ARCHIVES
OF BIOCHEMISTRY
AND
BIOPHYSICS
Vol. 299, No. 2, December, pp. 361-367, 1992
Fatty Acid Radical Formation in Rats Administered Oxidized Fatty Acids: In Viva Spin Trapping Investigation Walee Chamulitrat,’
Sandra
Laboratory of Molecular Biophysics, P.0. Box 12233, Research Triangle
J. Jordan,
and Ronald
P. Mason
National Znstitute of Environmental Park, North Carolina 27709
Health
Sciences,
National
Institutes
of Health,
Received May 28, 1992, and in revised form August 19, 1992
We report in vivo evidence for fatty acid-derived free radical metabolite formation in bile of rats dosed with spin traps and oxidized polyunsaturated fatty acids (PUFA). When rats were dosed with the spin trap 5,5dimethyll-pyrroline N-oxide (DMPO) and oxidized PUFA, the DMPO thiyl radical adduct was formed due to a reaction between oxidized PUFA and/or its metabolites with biliary glutathione. In vitro experiments were performed to determine the conditions necessary for the elimination of radical adduct formation by ex vivo reactions. Fatty acid-derived radical adducts of cY-(4-pyridyl-1-oxide)-N-tert-butylnitrone (4-POBN) were detected in vivo in bile samples collected into a mixture of iodoacetamide, desferrioxamine, and glutathione peroxidase. Upon the administration of oxidized 13C-algal fatty acids and 4-POBN, the EPR spectrum of the radical adducts present in the bile exhibited hyperfine couplings due to ‘3C. Our data demonstrate that the carbon-centered radical adducts observed in in vivo experiments are unequivocally derived from oxidized PUFA. This in vivo evidence for PUFA-derived free radical formation supports the proposal that processes involving free radicals may be the molecular basis for the previously described cytotoxicity of dietary oxidized PUFA. B 1992 Academic
Press,
Inc.
Numerous studies have focused on high dietary fat intake and its association with chronic diseases in humans. Polyunsaturated fatty acids in foodstuffs are easily oxidized and the resultant lipid peroxidation yields a variety of mutagens, promoters, and carcinogens (1, 2). The involvement of lipid peroxidation in dietary fats has been demonstrated by the increased production of fragmented products from fatty acid hydroperoxides, i.e., pentane (3), ethane (4), urinary malondialdehyde (5), and low molec’ To whom correspondence
should be addressed. Fax: (919) 541.7880.
0003.9861/92 $6.00 Copyright fi 1992 hy Academic Press, All rights of reproduction in any form
Inc. reserved.
ular weight aldehydes (6). It is well known that fatty acid peroxidation and fatty acid hydroperoxide decomposition are free radical-mediated processes (I, 2). Furthermore, fatty acid-derived free radicals are thought to be intermediates that cause damage to biomembranes, proteins, and DNA (1, 2, 7, 8). Evidence for in uivo free radical formation from dietary oxidized fatty acids includes increases in exhaled pentane gas (3), TBARS (9), glutathione peroxidase activity (lo), and chemiluminescence (11). There has been only one in uivo report using the spin-trapping technique, where the EPR spectrum of fatty acid-derived radical adduct in the liver from rats dosed with oxidized fatty acids was reported (12). The spin trap, phenyl-N-tert-butylnitrone (PBN),2 and oxidized methyl linoleate were intraperitoneally injected (12). The EPR analysis was done on the Folch extract of the rat livers obtained 30 min after the injections, and the detected PBN radical adduct was assigned to the fatty acid alkoxyl radical adduct. In addition, the 4-POBN radical adduct obtained from a mixture of 4-POBN and liver microsomes from rats which had been dosed with oxidized methyl linoleate was assigned to the fatty acid peroxyl radical adduct (12). The assignments for the identity of these radical adducts were based upon the comparison of hyperfine coupling constants obtained from in vitro experiments reported earlier (13-15). In this report, we employed an in viuo spin-trapping technique that has been developed in our laboratory (16) to reinvestigate the in uiuo free radical formation from oxidized fatty acids. Radical adducts from bile samples detected from the anesthetized rats, which had been dosed with a spin trap and oxidized fatty acids, were successfully obtained and characterized. Upon the administration of oxidized ‘“C-algal fatty acids, the EPR spectrum of the ’ Abbreviations used: PUFA, polyunsaturated fatty acids; LOOH, fatty acid hydroperoxides; DMPO, 5,5-dimethyl-1-pyrroline N-oxide; PBN, phenyl-N-tert-butylnitrone; 4-POBN, n-(4.pyridyl-l-oxide)-N-tert-butylnitrone. 361
362
CHAMULITRAT,
JORDAN,
radical adduct present in the bile exhibited hyperfine couplings due to 13C.These findings indicate that the radical adducts observed in in viva experiments are indeed formed from carbon-centered radicals of the administered oxidized PUFA. EXPERIMENTAL Oleic acid, DMPO, 4-POBN, deferoxamine mesylate (desferrioxamine), iodoacetamide, and glutathione peroxidase (from bovine erythrocytes, 135 units/mg solid, 680 units/mg prot., Lot 70H9311) were purchased from Sigma. DMPO was vacuum distilled at room temperature and stored under nitrogen at -70°C prior to use. Linoleic acid, linolenic acid, and arachidonic acid (>99%) were purchased from Nu-chek Prep (Elysian, MN). 13C-algal fatty acid mixture (CLM-2063, 13C > 98%) containingpalmitic acid (16:O) (20%), palmitoleic acid (16:l) (7%), hexadecadienoic acid (16:2) (15%), oleic acid (181) (IS%), linoleic acid (l&2) (26%), and linolenic acid (183) (11%) was obtained from Cambridge Isotope, Inc. (Woburn, MA). Oxidized fatty acids were obtained by bubbling the neat fatty acids with oxygen at room temperature to form the corresponding fatty acid hydroperoxides (LOOH) and other oxidation products. The conjugated diene formation was measured with an HP 8451A spectrophotometer. The absorbance at 234 nm was monitored periodically until the conjugated diene concentration reached about 0.9-1.1 Musing an extinction coefficient of 2.3 X lo4 cm-’ Mm’ (17). By this preparation, the hydroperoxide value was maximized and only minimal breakdown of fatty acid hydroperoxide by oxygen occurred (18). The same oxidized PLJFA batch was used for a given set of experiments. Similarly, neat 13C-algal fatty acid mixture was oxidized by bubbling with oxygen until there was no increase in 234.nm absorbance. In viva studies. Male Sprague-Dawley rats (300-400 g) which were allowed free access to both food and water were used. Nonfasted rats were anesthetized with Nembutal, and their bile ducts were cannulated using 7- to lo-cm segments of PEIO tubing (16). Anesthesia was maintained during the surgery and bile collection. 4-PORN (100 mg/kg), which had been dissolved in deionized water, and neat DMPO (500 pl/ kg) were administered by intraperitoneal injections, and oxidized PUFA was injected directly into the stomach. Bile samples were collected into plastic Eppendorf tubes during successive 20-min intervals for 2 h. Bile samples were frozen immediately on dry ice and stored at -70°C. The EPR analysis was performed within 2 days. The signal intensity of radical adducts increased with the collection t,ime maximizing with the sixth bile sample (120-140 min aft.er the administration of an oxidized PUFA). We therefore chose these bile samples to represent our in uiuo experiments In vitro studies. Chemical reactions in bile can occur during bile collection, thereby causing ex viuo free radical formation. To distinguish e.r uiuo radical adducts from those of in uivo signals, in vitro experiments were performed to investigate the possible chemical reactions and their inhibition. In vitro reactions were initiated by the addition of an aliquot of oxidized PUFA stock solutions, which had been dissolved in bile, to an Eppendorf tube containing 4.POBN and bile (obtained from untreated rats). The compounds tested for inhibition effects were desfer rioxamine, iodoacetamide, and glutathione peroxidase. Individually and in combination, desferrioxamine, iodoacetamide, and glutathione peroxidase were added to the bile and spin-trap mixture, which was followed by addition of the oxidized PUFA. After mixing, the incubation mixtures were analyzed by EPR spectroscopy for radical adduct formation. EPR measurements. The EPR spectra were recorded at room temperature on either a Bruker ESPSOO or ER 200D spectrometer operated, respectively, at 9.87 and 9.77 GHz with a IOO-kHz modulation frequency. The samples were pipetted to a quartz flat cell, which was then centered in a TM,,, cavity. The data were transferred from the ESP1600 data system to an IBM/AT where the computer simulation analyses were done using the software developed at this laboratory by D. R. Duling.
AND
MASON
RESULTS
The bile from a rat that had been administered DMPO and oxidized linoleic acid (LOOH) gave a four-line EPR spectrum with hyperfine coupling constants of uN = 15.2 G and c$ = 16.3 G (Fig. 1A). This DMPO radical adduct was identified as DMPO/GS’by comparison with the literature values (19). The doublet with a 1.9-G hyperfine coupling constant at the center of the spectrum is due to the ascorbate semidione radical from the oxidation of ascorbate in the bile. The omission of oxidized linoleic acid resulted in a very weak signal (Fig. 1B). The replacement of oxidized linoleic acid with neat linoleic acid resulted in a weaker signal (Fig. 1C). To determine if this radical adduct was being formed in the collection tube during bile collection, the sulfhydryl blocking reagent iodoacetamide (1.6 mM final concentration) was added to the collection tube. Iodoacetamide indeed inhibited the formation of DMPO/GS’ (Fig. 1D). When sufficient iodoacetamide (2.5 mM final concentration) was added, the inhibition was total (data not shown), suggesting that the DMPO/GS formation was due to the ex vivo reactions of oxidized linoleic acid and/or its metabolites with biliary glutathione. The formation of DMPO/GS’ clearly was oxidized linoleic acid-dependent and may depend on catalysis by
LOOH
-LOOH
FIG. 1. EPR spectra of radical adducts detected in bile from rats administered oxidized linoleic acid (LOOH) (520 mg/kg, ig) and DMPO (500 pi/kg, ip). (A) Complete system. (B) Omission of LOOH. (C) LOOH was replaced with linoleic acid (520 mg/kg). (D) Complete system with iodoacetamide (1.6 mM, final concentration) added into the collection tubes. Spectrometer conditions were: modulation amplitude, 0.72 G; microwave power, 20 mW; time constant, 0.66 s; scan range, 80 G; scan time, 335 s.
In Viuo FATTY
ACID
FREE
biliary metals. Note that the weak signal observed in Fig. 1B indicates that there was an er uiuo oxidation of GSH independent of the administered oxidized PUFA. When DMPO was replaced by 4-POBN, a six-line 4POBN radical adduct was detected in bile samples collected 2 h after the injections (Fig. 2A). Administration of 4-POBN alone resulted in a much weaker signal (Fig. 2B). When oxidized arachidonic acid was replaced with neat fatty acid, only a weak signal was observed (Fig. 2C). These results, including the control experiments, were also reproducible for oxidized linoleic acid (520 mg/kg) or linolenic acid (390 mg/kg) (data not shown). In contrast, the administration of the oxygenated oleic acid (1.1 g/kg) did not result in radical adduct formation (data not shown). The in vitro investigation of the formation of free radicals that are trapped by 4-POBN in bile is shown in Fig. 3. Chemical reactions involving adventitious transition metals and GSH are plausible sources for this free radical formation in bile. It was observed that both desferrioxamine and iodoacetamide inhibited free radical formation (Figs. 3B and 3C), and when combined, desferrioxamine (Fig. 3D). and iodoacetamide caused -100% inhibition When the in vitro experiment was carried out in the presence of added oxidized linoleic acid. desferrioxamine and
RADICAL
363
FORMATION in vitro
FIG. 3. EPR spectra of radical adducts detected in bile obtained from untreated rats. (A) bile (86%) mixed with 4-POHN (20 mM). (B) As in (A) with the addition of desferrioxamine (160 PM). (C) As in (A) with the addition of iodoacetamide (20 mM). (D) As in (A) with the addition ofdesferrioxamine (160 fiM) and iodoacetamide (20 mM). Spectrometer conditions were: modulation amplitude, 0.7 G; microwave power, 20 mW; time constant, 1.3 s; scan range, 55 G; scan time, 500 s.
LOOH 4-PORN
B
ij
-LOO11
-LOOH +arach!donic
acid
FIG. 2. EPR spectra of’ radical adducts detected in bile from rats administered oxidized arachidonic acid (LOOH) (260 mg/kg, ig) and 4. PORN (100 n&kg, ip). (A) Complete system. (B) Omission of LOOH. (C) IJOOH was replaced with arachidonic acid (260 mg/kg). Spectrometer conditions were: modulation amplitude, 0.72 G; microwave power, 20 mW; time constant, 1.3 s; scan range, 80 G; scan time; 671 s averaged over two scans.
iodoacetamide together inhibited, but not totally (Figs. 4A and 4B). Results shown in Figs. 1A and 1B indicate that the oxidized linoleic acid reacted with the GSH in the bile sample resulting in the Gs’ formation; therefore, glutathione peroxidase was added to the collection tube to reduce biliary fatty acid hydroperoxides to their corresponding alcohols. The glutathione peroxidase concentration was increased until the radical adduct signal was totally inhibited. As a result, the addition of the three components, iodoacetamide, desferrioxamine, and glutathione peroxidase, completely inhibited the radical adduct formation (Fig. 4C). Once it was determined that the ex: uiuo radical adduct formation could be suppressed by the addition of iodoacetamide, desferrioxamine, and glutathione peroxidase to the collection tubes, the experiment performed in Fig. 2A was repeated (Fig. 5A). With this protocol, the radical adducts detected (Fig. 5A) presumably result from the trapping of the radical(s) which were formed and trapped by 4-POBN in. uiuo. We did not know the chemical structure of the trapped radical(s), but we thought that fatty acid-derived carbon-centered radical adducts would be the most likely because these have previously been detected in vitro with 4-POBN (20-22), although the possibility that the products were adduct of oxygen-centered radical(s), e.g., fatty acid-alkoxyl radicals, cannot be excluded. We therefore represent these radical adducts as 4-POBN/
364
CHAMULITRAT,
JORDAN,
in vim LOOH 4-POBN
-“collected into iodoacetamide desfetioxamine
collected into iodoacetamide desfenioxamine glutathione peroxidase
v
10 Gauss
FIG. 4. EPR spectra of radical adducts detected in bile obtained from untreated rats and oxidized linoleic acid (LOOH) in the presence of 4POBN. (A) bile (95%) mixed with LOOH (120 CM) and 4-POBN (20 mM). (B) As in (A) with prior addition of desferrioxamine (160 PM) and iodoacetamide (25 mM). (C) As in (B) with prior addition of glutathione peroxidase (150 units/ml). Spectrometer conditions were: modulation amplitude, 0.7 G; microwave power, 20 mW; time constant, 1.3 s; scan range, 50 G; scan time, 500 s.
‘R. The simulation of the spectrum in Fig. 5A yielded a composite of ascorbate semidione (0.26 mol ratio) and 4POBN/‘R (0.74 mol ratio) with the hyperfine coupling constants shown in Table I. The omission of oxidized linoleic acid resulted in no detectable radical adduct formation, but a higher concentration of the ascorbate semidione radical (Fig. 5B). When oxidized linoleic acid was replaced with neat linoleic acid, only a very weak spectrum was detected (Fig. 5C). When bile was collected into water with an equivalent volume to that of the treatment performed in Fig. 5A, the radical adduct concentration increased by -2.5-fold, demonstrating that radical adducts from ex uiuo reactions were the major contribution to the spectrum (Fig. 5D). This spectrum was simulated as a composite of three species: ascorbate semidione (0.05 mol ratio), 4-POBN/‘R (0.71 mol ratio), and 4-POBN/GS (0.24 mol ratio) (Table I). The formation of 4-POBN/ GS’ was clearly ex uiuo (Fig. 5D), since this signal was inhibited when GSH was depleted by iodoacetamide. Since 4-POBN/a-hydroxyalkyl and other radical adducts exhibit similar hyperfine coupling constants (Table I), our assignment for 4-POBN/GS’ is based on the iodoacetamide effect. Experiments performed in Figs. 3 and 4 were essential to ensure that iron-, GSH-, and hydroperoxide-dependent ex uiuo reactions do not occur during sample collection. These er uiuo chemical reactions could account for the
AND
MASON
majority of radical adducts formed during liver homogenization and/or organic extraction which were performed in the earlier spin-trapping reports of free radical formation by oxidized fatty acids (12, 13). When treatments to suppress ex uiuo free radical reactions are not used, radical adduct formed ex vivo may be mistaken as evidence for in uivo free radical formation. An in viuo experiment with oxidized i3C-algal fatty acid was performed to determine if the 4-POBN radical adduct was indeed from the carbon-centered radicals derived from the oxidized fatty acids. Since this sample contained only 75% unsaturated fatty acids, a higher concentration of 4POBN was used. The in uiuo 4-POBN radical adduct from the administration of oxidized 13C-algal fatty acid with iodoacetamide, desferrioxamine, and glutathione peroxidase in the collection tubes exhibited the la-line signal due to the 13C (nuclear spin = 1) from 13C-labeled fatty acids (Fig. 6A). The simulation (Fig. 6B) to fit the experimental spectrum exhibited a composite of three species (Table I): 13C-labeled (4-POBN/13C-L’) (0.67 mol ratio) (Fig. 6C) and an unlabeled (4-POBN/‘R) (0.28 mol
in wvo
LOOH 4.POBN
-LOOH
-LOOH +linoleic
acid
collected
into water
, lOGauss+ FIG. 5. EPR spectra of radical adducts detected in bile from rats administered oxidized linoleic acid (LOOH) (520 mg/kg, ig) and 4-POBN (100 mg/kg, ip) with desferrioxamine (160 GM), iodoacetamide (20 mM), and glutathione peroxidase (150 units/ml) added to the bile collection tubes. (A) Complete system. (B) Omission of LOOH. (C) LOOH was replaced with linoleic acid (520 mg/kg). (D) Complete system but collected into deionized water. Spectrometer conditions were: modulation amplitude, 0.8 G; microwave power, 20 mW; time constant, 1.3 s; scan range, 50 G; scan time, 500 s averaged over two scans.
Jn Vim FATTY TABLE
ACID
FREE
I
Hyperfine Coupling Constants of Radical Adducts Obtained from Administration of Oxidized Fatty Acids coupling constants (G)
Hyperfine
Radical adduct DMPO/GS DMPO/GS 4-POBN/GS 4-POBN/GS 4-POBN/CH(OH)CH, 4-POBN/‘R 4-POBN/C 4-POBN/L 4-POBN/L 4-POBN/i3C-L
aN
4
15.2 15.4 15.34 15.0 15.50 15.87 15.80 15.63 15.38 15.64
16.3 16.2 2.62 2.3 2.50 2.58 2.56 2.66 2.50 2.73
Source
-
2.04
This Ref. This Ref. Ref. This Ref. Ref. Ref. This
work (19) work (23) (24) work (20) (25) (26) work
RADICAL
365
FORMATION
Spectroscopic assignment based on the comparison of hyperfine coupling constants alone can be misleading, especially when the assignments from the original reports were based upon experiments performed without employing any isotopes, e.g., 1702. As for the identity of fatty acid-derived carbon-centered radicals, 4-POBN radical adducts from two classes of such radicals have been reported: (i) the primary carbon-centered radical (al), where the adduct is formed from the fatty acid molecules by hydrogen abstraction, and (ii) secondary carbon-centered radicals (22, 25), where the alkoxyl fatty acid radicals have undergone P-scission or cyclization to the adjacent double bond to respectively generate alkyl and epoxy alkyl radicals. The hyperfine coupling constants (Fig. 5 and Table I) alone cannot distinguish the difference in the chemical structure of the carbon-centered radicals, and In “IV0
‘3C-algal
ratio) (Fig. 6D) radical adduct as well as the ascorbate semidone (0.05 mol ratio) (Fig. 6E). This result indicated that at least 67% of the total 4-POBN radical adducts was due to the trapping of the fatty acid-derived carboncentered radical(s) from the administered oxidized PUFA. The remaining 28% of the total radical adducts could be due to the trapping of carbon-centered radicals which did not originate from the administered oxidized PUFA or could be due to oxygen-centered radicals either from the oxidized PUFA or another source.
fatty acid hydroperoxlde
4-I’OBN
DISCUSSION Although it has been proposed by several studies on the cytotoxicity of fatty acid hydroperoxides that free radical formation is responsible for tissue damage, there has been only limited EPR evidence of free radical metabolites in uivo (12). In this paper, we have reported EPR evidence for fatty acid-derived radical adducts in bile obtained from anesthetized rats which had been administered oxidized fatty acids and a spin trap (Fig. 5, Table I). The radical adducts obtained from ex viva signals can be suppressed by adding iodoacetamide, desferrioxamine, and glutathione peroxidase to the collection tube to prevent chemical reactions during collection (Figs. 3 and 4). The assignment for fatty acid-derived carbon-centered radical adducts was confirmed by administration of oxidized i3C-labeled fatty acids to rats, where the hyperfine interaction from 13C was observed (Fig. 6). The hyperfine coupling constants for the fatty acidderived carbon-centered radical (Table I) are similar to those reported by Connor et al. (20). They reported that a fatty acid carbon-centered radical adduct of 4-POBN spectra exhibiting hyperfine coupling constants of aN = 15.8 G and ur = 2.6 G should be reassigned to a carboncentered radical adduct, not to an oxygen-centered radical adduct as previously assigned by Rosen et al. (14, 15).
2
Gauss
FIG. 6. (A) An EPR spectrum of radical adducts detected in bile from rats administered oxidized i3C-algal fatty acids (2.62 g/kg, ig) and 4POBN (500 mg/kg, ip) with desferrioxamine (160 PM), iodoacetamide (20 mM), and glutathione peroxidase (150 units/ml) added to the bile collection tubes. Spectrometer conditions were: modulation amplitude, 0.80 G; microwave power, 20 mW; time constant, 5 s; scan range, 55 G; scan time, 1800 s. (B) Composite simulated spectrum of spectra (C-E), where the hyperfine coupling constants of each radical adduct are tabulated in Table I. (C) Simulated spectrum of 4.POBN/W-L’: line width, 0.73 G; lineshape, 15% Lorentzian-85% Gaussian; mol ratio, 0.67. (D) Simulated spectrum of 4-POBNrR: line width, 0.70 G; lineshape, 16% Lorentzian-84% Gaussian; mol ratio, 0.28. (E) Simulated spectrum of ascorbate semidone radical: aH = 1.9 G; line width, 0.70 G; lineshape, 0% Lorentzian-100% Gaussian; mol ratio, 0.05.
366
CHAMULITRAT.
JORDAN.
high-pressure liquid chromatography/EPR/mass spectrometry must be done to accomplish this (22). The in U~UO spectrum Fig. 5A is probably of a mixture of primary and secondary oxidized PUFA-derived carbon-centered radicals and their corresponding oxygen-centered radicals. From our in vivo data from Fig. 6, we cannot identify which type of carbon-centered radicals were spin trapped because 13C-algal fatty acids used in our experiments were a mixture of several types of saturated and unsaturated fatty acids and were not specifically labeled. However, the observed 4-POBN radical adducts in viuo (Fig. 6), at least -70% of the total radical adducts, unequivocally result from the trapping of carbon-centered radicals from the administered oxidized PUFA. DMPO/GS’ may be formed in viuo as well, but we have no evidence of this, possibly due to the instability of this radical adduct. Oxidized fatty acids are absorbed and transported to the liver where the hydroperoxides will react with glutathione peroxidase and heme-containing proteins. Glutathione peroxidase reduces fatty acid hydroperoxides to their corresponding alcohol during the detoxification of autoxidized fatty acids (10). This reaction is a two-electron process utilizing GSH as a cofactor. Alternatively, the one-electron metabolism of fatty acid hydroperoxides by tissue iron, hematin, or hemoproteins will result in free radical metabolites (2,27). By these latter pathways, it is plausible that oxidized PUFA is metabolized in uiuo to free radical metabolites which are trapped by 4-POBN (Fig. 5 and Table I). As for experiments with DMPO (Fig. l), no evidence for in uiuo fatty acid-derived radical intermediates was found, i.e., the ex viuo reaction between LOOH and/or its metabolites and GSH forms GS’, which is subsequently trapped by DMPO. The biological reducing agents, ascorbic acid and GSH, serve as radical scavengers, which are oxidized to EPR detectable species (Figs. 5A and 5D). Although it has been reported that fatty acid hydroperoxides administered orally to rats exhibited acute toxicity on the intestines exemplified by hemorrhage and diarrhea and that they were easily decomposed or reduced in the digestive tract (28), our present data suggest that some administered fatty acid hydroperoxides survived such degradation. Both vegetable and animal oils rich in polyunsaturated fatty acids easily undergo oxidative deterioration during processing and cooking. Fatty acid hydroperoxides, products from oxidized oils, are known for their toxicity. Linoleic hydroperoxide has been shown to cause irreversible damage to porcine pulmonary artery endothelial cells in vitro (29). Atherosclerotic-like lesions have been observed using electron microscopy from tissues of rabbits injected with lipid hydroperoxide or oxidized cholesterol (30). Peroxidized polyunsaturated lipids cause the oxidation of benzo[a]pyrene leading to products which are mutagenic (31). In the context of cooked foods, pathological changes have been observed in the cell cultures incubated with heated fats (32) and in rats fed with deep fry fats (9, 33).
AND
MASON
It is concluded that oxidized lipid components interfere with the normal cell functions, resulting in pathological changes leading to cell injury. It was shown in vitro that biological membrane peroxidation leads to a molar ratio of l/15/50 for 4-hydroxynonenals/malondialdehyde/lipid peroxides (34). In addition, lipid hydroperoxides are more cytotoxic to human fibroblasts than 4-hydroxynonenals and much more than malondialdehyde (35). Although orally administered linoleic acid hydroperoxide and its secondary chemical products cause different responses in hepatic lipid metabolism in rats (36), fatty acid hydroperoxide seems to be the main component responsible for cell damage (34, 35). In this report, we have used the spin-trapping technique to show that oxidized fatty acids are metabolized in uiuo to fatty acid-derived carbon-centered radical(s). Since glutathione functions directly in the destruction of fatty acid hydroperoxides by glutathione peroxidase, further investigations are being carried out with our experimental protocols to obtain the effects of glutathione depleting agents. REFERENCES 1. Ames, B. N. (1983) Science 221, 125661264. 2. Kubow, S. (1992) Free Rod. Biol. Med. 12, 63-81. 3. Tappel, A. I,., and Dillard, C. J. (1981) Fed. Proc. 40, 174-178. 4. Dougherty, J. J., Croft, W. A., and Hoekstra, W. G. (1981) J. Nutr. 111, 1784-1796.
OF BIOCHEMISTRY
AND
BIOPHYSICS
Vol. 299, No. 2, December, pp. 361-367, 1992
Fatty Acid Radical Formation in Rats Administered Oxidized Fatty Acids: In Viva Spin Trapping Investigation Walee Chamulitrat,’
Sandra
Laboratory of Molecular Biophysics, P.0. Box 12233, Research Triangle
J. Jordan,
and Ronald
P. Mason
National Znstitute of Environmental Park, North Carolina 27709
Health
Sciences,
National
Institutes
of Health,
Received May 28, 1992, and in revised form August 19, 1992
We report in vivo evidence for fatty acid-derived free radical metabolite formation in bile of rats dosed with spin traps and oxidized polyunsaturated fatty acids (PUFA). When rats were dosed with the spin trap 5,5dimethyll-pyrroline N-oxide (DMPO) and oxidized PUFA, the DMPO thiyl radical adduct was formed due to a reaction between oxidized PUFA and/or its metabolites with biliary glutathione. In vitro experiments were performed to determine the conditions necessary for the elimination of radical adduct formation by ex vivo reactions. Fatty acid-derived radical adducts of cY-(4-pyridyl-1-oxide)-N-tert-butylnitrone (4-POBN) were detected in vivo in bile samples collected into a mixture of iodoacetamide, desferrioxamine, and glutathione peroxidase. Upon the administration of oxidized 13C-algal fatty acids and 4-POBN, the EPR spectrum of the radical adducts present in the bile exhibited hyperfine couplings due to ‘3C. Our data demonstrate that the carbon-centered radical adducts observed in in vivo experiments are unequivocally derived from oxidized PUFA. This in vivo evidence for PUFA-derived free radical formation supports the proposal that processes involving free radicals may be the molecular basis for the previously described cytotoxicity of dietary oxidized PUFA. B 1992 Academic
Press,
Inc.
Numerous studies have focused on high dietary fat intake and its association with chronic diseases in humans. Polyunsaturated fatty acids in foodstuffs are easily oxidized and the resultant lipid peroxidation yields a variety of mutagens, promoters, and carcinogens (1, 2). The involvement of lipid peroxidation in dietary fats has been demonstrated by the increased production of fragmented products from fatty acid hydroperoxides, i.e., pentane (3), ethane (4), urinary malondialdehyde (5), and low molec’ To whom correspondence
should be addressed. Fax: (919) 541.7880.
0003.9861/92 $6.00 Copyright fi 1992 hy Academic Press, All rights of reproduction in any form
Inc. reserved.
ular weight aldehydes (6). It is well known that fatty acid peroxidation and fatty acid hydroperoxide decomposition are free radical-mediated processes (I, 2). Furthermore, fatty acid-derived free radicals are thought to be intermediates that cause damage to biomembranes, proteins, and DNA (1, 2, 7, 8). Evidence for in uivo free radical formation from dietary oxidized fatty acids includes increases in exhaled pentane gas (3), TBARS (9), glutathione peroxidase activity (lo), and chemiluminescence (11). There has been only one in uivo report using the spin-trapping technique, where the EPR spectrum of fatty acid-derived radical adduct in the liver from rats dosed with oxidized fatty acids was reported (12). The spin trap, phenyl-N-tert-butylnitrone (PBN),2 and oxidized methyl linoleate were intraperitoneally injected (12). The EPR analysis was done on the Folch extract of the rat livers obtained 30 min after the injections, and the detected PBN radical adduct was assigned to the fatty acid alkoxyl radical adduct. In addition, the 4-POBN radical adduct obtained from a mixture of 4-POBN and liver microsomes from rats which had been dosed with oxidized methyl linoleate was assigned to the fatty acid peroxyl radical adduct (12). The assignments for the identity of these radical adducts were based upon the comparison of hyperfine coupling constants obtained from in vitro experiments reported earlier (13-15). In this report, we employed an in viuo spin-trapping technique that has been developed in our laboratory (16) to reinvestigate the in uiuo free radical formation from oxidized fatty acids. Radical adducts from bile samples detected from the anesthetized rats, which had been dosed with a spin trap and oxidized fatty acids, were successfully obtained and characterized. Upon the administration of oxidized ‘“C-algal fatty acids, the EPR spectrum of the ’ Abbreviations used: PUFA, polyunsaturated fatty acids; LOOH, fatty acid hydroperoxides; DMPO, 5,5-dimethyl-1-pyrroline N-oxide; PBN, phenyl-N-tert-butylnitrone; 4-POBN, n-(4.pyridyl-l-oxide)-N-tert-butylnitrone. 361
362
CHAMULITRAT,
JORDAN,
radical adduct present in the bile exhibited hyperfine couplings due to 13C.These findings indicate that the radical adducts observed in in viva experiments are indeed formed from carbon-centered radicals of the administered oxidized PUFA. EXPERIMENTAL Oleic acid, DMPO, 4-POBN, deferoxamine mesylate (desferrioxamine), iodoacetamide, and glutathione peroxidase (from bovine erythrocytes, 135 units/mg solid, 680 units/mg prot., Lot 70H9311) were purchased from Sigma. DMPO was vacuum distilled at room temperature and stored under nitrogen at -70°C prior to use. Linoleic acid, linolenic acid, and arachidonic acid (>99%) were purchased from Nu-chek Prep (Elysian, MN). 13C-algal fatty acid mixture (CLM-2063, 13C > 98%) containingpalmitic acid (16:O) (20%), palmitoleic acid (16:l) (7%), hexadecadienoic acid (16:2) (15%), oleic acid (181) (IS%), linoleic acid (l&2) (26%), and linolenic acid (183) (11%) was obtained from Cambridge Isotope, Inc. (Woburn, MA). Oxidized fatty acids were obtained by bubbling the neat fatty acids with oxygen at room temperature to form the corresponding fatty acid hydroperoxides (LOOH) and other oxidation products. The conjugated diene formation was measured with an HP 8451A spectrophotometer. The absorbance at 234 nm was monitored periodically until the conjugated diene concentration reached about 0.9-1.1 Musing an extinction coefficient of 2.3 X lo4 cm-’ Mm’ (17). By this preparation, the hydroperoxide value was maximized and only minimal breakdown of fatty acid hydroperoxide by oxygen occurred (18). The same oxidized PLJFA batch was used for a given set of experiments. Similarly, neat 13C-algal fatty acid mixture was oxidized by bubbling with oxygen until there was no increase in 234.nm absorbance. In viva studies. Male Sprague-Dawley rats (300-400 g) which were allowed free access to both food and water were used. Nonfasted rats were anesthetized with Nembutal, and their bile ducts were cannulated using 7- to lo-cm segments of PEIO tubing (16). Anesthesia was maintained during the surgery and bile collection. 4-PORN (100 mg/kg), which had been dissolved in deionized water, and neat DMPO (500 pl/ kg) were administered by intraperitoneal injections, and oxidized PUFA was injected directly into the stomach. Bile samples were collected into plastic Eppendorf tubes during successive 20-min intervals for 2 h. Bile samples were frozen immediately on dry ice and stored at -70°C. The EPR analysis was performed within 2 days. The signal intensity of radical adducts increased with the collection t,ime maximizing with the sixth bile sample (120-140 min aft.er the administration of an oxidized PUFA). We therefore chose these bile samples to represent our in uiuo experiments In vitro studies. Chemical reactions in bile can occur during bile collection, thereby causing ex viuo free radical formation. To distinguish e.r uiuo radical adducts from those of in uivo signals, in vitro experiments were performed to investigate the possible chemical reactions and their inhibition. In vitro reactions were initiated by the addition of an aliquot of oxidized PUFA stock solutions, which had been dissolved in bile, to an Eppendorf tube containing 4.POBN and bile (obtained from untreated rats). The compounds tested for inhibition effects were desfer rioxamine, iodoacetamide, and glutathione peroxidase. Individually and in combination, desferrioxamine, iodoacetamide, and glutathione peroxidase were added to the bile and spin-trap mixture, which was followed by addition of the oxidized PUFA. After mixing, the incubation mixtures were analyzed by EPR spectroscopy for radical adduct formation. EPR measurements. The EPR spectra were recorded at room temperature on either a Bruker ESPSOO or ER 200D spectrometer operated, respectively, at 9.87 and 9.77 GHz with a IOO-kHz modulation frequency. The samples were pipetted to a quartz flat cell, which was then centered in a TM,,, cavity. The data were transferred from the ESP1600 data system to an IBM/AT where the computer simulation analyses were done using the software developed at this laboratory by D. R. Duling.
AND
MASON
RESULTS
The bile from a rat that had been administered DMPO and oxidized linoleic acid (LOOH) gave a four-line EPR spectrum with hyperfine coupling constants of uN = 15.2 G and c$ = 16.3 G (Fig. 1A). This DMPO radical adduct was identified as DMPO/GS’by comparison with the literature values (19). The doublet with a 1.9-G hyperfine coupling constant at the center of the spectrum is due to the ascorbate semidione radical from the oxidation of ascorbate in the bile. The omission of oxidized linoleic acid resulted in a very weak signal (Fig. 1B). The replacement of oxidized linoleic acid with neat linoleic acid resulted in a weaker signal (Fig. 1C). To determine if this radical adduct was being formed in the collection tube during bile collection, the sulfhydryl blocking reagent iodoacetamide (1.6 mM final concentration) was added to the collection tube. Iodoacetamide indeed inhibited the formation of DMPO/GS’ (Fig. 1D). When sufficient iodoacetamide (2.5 mM final concentration) was added, the inhibition was total (data not shown), suggesting that the DMPO/GS formation was due to the ex vivo reactions of oxidized linoleic acid and/or its metabolites with biliary glutathione. The formation of DMPO/GS’ clearly was oxidized linoleic acid-dependent and may depend on catalysis by
LOOH
-LOOH
FIG. 1. EPR spectra of radical adducts detected in bile from rats administered oxidized linoleic acid (LOOH) (520 mg/kg, ig) and DMPO (500 pi/kg, ip). (A) Complete system. (B) Omission of LOOH. (C) LOOH was replaced with linoleic acid (520 mg/kg). (D) Complete system with iodoacetamide (1.6 mM, final concentration) added into the collection tubes. Spectrometer conditions were: modulation amplitude, 0.72 G; microwave power, 20 mW; time constant, 0.66 s; scan range, 80 G; scan time, 335 s.
In Viuo FATTY
ACID
FREE
biliary metals. Note that the weak signal observed in Fig. 1B indicates that there was an er uiuo oxidation of GSH independent of the administered oxidized PUFA. When DMPO was replaced by 4-POBN, a six-line 4POBN radical adduct was detected in bile samples collected 2 h after the injections (Fig. 2A). Administration of 4-POBN alone resulted in a much weaker signal (Fig. 2B). When oxidized arachidonic acid was replaced with neat fatty acid, only a weak signal was observed (Fig. 2C). These results, including the control experiments, were also reproducible for oxidized linoleic acid (520 mg/kg) or linolenic acid (390 mg/kg) (data not shown). In contrast, the administration of the oxygenated oleic acid (1.1 g/kg) did not result in radical adduct formation (data not shown). The in vitro investigation of the formation of free radicals that are trapped by 4-POBN in bile is shown in Fig. 3. Chemical reactions involving adventitious transition metals and GSH are plausible sources for this free radical formation in bile. It was observed that both desferrioxamine and iodoacetamide inhibited free radical formation (Figs. 3B and 3C), and when combined, desferrioxamine (Fig. 3D). and iodoacetamide caused -100% inhibition When the in vitro experiment was carried out in the presence of added oxidized linoleic acid. desferrioxamine and
RADICAL
363
FORMATION in vitro
FIG. 3. EPR spectra of radical adducts detected in bile obtained from untreated rats. (A) bile (86%) mixed with 4-POHN (20 mM). (B) As in (A) with the addition of desferrioxamine (160 PM). (C) As in (A) with the addition of iodoacetamide (20 mM). (D) As in (A) with the addition ofdesferrioxamine (160 fiM) and iodoacetamide (20 mM). Spectrometer conditions were: modulation amplitude, 0.7 G; microwave power, 20 mW; time constant, 1.3 s; scan range, 55 G; scan time, 500 s.
LOOH 4-PORN
B
ij
-LOO11
-LOOH +arach!donic
acid
FIG. 2. EPR spectra of’ radical adducts detected in bile from rats administered oxidized arachidonic acid (LOOH) (260 mg/kg, ig) and 4. PORN (100 n&kg, ip). (A) Complete system. (B) Omission of LOOH. (C) IJOOH was replaced with arachidonic acid (260 mg/kg). Spectrometer conditions were: modulation amplitude, 0.72 G; microwave power, 20 mW; time constant, 1.3 s; scan range, 80 G; scan time; 671 s averaged over two scans.
iodoacetamide together inhibited, but not totally (Figs. 4A and 4B). Results shown in Figs. 1A and 1B indicate that the oxidized linoleic acid reacted with the GSH in the bile sample resulting in the Gs’ formation; therefore, glutathione peroxidase was added to the collection tube to reduce biliary fatty acid hydroperoxides to their corresponding alcohols. The glutathione peroxidase concentration was increased until the radical adduct signal was totally inhibited. As a result, the addition of the three components, iodoacetamide, desferrioxamine, and glutathione peroxidase, completely inhibited the radical adduct formation (Fig. 4C). Once it was determined that the ex: uiuo radical adduct formation could be suppressed by the addition of iodoacetamide, desferrioxamine, and glutathione peroxidase to the collection tubes, the experiment performed in Fig. 2A was repeated (Fig. 5A). With this protocol, the radical adducts detected (Fig. 5A) presumably result from the trapping of the radical(s) which were formed and trapped by 4-POBN in. uiuo. We did not know the chemical structure of the trapped radical(s), but we thought that fatty acid-derived carbon-centered radical adducts would be the most likely because these have previously been detected in vitro with 4-POBN (20-22), although the possibility that the products were adduct of oxygen-centered radical(s), e.g., fatty acid-alkoxyl radicals, cannot be excluded. We therefore represent these radical adducts as 4-POBN/
364
CHAMULITRAT,
JORDAN,
in vim LOOH 4-POBN
-“collected into iodoacetamide desfetioxamine
collected into iodoacetamide desfenioxamine glutathione peroxidase
v
10 Gauss
FIG. 4. EPR spectra of radical adducts detected in bile obtained from untreated rats and oxidized linoleic acid (LOOH) in the presence of 4POBN. (A) bile (95%) mixed with LOOH (120 CM) and 4-POBN (20 mM). (B) As in (A) with prior addition of desferrioxamine (160 PM) and iodoacetamide (25 mM). (C) As in (B) with prior addition of glutathione peroxidase (150 units/ml). Spectrometer conditions were: modulation amplitude, 0.7 G; microwave power, 20 mW; time constant, 1.3 s; scan range, 50 G; scan time, 500 s.
‘R. The simulation of the spectrum in Fig. 5A yielded a composite of ascorbate semidione (0.26 mol ratio) and 4POBN/‘R (0.74 mol ratio) with the hyperfine coupling constants shown in Table I. The omission of oxidized linoleic acid resulted in no detectable radical adduct formation, but a higher concentration of the ascorbate semidione radical (Fig. 5B). When oxidized linoleic acid was replaced with neat linoleic acid, only a very weak spectrum was detected (Fig. 5C). When bile was collected into water with an equivalent volume to that of the treatment performed in Fig. 5A, the radical adduct concentration increased by -2.5-fold, demonstrating that radical adducts from ex uiuo reactions were the major contribution to the spectrum (Fig. 5D). This spectrum was simulated as a composite of three species: ascorbate semidione (0.05 mol ratio), 4-POBN/‘R (0.71 mol ratio), and 4-POBN/GS (0.24 mol ratio) (Table I). The formation of 4-POBN/ GS’ was clearly ex uiuo (Fig. 5D), since this signal was inhibited when GSH was depleted by iodoacetamide. Since 4-POBN/a-hydroxyalkyl and other radical adducts exhibit similar hyperfine coupling constants (Table I), our assignment for 4-POBN/GS’ is based on the iodoacetamide effect. Experiments performed in Figs. 3 and 4 were essential to ensure that iron-, GSH-, and hydroperoxide-dependent ex uiuo reactions do not occur during sample collection. These er uiuo chemical reactions could account for the
AND
MASON
majority of radical adducts formed during liver homogenization and/or organic extraction which were performed in the earlier spin-trapping reports of free radical formation by oxidized fatty acids (12, 13). When treatments to suppress ex uiuo free radical reactions are not used, radical adduct formed ex vivo may be mistaken as evidence for in uivo free radical formation. An in viuo experiment with oxidized i3C-algal fatty acid was performed to determine if the 4-POBN radical adduct was indeed from the carbon-centered radicals derived from the oxidized fatty acids. Since this sample contained only 75% unsaturated fatty acids, a higher concentration of 4POBN was used. The in uiuo 4-POBN radical adduct from the administration of oxidized 13C-algal fatty acid with iodoacetamide, desferrioxamine, and glutathione peroxidase in the collection tubes exhibited the la-line signal due to the 13C (nuclear spin = 1) from 13C-labeled fatty acids (Fig. 6A). The simulation (Fig. 6B) to fit the experimental spectrum exhibited a composite of three species (Table I): 13C-labeled (4-POBN/13C-L’) (0.67 mol ratio) (Fig. 6C) and an unlabeled (4-POBN/‘R) (0.28 mol
in wvo
LOOH 4.POBN
-LOOH
-LOOH +linoleic
acid
collected
into water
, lOGauss+ FIG. 5. EPR spectra of radical adducts detected in bile from rats administered oxidized linoleic acid (LOOH) (520 mg/kg, ig) and 4-POBN (100 mg/kg, ip) with desferrioxamine (160 GM), iodoacetamide (20 mM), and glutathione peroxidase (150 units/ml) added to the bile collection tubes. (A) Complete system. (B) Omission of LOOH. (C) LOOH was replaced with linoleic acid (520 mg/kg). (D) Complete system but collected into deionized water. Spectrometer conditions were: modulation amplitude, 0.8 G; microwave power, 20 mW; time constant, 1.3 s; scan range, 50 G; scan time, 500 s averaged over two scans.
Jn Vim FATTY TABLE
ACID
FREE
I
Hyperfine Coupling Constants of Radical Adducts Obtained from Administration of Oxidized Fatty Acids coupling constants (G)
Hyperfine
Radical adduct DMPO/GS DMPO/GS 4-POBN/GS 4-POBN/GS 4-POBN/CH(OH)CH, 4-POBN/‘R 4-POBN/C 4-POBN/L 4-POBN/L 4-POBN/i3C-L
aN
4
15.2 15.4 15.34 15.0 15.50 15.87 15.80 15.63 15.38 15.64
16.3 16.2 2.62 2.3 2.50 2.58 2.56 2.66 2.50 2.73
Source
-
2.04
This Ref. This Ref. Ref. This Ref. Ref. Ref. This
work (19) work (23) (24) work (20) (25) (26) work
RADICAL
365
FORMATION
Spectroscopic assignment based on the comparison of hyperfine coupling constants alone can be misleading, especially when the assignments from the original reports were based upon experiments performed without employing any isotopes, e.g., 1702. As for the identity of fatty acid-derived carbon-centered radicals, 4-POBN radical adducts from two classes of such radicals have been reported: (i) the primary carbon-centered radical (al), where the adduct is formed from the fatty acid molecules by hydrogen abstraction, and (ii) secondary carbon-centered radicals (22, 25), where the alkoxyl fatty acid radicals have undergone P-scission or cyclization to the adjacent double bond to respectively generate alkyl and epoxy alkyl radicals. The hyperfine coupling constants (Fig. 5 and Table I) alone cannot distinguish the difference in the chemical structure of the carbon-centered radicals, and In “IV0
‘3C-algal
ratio) (Fig. 6D) radical adduct as well as the ascorbate semidone (0.05 mol ratio) (Fig. 6E). This result indicated that at least 67% of the total 4-POBN radical adducts was due to the trapping of the fatty acid-derived carboncentered radical(s) from the administered oxidized PUFA. The remaining 28% of the total radical adducts could be due to the trapping of carbon-centered radicals which did not originate from the administered oxidized PUFA or could be due to oxygen-centered radicals either from the oxidized PUFA or another source.
fatty acid hydroperoxlde
4-I’OBN
DISCUSSION Although it has been proposed by several studies on the cytotoxicity of fatty acid hydroperoxides that free radical formation is responsible for tissue damage, there has been only limited EPR evidence of free radical metabolites in uivo (12). In this paper, we have reported EPR evidence for fatty acid-derived radical adducts in bile obtained from anesthetized rats which had been administered oxidized fatty acids and a spin trap (Fig. 5, Table I). The radical adducts obtained from ex viva signals can be suppressed by adding iodoacetamide, desferrioxamine, and glutathione peroxidase to the collection tube to prevent chemical reactions during collection (Figs. 3 and 4). The assignment for fatty acid-derived carbon-centered radical adducts was confirmed by administration of oxidized i3C-labeled fatty acids to rats, where the hyperfine interaction from 13C was observed (Fig. 6). The hyperfine coupling constants for the fatty acidderived carbon-centered radical (Table I) are similar to those reported by Connor et al. (20). They reported that a fatty acid carbon-centered radical adduct of 4-POBN spectra exhibiting hyperfine coupling constants of aN = 15.8 G and ur = 2.6 G should be reassigned to a carboncentered radical adduct, not to an oxygen-centered radical adduct as previously assigned by Rosen et al. (14, 15).
2
Gauss
FIG. 6. (A) An EPR spectrum of radical adducts detected in bile from rats administered oxidized i3C-algal fatty acids (2.62 g/kg, ig) and 4POBN (500 mg/kg, ip) with desferrioxamine (160 PM), iodoacetamide (20 mM), and glutathione peroxidase (150 units/ml) added to the bile collection tubes. Spectrometer conditions were: modulation amplitude, 0.80 G; microwave power, 20 mW; time constant, 5 s; scan range, 55 G; scan time, 1800 s. (B) Composite simulated spectrum of spectra (C-E), where the hyperfine coupling constants of each radical adduct are tabulated in Table I. (C) Simulated spectrum of 4.POBN/W-L’: line width, 0.73 G; lineshape, 15% Lorentzian-85% Gaussian; mol ratio, 0.67. (D) Simulated spectrum of 4-POBNrR: line width, 0.70 G; lineshape, 16% Lorentzian-84% Gaussian; mol ratio, 0.28. (E) Simulated spectrum of ascorbate semidone radical: aH = 1.9 G; line width, 0.70 G; lineshape, 0% Lorentzian-100% Gaussian; mol ratio, 0.05.
366
CHAMULITRAT.
JORDAN.
high-pressure liquid chromatography/EPR/mass spectrometry must be done to accomplish this (22). The in U~UO spectrum Fig. 5A is probably of a mixture of primary and secondary oxidized PUFA-derived carbon-centered radicals and their corresponding oxygen-centered radicals. From our in vivo data from Fig. 6, we cannot identify which type of carbon-centered radicals were spin trapped because 13C-algal fatty acids used in our experiments were a mixture of several types of saturated and unsaturated fatty acids and were not specifically labeled. However, the observed 4-POBN radical adducts in viuo (Fig. 6), at least -70% of the total radical adducts, unequivocally result from the trapping of carbon-centered radicals from the administered oxidized PUFA. DMPO/GS’ may be formed in viuo as well, but we have no evidence of this, possibly due to the instability of this radical adduct. Oxidized fatty acids are absorbed and transported to the liver where the hydroperoxides will react with glutathione peroxidase and heme-containing proteins. Glutathione peroxidase reduces fatty acid hydroperoxides to their corresponding alcohol during the detoxification of autoxidized fatty acids (10). This reaction is a two-electron process utilizing GSH as a cofactor. Alternatively, the one-electron metabolism of fatty acid hydroperoxides by tissue iron, hematin, or hemoproteins will result in free radical metabolites (2,27). By these latter pathways, it is plausible that oxidized PUFA is metabolized in uiuo to free radical metabolites which are trapped by 4-POBN (Fig. 5 and Table I). As for experiments with DMPO (Fig. l), no evidence for in uiuo fatty acid-derived radical intermediates was found, i.e., the ex viuo reaction between LOOH and/or its metabolites and GSH forms GS’, which is subsequently trapped by DMPO. The biological reducing agents, ascorbic acid and GSH, serve as radical scavengers, which are oxidized to EPR detectable species (Figs. 5A and 5D). Although it has been reported that fatty acid hydroperoxides administered orally to rats exhibited acute toxicity on the intestines exemplified by hemorrhage and diarrhea and that they were easily decomposed or reduced in the digestive tract (28), our present data suggest that some administered fatty acid hydroperoxides survived such degradation. Both vegetable and animal oils rich in polyunsaturated fatty acids easily undergo oxidative deterioration during processing and cooking. Fatty acid hydroperoxides, products from oxidized oils, are known for their toxicity. Linoleic hydroperoxide has been shown to cause irreversible damage to porcine pulmonary artery endothelial cells in vitro (29). Atherosclerotic-like lesions have been observed using electron microscopy from tissues of rabbits injected with lipid hydroperoxide or oxidized cholesterol (30). Peroxidized polyunsaturated lipids cause the oxidation of benzo[a]pyrene leading to products which are mutagenic (31). In the context of cooked foods, pathological changes have been observed in the cell cultures incubated with heated fats (32) and in rats fed with deep fry fats (9, 33).
AND
MASON
It is concluded that oxidized lipid components interfere with the normal cell functions, resulting in pathological changes leading to cell injury. It was shown in vitro that biological membrane peroxidation leads to a molar ratio of l/15/50 for 4-hydroxynonenals/malondialdehyde/lipid peroxides (34). In addition, lipid hydroperoxides are more cytotoxic to human fibroblasts than 4-hydroxynonenals and much more than malondialdehyde (35). Although orally administered linoleic acid hydroperoxide and its secondary chemical products cause different responses in hepatic lipid metabolism in rats (36), fatty acid hydroperoxide seems to be the main component responsible for cell damage (34, 35). In this report, we have used the spin-trapping technique to show that oxidized fatty acids are metabolized in uiuo to fatty acid-derived carbon-centered radical(s). Since glutathione functions directly in the destruction of fatty acid hydroperoxides by glutathione peroxidase, further investigations are being carried out with our experimental protocols to obtain the effects of glutathione depleting agents. REFERENCES 1. Ames, B. N. (1983) Science 221, 125661264. 2. Kubow, S. (1992) Free Rod. Biol. Med. 12, 63-81. 3. Tappel, A. I,., and Dillard, C. J. (1981) Fed. Proc. 40, 174-178. 4. Dougherty, J. J., Croft, W. A., and Hoekstra, W. G. (1981) J. Nutr. 111, 1784-1796.
