one reaction. On the other hand, C decays in a quasi-instantaneous
rate to the fundamental state (t1/2 > 5 ns) so that the photon emission rate matches directly the rate of formation of compound C. The photon count is a differential measure and so directly related to the rate of variation of C but not to its accumulation, like in the usual concentration measurement in kinetics experiments: cps = dC/dt = k2B
(2)
Integration of data is therefore not necessary for obtaining B from C, because c.p.s. is directly proportional to the concentration of B. So the value of B of Eqn. 2 can be substituted in Eqn. 1 so obtaining a direct relationship between the photon count and concentration of A: cps = A * k2 * (k!/(k2 - kO) * (e~ k,t - e"k2t)
(3)
This equation describes the relationship between photon emission and concentration of lipid hydroperoxides. Since different free radical scavengers, possibly present in the sample, could affect the reaction rate and thus the complex constants k{ and k2, an internal standardization was necessary. The concentration of lipid hydroperoxide in the sample, corrected for the effect of inhibitors of the chemiluminescent reaction, was extrapolated at zero internal standard concentration. Notably this correction is independent from the nature of inhibitor, and does not require a precise knowledge of the mechanism of inhibition. The photon emission rates are fitted starting from an initial time identified as the last point giving background reading before the sudden increase of cps. Data were processed using Eqn. 3, to which an offset value can be eventually added, by nonlinear regression analysis using a program operating by successive iteration. The program used starts from initial provisional estimates of kj and k2 that were determined in a series of separate experiments as k{ = 0.71 ± 0.28 and k2 = 9.47 X 10~2 ± 7.40 X 10~3 (s _ 1 ± SD), using purified phospholipid hydroperoxides. From these, for every iteration the program finds the best parameters that satisfy Eqn. 3, then computes the chi square (χ2). When the χ2 difference between two successive iterations is less then a critical value (usually < 1 %) the iteration stops and the parameters of the last iteration are considered those that best satisfy Eqn. 3. As expected, the calculated values of A as cps are not a linear function of the actual hydroperoxide content, while data are best fit by an exponential equation, consistent with the concerted mechanism described above: A = qx11
(4)
Kinetics of photon emission
where x is the hydroperoxide concentration and q a constant. The exponential constant n was determined in a series of measurements on pure lipid hydroperoxides preparation as 2.03 ± 0.27. The value of both q and n in Eqn. 4 is influenced by radical quenchers and decreases as the quencher concentration increases. The mathematical simulation of the equation showed that q and n influence the shape of the calibration curve in a similar way so that it cannot be discriminated which parameter is modified by any given quencher. In the presence of quenchers and inhibitors of the chemiluminescent reaction, the calibration curve approximates to a straight line and the average slope decreases. Because the hydroperoxides of the samples and that of the internal standard are subjected to the same quenching effect, the actual value of n and q is obtained for each sample. Introducing an internal standardization, Eqn. 4 becomes A = q(x + z) n
(5)
where z is the hydroperoxide content of the sample and x the internal standard. Due to the low level of lipid hydroperoxides in lipoproteins and the presence in plasma of efficient antioxidants such as ascorbate and urate, samples are prepared by gel filtration of whole plasma with low molecular weight cut-off "desalting" columns, or affinity chromatography with heparin-Sepharose to isolate apoBcontaining lipoproteins [24]. Suitable storage of the samples requires immediate freezing in liquid nitrogen and storage at — 80°C. This "mild" sample preparation procedure guarantees that no measurable loss or generation of lipid hydroperoxides takes place within 2 weeks. CONCLUSIONS
The described kinetic procedure for measuring low amounts of lipid hydroperoxides is suitable for analysis of human plasma. The biases related to unspecific or side reactions have been eliminated by optimizing the reaction mixture composition, and this is tested at each measurement by the quality of the fitting of the model reaction. The major disadvantage of the procedure is that LOOH of different lipid classes are measured together. However, the procedure can be applied to single lipid classes or species when a procedure for isolating lipid classes without any artificial loss or generation of hydroperoxides is developed and validated. In our opinion, the HPLC procedures for measuring lipid hydroperoxides suffer the poorly reproducible ex-
161
traction of lipid hydroperoxides. Moreover, in HPLC-CL procedures the detection and integration take place in a flow cell, and chemiluminescence data are processed as absorbance or fluorescence of a stable product. This is not fully correct, in our opinion, since the photon emission is a dynamic event, which is completely meaningful only when analyzed kinetically. The effect of the "window of reading," generated by the use of a flow cell could produce a pseudolinear calibration curve of photon emission vs. lipid hydroperoxides, which is in disagreement with the basic chemistry of the reaction. In fact, kinetic analysis suggests a concerted mechanism and a secondorder kinetics of photon emission rate as a function of hydroperoxide concentration. We recently applied this procedure to demonstrate unequivocally that a "regular" breakfast increases the plasma level of lipid hydroperoxides and that antioxidants such as those of wine, taken with food, dampens the postprandial increase of plasma lipid hydroperoxides [28]. This evidence could be relevant for studying the definition of the impact of different foods containing lipid hydroperoxides and antioxidants on oxidation and oxidability of plasma lipoproteins. REFERENCES [1] Terao, J. Reactions of lipid hydroperoxides. In: Vigo-Pelfrey, C , ed. Membrane lipid oxidation (Vol. I). Boca Raton: CRC Press; 1990:219-238. [2] Clark, I. A.; Cowden, W. B.; Hunt, N. H. Free radical-induced pathology. Med. Res. Rev. 5:297-332; 1985. [3] Halliwell, B.; Gutteridge, J. M. C. Role of free radical and catalytic metal ions in human disease. Methods Enzymol. 186:1— 85; 1990. [4] Barthel, G.; Grosh, W. Peroxide value determination—comparison of some methods. J. Am. Oil. Chem. Soc. 51:540-544; 1974. [5] Janero, D. R. Malondialdehyde and thiobarbituric acid-reactivity as a diagnostic indices of lipid peroxidation and peroxidative tissue injury. Free Radic. Biol. Med. 9:515-541; 1990. [6] Pry or, W. A.; Castle, L. Chemical methods for the detection of lipid hydroperoxides. Methods Enzymol. 105:293-299; 1984. [7] Slater, T. F. Overview of methods used for detecting lipid peroxidation. Methods Enzymol 105:283-293; 1984. [8] Berliner, J. A.; Heinecke, J. W. The role of oxidized lipoproteins in atherogenesis. Free Radic. Biol. Med. 20:707-727; 1996. [9] Halliwell, B. Antioxidant characterization. Methodology and mechanism. Biochem. Pharmacol. 49:1341-1348; 1995. [10] Ursini, F.; Maiorino, M.; Sevanian, A. Membrane hydroperoxides. In: Sies, H., ed. Oxidative stress: oxidants and antioxidants. London: Academic Press; 1994:319-336. [11] Kulmacz, R. J.; Lands, W. E. M. Requirement for hydroperoxide by the cyclooxigenase and peroxidase activities of prostaglandin H synthase. Prostaglandins 25:531-540; 1983. [12] Schnurr, K.; Belkner, J.; Ursini, F.; Schewe, T.; Kühn, H. The selenoenzyme Phospholipid Hydroperoxide Glutathione Peroxidase controls the activity of the 15-Lipoxygenase with complex substrates and mantains the specificity of the oxygenation products. J. Biol. Chem. 271:4653-4658; 1996. [13] Abate, C ; Patel, L.; Rauscer, F. J. Ill; Curran, T. Redox regulation of fos and jun DNA-binding activity in vitro. Science 249: 1157-1161; 1990. [14] Pahl, H. L.; Baeuerle, P. A. Oxygen and the control of gene expression. BioEssays 16:497-502; 1994. [15] Schulze-Osthoff, K.; Los, M.; Bauerle, P. A. Redox signalling by
162
[16] [17] [18] [19] [20]
[21] [22] [23]
[24]
A. M. PASTORINO et al.
transcription factors NF-κ B and AP-1 in lymphocytes. Biochem. Pharmacol. 50:735-741; 1995. Suzuki, Y. J.; Forman, H. J.; Sevanian, A. Oxidants as stimulators of signal transduction. Free Radio. Biol. Med. 22:269-285; 1997. Ursini, F.; Maiorino, M.; Brigelius-Flohe, R.; Aumann, K. D.; Roveri, A.; Schomburg, D.; Flohe, L. The diversity of glutathione peroxidases. Methods Enzymol. 252:38-53; 1995. Ursini, F.; Heim, S.; Kiess, M.; Maiorino, M.; Roveri, A.; Wissing, J.; Flohe, L. Dual function of the selenoprotein PHGPx during sperm maturation. Science 285:1393-1396; 1999. Stine, C. M ; Harland, H. A.; Coulter, S. T.; Jenness, R. A. Modified peroxide test for detection of lipid oxidation in dairy products. J. Dairy Sei. 37:202-208; 1954. Nourooz-Zadeh, J.; Tajaddini-Sarmadi, J.; Wolff, S. P. Measurement of plasma hydroperoxide concentrations by the ferrous oxidation-xylenol orange assay in conjunction with triphenilphosphine. Anal. Biochem. 220:403-409; 1994. Cramer, G. L.; Miller, J. F.; Pendleton, R. B.; Lands, E. M. Iodometric measurement of lipid hydroperoxides in human plasma. Anal. Biochem. 193:204-211; 1991. Maiorino, M.; Roveri, A.; Ursini, F.; Gregolin, C. Enzymatic determination of membrane lipid peroxidation. Free Radic. Biol. Med. 1:203-209; 1985. Zamburlini, A.; Maiorino, M.; Barbera, P.; Pastorino, A. M.; Roveri, A.; Cominacini, L.; Ursini, F. Measurement of lipid hydroperoxides in plasma lipoproteins by a new highly-sensitive "single photon counting" luminometer. Biochim. Biophys. Acta 1256:233-240; 1995. Zamburlini, A.; Maiorino, M.; Barbera, P.; Roveri, A.; Ursini, F.
[25]
[26] [27] [28]
Direct measurement by single photon counting of lipid hydroperoxides in human plasma and lipoproteins. Anal. Biochem. 232: 107-113; 1995. Pastorino, A. M.; Zamburlini, A.; Zennaro, L.; Maiorino, M.; Ursini, F. Measurement of lipid hydroperoxides in human plasma and lipoproteins by kinetic analysis of photon emission. Methods Enzymol. 300:33-43; 1998. Faulkner, K.; Fridovich, I. Luminol and Lucigenin as detectors for Superoxide. Free Radic. Biol. Med. 15:447-451; 1993. Brown, S. B.; Dean, T. C ; Jones, P. Catalytic activity of iron(III)centered catalysts. Role of dimerization in the catalytic action of ferrihemes. Biochem. J. 117:741-744; 1970. Ursini, F.; Zamburlini, A.; Cazzolato, G.; Maiorino, M.; BittoloBon, G.; Sevanian, A. Post-prandial lipid hydroperoxides: a possible link between diet and atherosclerosis. Free Radic. Biol. Med. 25:250-252; 1998.
ABBREVIATIONS
LOOH—lipid hydroperoxide LDL—low-density lipoprotein PHGPx—phospholipid hydroperoxide glutathione peroxidase HPLC-CL—high-performance liquid chromatography; chemiluminescence detection
157
KINETIC ANALYSIS OF LIPID-HYDROPEROXIDES IN PLASMA ANTONIO M. PASTORINO, MATILDE MAIORINO, and FULVIO URSINI
Department of Biological Chemistry, University of Padova, Padova, Italy (Received 24 January 2000; Revised 18 April 2000; Accepted 20 April 2000) Abstract—We increased the precision of chemiluminescent procedure for measuring lipid hydroperoxides in plasma or lipoproteins by (i) escaping from extraction and chromatography of lipids, (ii) using detergent dispersed lipids, and (iii) calculating the results by fitting the photon emission rate with the integrated equation, which describes the model of the series of reactions. The use of kinetics instead of the crude integration of cps increases precision because at each measurement the correct reaction pathway is tested. This was relevant for the optimization of the analytical procedure, contributing to the elimination of possible side reactions. The relationship between lipid hydroperoxide content in the sample and cps is not linear; thus, the calculation of results through internal calibration is carried out using an exponential equation. This is in agreement with the reaction mechanism and raises the point of the linear calibration previously reported in other chemiluminescent procedures. Although sensitive and precise, this procedure suffers for being time consuming, requiring approximately 30 min per sample. Moreover, since no chromatography is used, information about the hydroperoxides in different lipid classes is missing. Obviously this will be solved when a validated procedure for quantitatively extracting lipid hydroperoxides is available. © 2000 Elsevier Science Inc. Keywords—Free radical, Lipid hydroperoxide, Luminol, Chemiluminescence
LIPID HYDROPEROXIDES Lipid hydroperoxides (LOOH) are the main product of autoxidation of fatty acids and are present in virtually all foods containing fats, of which, above a given threshold, they account for spoilage and degradation [1]. Lipid hydroperoxides are also present, obviously at a much lower concentration, in living organisms where they have Antonio M. Pastorino has degrees in Chemistry (1965) and Pharmacy (1984) from the University of Padova, Italy. He served for 30 years as a research scientist at the Glaxo-Wellcome Laboratory for Biomedical Research in Verona. In the framework of studies on the anti-atherosclerotic activity of drugs, he had a long-term collaboration with the Department of Biological Chemistry of Padova, where he now serves as guest scientist. Matilde Maiorino is Associate Professor of Biochemistry at the School of Medicine of the University of Padova. Her scientific experience ranges from mechanisms of lipid peroxidation and antioxidant defense to enzymology and regulation of gene expression. Fulvio Ursini is Full Professor of Biochemistry at the School of Medicine of the University of Padova. He has experience on enzymology, lipid peroxidation, and antioxidant mechanism. His major fields are now the function of selenoenzymes and role of nutrition on oxidative resistance of plasma lipoproteins. Address correspondence to: Fulvio Ursini, M.D., Department of Biological Chemistry, University of Padova, viale G. Colombo 3, 1-35121, Padova, Italy; Fax: +39 (049) 8073310; E-Mail: ursini@civ. bio.unipd.it.
been suggested to play a role in the molecular mechanism of several pathological processes [2,3]. In fact, the concentration of LOOH and their decomposition products increases in diseases where oxidant free radicals are produced and, further, the level of lipid peroxidation has been adopted as a marker to support the claim about the free radical nature of a given pathophysiological condition [2-7]. The presence of lipid hydroperoxides in plasma lipoproteins is required for further oxidation of lipoprotein particles in the presence of both transition metals and lipoxygenase. Oxidatively modified lipoproteins seem to be involved in the activation of the series of cellular events ultimately resulting in atherosclerosis [8]. This brought to the focus the issue of understanding the source of lipid hydroperoxides in plasma lipoproteins, while the biological effect of oxidatively modified lipoproteins on vascular cells became a major issue in cardiovascular research. Hydroperoxidation of unsaturated fats is a thermodynamically favorable event, which is kinetically controlled in the foodstuff, as well as in vivo, by several elements which, because of this, are classified as antioxidants. Lipid hydroperoxides have a 2-fold relation-
Reprinted from: Free Radical Biology & Medicine, Vol. 29, No. 5, pp. 397-402, 2000
158
A. M. PASTORINO et al.
ship with antioxidants. Chain-breaking free radical scavengers limit the peroxidative chain reaction by reducing by one electron a lipid peroxy radical to a hydroperoxide, while peroxidolytic compounds, or enzymes, by reducing by two electrons hydroperoxides, prevent their decomposition leading to new initiations [9,10]. Thus, lipid hydroperoxides are the central molecules on which the major antioxidant systems operate and integrate. Following the discovery of lipoxygenases, a physiological role has been identified for lipid hydroperoxides and the concept of "peroxide tone" has been introduced, which implies that a steady state concentration of hydroperoxides is functionally relevant in maintaining the optimal cellular function, as far as activation of cycloxygenases and lipoxygenases is concerned [11,12]. Finally, the rapidly expanding field of redox regulation of elements involved in the control of gene expression and cellular signaling [13-16] highlighted the concept that, in some disease conditions, a poorly controlled balance between production and reduction of hydroperoxides could generate distorted cellular responses. REACTIONS OF LIPID HYDROPEROXIDES
The oxidizing potential of lipid hydroperoxides, through one-electron or two-electron redox transitions, drives reactions involved in (i) the damaging effect, (ii) the possible physiological effect, (iii) the antioxidant defense, and (iv) the analytic procedures. The easy molecule-assisted homolysis related to the low dissociation energy of the 0 - 0 bond (44 kcal/mol -1 ) or, more likely, the one-electron transfer in the presence of metal ions, produces radicals that cause further decomposition of fatty acid hydroperoxide, finally leading to a series of aldehydes [1]. These contribute to cellular damage and are measured in the popular thiobarbituric acid (TBA) test for lipid peroxidation. The oxidation of transition metal complexes in the presence of hydroperoxides has been used in some analytical procedures, where the chromogen is provided by the change of the redox status of the metal. The heterolytic displacement in the presence of a nucleophile—a two-electron redox transition—accounts for biological damage when thiols or methionine residues in proteins are oxidized. The same reaction provides antioxidant protection when the nucleophile belongs to the active center of an enzyme, which, in order to be fully active, has to react much faster than the molecule to be protected and the native form of which is rapidly regenerated by a suitable reductant [17]. It appears reasonable that also the possible physiological effect of hydroperoxides has to deal with oxidation of some targets via nucleophilic displacement, without free radical intermediates, possibly through an enzyme providing specificity to the reaction. A relevant example of
this mechanism has been recently described to take place in the late phase of spermatogenesis, when the mitochondrial capsule of spermatozoa is built up by oxidation of protein thiols catalyzed by the selenoenzyme phospholipid hydroperoxide glutathione peroxidase (PHGPx), which finally becomes a cross-linked component of the structure [18]. The nucleophilic displacement reaction of lipid hydroperoxides is the reaction involved in analytical procedures such as the enzymatic determination with selenium dependent peroxidases and the iodometric titration. ANALYSIS OF LIPID HYDROPEROXIDES
The analytical procedures based on one-electron redox transition have been widely used in both food chemistry and biomedical research. The oxidation of different ferrous iron complex [4,19,20] by lipid hydroperoxide generates an easily measurable colored complex, but the procedure suffers the limited specificity of the reaction. In fact, the oxidation potential of the ferrous iron complexes allows the electron transfer to other compounds, different from lipid hydroperoxides, possibly present in the reaction mixture, including oxygen. The kinetics of the reaction actually favors the reaction with LOOH, thus allowing their measurement, although in the presence of oxygen. Nevertheless, in our experience the approach provides limited precision when samples, such as plasma, containing a very low level of LOOH, are analyzed. This approach is further biased by the fact that, in the presence of ferrous iron complexes, free radicals are produced from lipid hydroperoxides, which could promote oxidation of lipids, thus leading to an overestimation of the peroxide title. Granted these radicals could be reduced by the excess of ferrous iron and that the reaction with hydroperoxides is rather fast, thus limiting the occurrence of side reaction, the intrinsic bias of the procedure is, in our opinion, not fully overcome. This is in agreement with the fact that plasma values obtained by these procedures are higher than those obtained by our chemiluminescent procedure (see below). The iodometric titration [21], taking place by a twoelectron nucleophilic displacement, is not biased by the production of new radicals. However, the problem of the easy oxidability of the donor substrate is even more relevant. Moreover, the iodine generated in the reaction could be variably adsorbed by proteins present in the sample, thus introducing a bias difficult to overcome also adopting an internal calibration. The enzymatic titration of lipid hydroperoxides with a selenium-dependent peroxidase active on lipids provides an excellent specificity, but the sensitivity is limited to that of detection of oxidized glutathione [22]. This, together with the absolute requirement of a solution where both the enzymes are active and the lipid substrate is
159
Kinetics of photon emission
fully soluble or homogeneously dispersed, limits the applicability of the procedure. Practically, using plasma or isolated lipoproteins dispersed with a detergent, the detection limit is, by our calculation, approximately 10 nmol per ml of plasma or per mg of LDL cholesterol, respectively.
NH 2
OH
NH 2
O
DETECTION OF LIPID HYDROPEROXIDES BY CHEMILUMINESCENCE
For all the above reasons, in the framework of our study on plasma lipid hydroperoxides we set up a sensitive procedure that is specific and precise as well [23-25]. The analysis of chemiluminescence emission of luminol oxidized in the presence of lipid hydroperoxides and hemin provides the required sensitivity. The specificity is substantially increased by the use of a ferric iron complex, which, practically, reacts only with hydroperoxides. This reaction generates both a hemin hydroxyl radical and a lipid alkoxy radical [1]. The latter rearranges to an epoxy ally lie radical that, on oxygen addition, generates a peroxy radical [1,25]. This is the final oxidant most likely involved in the formation of luminol radical. The limited reactivity of hemin is the first element of specificity of this test, in comparison with procedures involving ferrous iron. Moreover, in setting up and validating the analytical procedure we used PHGPx [24], a selenium peroxidase active on all lipid hydroperoxides [17]. The enzymatic reduction of LOOH completely prevented the photon emission, confirming the specificity for LOOH of the chemiluminescent reaction. We also have evidence that the chemiluminescent reaction takes places with the same kinetics for hydroperoxide derivatives of phospholipids, triglycerides, and cholesterol esters. It appears reasonable that endoperoxides must react as well, although we did not address this issue specifically. Hydrogen peroxide is at least two order of magnitude less efficient than LOOH in producing photon emission. A plausible reason for this is that the homolytic breakdown of the peroxide gives rise to an alkoxy radical and then a peroxy radical in the case of a lipid hydroperoxide, but to a hydroxyl radical on the case of hydrogen peroxide. The very high reactivity of the latter with several molecules could actually limit the oxidation of luminol. However, a different reaction pathway cannot be excluded. The oxidant generated by the interaction of hemin with the lipid hydroperoxide oxidized the luminol monoanion to the corresponding radical (Fig. 1). This can both reduce oxygen to Superoxide and add Superoxide in a concerted mechanism where two luminol radicals are required to generate the endoperoxide derivative of luminol [26]. The latter decomposes immediately giving rise to nitrogen and the excited form of aminopthalate. This excited species decays to the ground state by emitting a photon. The precision of the measurement was optimized by
+ No + hv
Fig. 1. Reactions of luminol oxidation: The dissociated enolic form of luminol is oxidized to the corresponding radical. This can both reduce oxygen to Superoxide and add Superoxide. The product of the concerted mechanism is an endoperoxide that decomposes, giving rise to nitrogen and the excited form of aminopthalate. The latter decays to ground state emitting a photon.
adopting a kinetic approach. In fact, the product of the reaction is not a given measurable compound but an electronically excited intermediate, the decay of which generates photons detected in the analytic procedure. Thus, an optimized analysis of LOOH has to take in account the kinetics of the reaction.
160
A. M. PASTORINO et al.
The photon emission rate is related to the hydroperoxide concentration and a nonlinear relationship is expected since more than one hydroperoxide is required to get the emission of one photon. In our procedure "peroxide free" Triton X-100 was used to obtain a homogeneous micellar dispersion of lipids and the hydrophobic milieu optimizes the superoxide-driven concerted reaction of luminol oxidation. Moreover, oxidative side reactions are minimized or prevented by the detergent, which operates a dilution of the peroxidic substrate in micelles [24]. A further advantage of the use of the detergent is the dissociation of hemin dimers, which are much less reactive than monomers in decomposing lipid hydroperoxides [27]. The detergent concentration has been optimized, using the kinetics of photon emission, which diverges from the model (see below) when side reactions take place. The detection hardware adopted includes a very sensitive, cooled, phototube capable of a highly efficient single photon counting phototube. The time course of the photon emission rate has the shape of a sharp peak with a rapid increase in few seconds, maximum after about 6 s and a smooth shoulder reaching background values after about 50 s. MATHEMATICAL MODEL OF PHOTON EMISSION RATE AND FITTING
To relate the photon counting profile to lipid hydroperoxide we adopted a mathematical fitting to a model [25] since a crude integration of counts not does permit the critical evaluation of the dynamics of the event. The photon emission is due to the decay of an unstable end product after a chain of reactions; therefore a simplified model of consecutive reaction can be worked out to account for the slowest reactions that represent the rate-limiting steps of the system:
where A = hydroperoxides, B = luminol-Superoxide intermediate, and C = aminopthalate excited state. The integration of the system of differential equations for two consecutive first-order reactions gives the time course of the transient intermediate B as B = A * (k!/(k2 - k t )) * (e~ klt - e~k2t)
(1)
where kj and k2 are complex constants actually accounting for more that one reaction. On the other hand, C decays in a quasi-instantaneous
rate to the fundamental state (t1/2 > 5 ns) so that the photon emission rate matches directly the rate of formation of compound C. The photon count is a differential measure and so directly related to the rate of variation of C but not to its accumulation, like in the usual concentration measurement in kinetics experiments: cps = dC/dt = k2B
(2)
Integration of data is therefore not necessary for obtaining B from C, because c.p.s. is directly proportional to the concentration of B. So the value of B of Eqn. 2 can be substituted in Eqn. 1 so obtaining a direct relationship between the photon count and concentration of A: cps = A * k2 * (k!/(k2 - kO) * (e~ k,t - e"k2t)
(3)
This equation describes the relationship between photon emission and concentration of lipid hydroperoxides. Since different free radical scavengers, possibly present in the sample, could affect the reaction rate and thus the complex constants k{ and k2, an internal standardization was necessary. The concentration of lipid hydroperoxide in the sample, corrected for the effect of inhibitors of the chemiluminescent reaction, was extrapolated at zero internal standard concentration. Notably this correction is independent from the nature of inhibitor, and does not require a precise knowledge of the mechanism of inhibition. The photon emission rates are fitted starting from an initial time identified as the last point giving background reading before the sudden increase of cps. Data were processed using Eqn. 3, to which an offset value can be eventually added, by nonlinear regression analysis using a program operating by successive iteration. The program used starts from initial provisional estimates of kj and k2 that were determined in a series of separate experiments as k{ = 0.71 ± 0.28 and k2 = 9.47 X 10~2 ± 7.40 X 10~3 (s _ 1 ± SD), using purified phospholipid hydroperoxides. From these, for every iteration the program finds the best parameters that satisfy Eqn. 3, then computes the chi square (χ2). When the χ2 difference between two successive iterations is less then a critical value (usually < 1 %) the iteration stops and the parameters of the last iteration are considered those that best satisfy Eqn. 3. As expected, the calculated values of A as cps are not a linear function of the actual hydroperoxide content, while data are best fit by an exponential equation, consistent with the concerted mechanism described above: A = qx11
(4)
Kinetics of photon emission
where x is the hydroperoxide concentration and q a constant. The exponential constant n was determined in a series of measurements on pure lipid hydroperoxides preparation as 2.03 ± 0.27. The value of both q and n in Eqn. 4 is influenced by radical quenchers and decreases as the quencher concentration increases. The mathematical simulation of the equation showed that q and n influence the shape of the calibration curve in a similar way so that it cannot be discriminated which parameter is modified by any given quencher. In the presence of quenchers and inhibitors of the chemiluminescent reaction, the calibration curve approximates to a straight line and the average slope decreases. Because the hydroperoxides of the samples and that of the internal standard are subjected to the same quenching effect, the actual value of n and q is obtained for each sample. Introducing an internal standardization, Eqn. 4 becomes A = q(x + z) n
(5)
where z is the hydroperoxide content of the sample and x the internal standard. Due to the low level of lipid hydroperoxides in lipoproteins and the presence in plasma of efficient antioxidants such as ascorbate and urate, samples are prepared by gel filtration of whole plasma with low molecular weight cut-off "desalting" columns, or affinity chromatography with heparin-Sepharose to isolate apoBcontaining lipoproteins [24]. Suitable storage of the samples requires immediate freezing in liquid nitrogen and storage at — 80°C. This "mild" sample preparation procedure guarantees that no measurable loss or generation of lipid hydroperoxides takes place within 2 weeks. CONCLUSIONS
The described kinetic procedure for measuring low amounts of lipid hydroperoxides is suitable for analysis of human plasma. The biases related to unspecific or side reactions have been eliminated by optimizing the reaction mixture composition, and this is tested at each measurement by the quality of the fitting of the model reaction. The major disadvantage of the procedure is that LOOH of different lipid classes are measured together. However, the procedure can be applied to single lipid classes or species when a procedure for isolating lipid classes without any artificial loss or generation of hydroperoxides is developed and validated. In our opinion, the HPLC procedures for measuring lipid hydroperoxides suffer the poorly reproducible ex-
161
traction of lipid hydroperoxides. Moreover, in HPLC-CL procedures the detection and integration take place in a flow cell, and chemiluminescence data are processed as absorbance or fluorescence of a stable product. This is not fully correct, in our opinion, since the photon emission is a dynamic event, which is completely meaningful only when analyzed kinetically. The effect of the "window of reading," generated by the use of a flow cell could produce a pseudolinear calibration curve of photon emission vs. lipid hydroperoxides, which is in disagreement with the basic chemistry of the reaction. In fact, kinetic analysis suggests a concerted mechanism and a secondorder kinetics of photon emission rate as a function of hydroperoxide concentration. We recently applied this procedure to demonstrate unequivocally that a "regular" breakfast increases the plasma level of lipid hydroperoxides and that antioxidants such as those of wine, taken with food, dampens the postprandial increase of plasma lipid hydroperoxides [28]. This evidence could be relevant for studying the definition of the impact of different foods containing lipid hydroperoxides and antioxidants on oxidation and oxidability of plasma lipoproteins. REFERENCES [1] Terao, J. Reactions of lipid hydroperoxides. In: Vigo-Pelfrey, C , ed. Membrane lipid oxidation (Vol. I). Boca Raton: CRC Press; 1990:219-238. [2] Clark, I. A.; Cowden, W. B.; Hunt, N. H. Free radical-induced pathology. Med. Res. Rev. 5:297-332; 1985. [3] Halliwell, B.; Gutteridge, J. M. C. Role of free radical and catalytic metal ions in human disease. Methods Enzymol. 186:1— 85; 1990. [4] Barthel, G.; Grosh, W. Peroxide value determination—comparison of some methods. J. Am. Oil. Chem. Soc. 51:540-544; 1974. [5] Janero, D. R. Malondialdehyde and thiobarbituric acid-reactivity as a diagnostic indices of lipid peroxidation and peroxidative tissue injury. Free Radic. Biol. Med. 9:515-541; 1990. [6] Pry or, W. A.; Castle, L. Chemical methods for the detection of lipid hydroperoxides. Methods Enzymol. 105:293-299; 1984. [7] Slater, T. F. Overview of methods used for detecting lipid peroxidation. Methods Enzymol 105:283-293; 1984. [8] Berliner, J. A.; Heinecke, J. W. The role of oxidized lipoproteins in atherogenesis. Free Radic. Biol. Med. 20:707-727; 1996. [9] Halliwell, B. Antioxidant characterization. Methodology and mechanism. Biochem. Pharmacol. 49:1341-1348; 1995. [10] Ursini, F.; Maiorino, M.; Sevanian, A. Membrane hydroperoxides. In: Sies, H., ed. Oxidative stress: oxidants and antioxidants. London: Academic Press; 1994:319-336. [11] Kulmacz, R. J.; Lands, W. E. M. Requirement for hydroperoxide by the cyclooxigenase and peroxidase activities of prostaglandin H synthase. Prostaglandins 25:531-540; 1983. [12] Schnurr, K.; Belkner, J.; Ursini, F.; Schewe, T.; Kühn, H. The selenoenzyme Phospholipid Hydroperoxide Glutathione Peroxidase controls the activity of the 15-Lipoxygenase with complex substrates and mantains the specificity of the oxygenation products. J. Biol. Chem. 271:4653-4658; 1996. [13] Abate, C ; Patel, L.; Rauscer, F. J. Ill; Curran, T. Redox regulation of fos and jun DNA-binding activity in vitro. Science 249: 1157-1161; 1990. [14] Pahl, H. L.; Baeuerle, P. A. Oxygen and the control of gene expression. BioEssays 16:497-502; 1994. [15] Schulze-Osthoff, K.; Los, M.; Bauerle, P. A. Redox signalling by
162
[16] [17] [18] [19] [20]
[21] [22] [23]
[24]
A. M. PASTORINO et al.
transcription factors NF-κ B and AP-1 in lymphocytes. Biochem. Pharmacol. 50:735-741; 1995. Suzuki, Y. J.; Forman, H. J.; Sevanian, A. Oxidants as stimulators of signal transduction. Free Radio. Biol. Med. 22:269-285; 1997. Ursini, F.; Maiorino, M.; Brigelius-Flohe, R.; Aumann, K. D.; Roveri, A.; Schomburg, D.; Flohe, L. The diversity of glutathione peroxidases. Methods Enzymol. 252:38-53; 1995. Ursini, F.; Heim, S.; Kiess, M.; Maiorino, M.; Roveri, A.; Wissing, J.; Flohe, L. Dual function of the selenoprotein PHGPx during sperm maturation. Science 285:1393-1396; 1999. Stine, C. M ; Harland, H. A.; Coulter, S. T.; Jenness, R. A. Modified peroxide test for detection of lipid oxidation in dairy products. J. Dairy Sei. 37:202-208; 1954. Nourooz-Zadeh, J.; Tajaddini-Sarmadi, J.; Wolff, S. P. Measurement of plasma hydroperoxide concentrations by the ferrous oxidation-xylenol orange assay in conjunction with triphenilphosphine. Anal. Biochem. 220:403-409; 1994. Cramer, G. L.; Miller, J. F.; Pendleton, R. B.; Lands, E. M. Iodometric measurement of lipid hydroperoxides in human plasma. Anal. Biochem. 193:204-211; 1991. Maiorino, M.; Roveri, A.; Ursini, F.; Gregolin, C. Enzymatic determination of membrane lipid peroxidation. Free Radic. Biol. Med. 1:203-209; 1985. Zamburlini, A.; Maiorino, M.; Barbera, P.; Pastorino, A. M.; Roveri, A.; Cominacini, L.; Ursini, F. Measurement of lipid hydroperoxides in plasma lipoproteins by a new highly-sensitive "single photon counting" luminometer. Biochim. Biophys. Acta 1256:233-240; 1995. Zamburlini, A.; Maiorino, M.; Barbera, P.; Roveri, A.; Ursini, F.
[25]
[26] [27] [28]
Direct measurement by single photon counting of lipid hydroperoxides in human plasma and lipoproteins. Anal. Biochem. 232: 107-113; 1995. Pastorino, A. M.; Zamburlini, A.; Zennaro, L.; Maiorino, M.; Ursini, F. Measurement of lipid hydroperoxides in human plasma and lipoproteins by kinetic analysis of photon emission. Methods Enzymol. 300:33-43; 1998. Faulkner, K.; Fridovich, I. Luminol and Lucigenin as detectors for Superoxide. Free Radic. Biol. Med. 15:447-451; 1993. Brown, S. B.; Dean, T. C ; Jones, P. Catalytic activity of iron(III)centered catalysts. Role of dimerization in the catalytic action of ferrihemes. Biochem. J. 117:741-744; 1970. Ursini, F.; Zamburlini, A.; Cazzolato, G.; Maiorino, M.; BittoloBon, G.; Sevanian, A. Post-prandial lipid hydroperoxides: a possible link between diet and atherosclerosis. Free Radic. Biol. Med. 25:250-252; 1998.
ABBREVIATIONS
LOOH—lipid hydroperoxide LDL—low-density lipoprotein PHGPx—phospholipid hydroperoxide glutathione peroxidase HPLC-CL—high-performance liquid chromatography; chemiluminescence detection