206 Measurement of Lipid Peroxidation In Vivo: A Comparison of Different Procedures Alfonso Pompella, Emilia Maellaro, Alessandro F. Casini, Marco Ferrali, Lucia Ciccoli and Mario Comporti* Istitutodi Patologia Generale deil'Universit~ di Siena, Via Laterino 8, 53100 Siena, Italy A study was undertaken to investigate whether some of the methods commonly used to detect lipid peroxidation of cellular membranes in vivo correlate with each other. The study was performed with the livers of bromoben- zene-intoxicated mice, in which lipid peroxidation devel- ops when the depletion of glutathione (GSH) reaches a threshold value. The methods tested and compared were the following: i) measurement of the malondialdehyde (MDA) content of the fiver; ii) detection of diene conjuga- tion absorption in liver phospholipids; iii) measurement of the loss of polyunsaturated fatty acids in liver phos- phofipids; and iv) determination of carbonyl functions formed in acyl residues of membrane phospholipids as a result of the peroxidative breakdown of phospholipid fat- ty acids. Correlations among the values obtained with these methods showed high statistical significances, in- dicating that the procedures measure lipid pcroxidation in vivo with comparable reliability. Analogously, the four methods appeared also to correlate when applied to in vitro microsomal lipid peroxidation. Lipids 22, 206-211 (1987). During the past 30 years, a great deal of experimental evidence has accumulated suggesting that lipid peroxida- tion in cellular membranes is implicated in a variety of pathological conditions. These include liver cell injury by a number of toxins (CCL and other halogenated hydrocar- bons [1-6]; the effect of bromobenzene and other aryl halides acting as glutathione [GSH] depleting agents [7-9], injury due to other hepatotoxins, probably in- cluding ethanol [10]); lung damage after exposure to nitrogen dioxide (11) and ozone (12) or intoxication with the herbicide paraquat (13); increased red blood cell permeability and hemolysis associated with vitamin E deficiency (14); retrolental fibroplasia (15); paroxysmal nocturnal hemoglobinuria (16); abetalipoproteinemia (17); cell damage caused by ionizing radiation (18); and several aspects of oxygen toxicity (19). Because of the growing importance of lipid peroxida- tion in the biomedical field, an increasing need for reliable methods to detect these processes has been perceived. Many methods are available to measure the extent of lipid peroxidation, but most are based upon the measurement of the products originating from the process at different stages. These methods include (20) i) the classic thiobar- bituric acid (TBA) reaction to measure malondialdehyde (MDA) (21); ii) detection of the UV absorption character- istic of conjugated dienes (22); iii) fluorescent analysis of lipid peroxidation products (23); iv) measurement of ethane and pentane formation (24,25); v) detection of chemiluminescence (26); vi) measurement of oxygen up- take (27); vii) measurement of the loss of polyunsaturated fatty acids in membrane phospholipids (28); viii) the detec- tion of lipid hydroperoxides {29); and ix) measurement of specific aldehydes such as alkenals (30). *To whom correspondence should be addressed. The detection of lipid peroxidation in biological systems in vitro is relatively simple since, when the reaction has been blocked, it is conceivable that the measurement of a certain product in the sample represents a reliable esti- mation. On the other hand, when detection of lipid perox- idation in vivo is attempted, many problems arise: the oxidation product may be rapidly metabolized or removed from the tissue under study; it may have interacted with different cellular substances so that it cannot be found in the free form; or it may be formed as an artifact dur- ing sampling of the tissue and other technical procedures. For these reasons, the detection of lipid peroxidation in vivo has often puzzled the biochemical pathologist who is primarily concerned with understanding the in vivo events. The present work was undertaken to determine whether the methods more commonly used in various laboratories to detect lipid peroxidation in vivo correlate and to test the reliability of each individual method. The study was performed with the livers of bromobenzene- intoxicated mice, in which lipid peroxidation develops when the depletion of GSH reaches a threshold value (9). Bromobenzene hepatotoxicity represents a good model for the study of in vivo lipid peroxidation, since in this experimental condition the level of detectable lipid perox- idation is far greater than in the case of CCL or BrCC13 hepatotoxicity (9,31). The following methods were compared: i) measurement of the MDA content of the liver; ii) detection of diene con- jugation absorption in liver phospholipids; iii) measure- ment of the loss of polyunsaturated fatty acids in liver phospholipids; and iv) measurement of the carbonyl func- tions formed in the acyl residues of membrane phos- pholipids as a result of the peroxidative breakdown of phospholipid fatty acids. The latter method was recently developed in our laboratory (32). MATERIALS AND METHODS Male NMRI albino mice (Ivanovas GmbH, Federal Re- public of Germany) weighing 20-30 g and maintained on a pellet diet (Altromin-Rieper, Bolzano, Italy) were used. In vivo experiments. The animals were fed a liquid glucose (20%) diet for two days before intoxication, ac- cording to the protocol of Wendel et al. (33), to decrease the hepatic GSH content. This regimen decreased hepatic GSH by about 50% as compared to laboratory chow-fed animals and increased the frequency of occurrence of lipid peroxidation in liver phospholipids. Bromobenzene (C. Erba, Milano, Italy) mixed with two volumes of mineral oil was administered intragastrically under light ether anesthesia, at a dose of 15 mmol/kg body weight. Control mice received mineral oil alone. All animals were fasted after intoxication. Eighteen hr after intoxication, the animals were killed by exsanguination under ether anesthesia, and the livers were quickly removed, rinsed in ice-cold saline, weighed, and divided into two portions. The first portion was LIPIDS, Vot, 22, No. 3 (1987)