Life Sciences, Vol. 50, pp. 875-882 Printed in the USA
Pergamon Press
IN VITRO LIPID PEROXIDATION OF HUMAN SERUM CATALYZED BY CUPRIC ION: ANTIOXlDANT RATHER THAN PROOXlDANT ROLE OF ASCORBATE Amitava Dasgupta and Tereslta Zdunek Department of Pathology, The University of Chicago Pritzker School of Medicine (Received in final form January 17, 1992)
The dual role of ascorbate as an antioxidant and a prooxidant has been clearly documented in the literature. Ascorbate acts as an antioxidant by protecting human serum from lipid peroxidation induced by azo dye-generated free radicals. On the other hand, ascorbate is readily oxidized in the presence of transition metal ions, (especially cupric ion) and accelerates lipid peroxidation in tissue homogenates by producing free radicals. Interestingly, we observed an antioxidant rather than an expected prooxidant role of ascorbate when human serum supplemented with 1.2mmol/L ascorbate underwent lipid peroxidations initiated by 2mmol/L copper sulfate. The antioxidant role of ascorbate was confirmed by studying the conventional thiobarbituric acid reactive substances (TBARS) as well as by observing the protective effect of ascorbate on the copper-induced peroxidation of unsaturated and polyunsaturated fatty acids. The antioxidation protection provided by ascorbate was comparable to that of equimolar (x-tocopherol when incubated for 24h. However, lipid peroxidation products were lower in serum supplemented with r,-tocopherol after 48h of incubation. This effect may be attributed to the binding of copper by serum proteins, thus preventing direct interaction between cupric ions and ascorbate. This proposed mechanism is based on the observation that the concentration of ascorbate decreased more slowly in serum than in phosphate buffer at physiological pH. Our results also indicate an outstanding anti-oxidant property of human serum due to the chelation of transition metal ions (even at high concentrations) by various serum proteins. While oxygen is essential to life, it can be potentially toxic to living organisms due to its participation in the formation of free radicals. Free radicals can damage all biological compounds; lipide, proteins, nucleic acids, carbohydrates and the macromolecules of connective tissues. Recently, free radical-mediated lipid peroxidations and high concentrations of lipid peroxidation products have been reported in several human diseases: atherosclerosis (1), multiple sclerosis (2), diabetes (3), cancer and aging (4). Knight et al also reported increased urinary lipoperoxides in drug abusers (5). It has been proposed that eliminating lipid hydroperoxides in plasma before they are taken up by peripheral tissue may improve the prognosis of such diseases (6). Ames and his coinvestigators described an important antioxidant role for ascorbate in human plasma when free radicals were generated by the decomposition of azo dyes (6,7). On the other hand, the literature reports clearly indicate that ascorbate can generate free radicals in the presence of transition metal ions, especially cupric ions, due to autooxidation. This prooxidant activity results in significant cytotoxicity (8-10). It is also known that ascorbate Address correspondence to Dr. Amitava Dasgupta, Clinical Laboratories, Box 146, The University of Chicago Hospitals, Chicago, IL 60637 0024-3205/92 $5.00 + .00 Copyright © 1992 Pergamon Press plc All rights reserved.
876
Lipid Peroxidation
and Ascorbate
Vol.
50, No.
12, 1992
can stimulate collagen synthesis in cultured fibroblast cells due to increased lipid peroxidation and that such effect can be inhibited by antioxidants (11). High concentrations of ascorbate also cause genetic damage. This effect is enhanced in the presence of cupric ions due to catalytic generation of free radicals (12). The U.S. recommended daily allowance (RDA) for ascorbate is based solely on its effect on collagen synthesis (13). The reluctance to raise the RDA for ascorbate is due to the reported prooxidation properties of ascorbate in tissue homogenates when catalyzed by transition metal ions (7), In order to understand the dual role of ascorbic acid which seems to be dependent on concentration and the presence or absence of high concentrations of transition metal ions, we studied the in vitro effects of high concentrations of ascorbate on cupric ion catalyzed lipid peroxidation of human serum. The physiological range of ascorbate in human serum is 0.034 0.11mmol/L, while that of copper is 3.14 - 10.99 Izmol/L. Subbiah et al recently described a protocol to test susceptibility of peroxidation of lipoproteins in human plasma and used cupric ions to initiate lipid peroxidation (29,30). Piriou and his coinvestigators also used high concentrations of copper sulfate (up to 2mmol/L) to initiate lipid peroxidation of human plasma in order to study antioxidant effect of hemolysate (17). The concentration of copper sulfate we selected for our study was also 2mmol/L. We decided to study copper-ascorbate system instead of the iron-ascorbate system because Winkler indicated that ascorbic acid oxidation is principally a cupric ion-dependent process at physiological pH (23). Now we would like to report an antioxidant role rather than the expected prooxidant effect of ascorbate in the cupric ion-catalyzed lipid peroxidation of human sera. MATERIALS AND METHODS
c~-Tocopherol, ascorbic acid, cupric sulfate and 2-thiobarbituric acid were purchased from Sigma Chemical Co (St. Louis, MO). Phosphotungstic acid was obtained from Merck and Company (Rahway, NJ), glacial acetic acid from Mallinckrodt (Paris, KY) and malonaldehyde bis(dimethyl acetal) from Aldrich (Milwaukee, WI). Spectroscopic measurements were done using a Digispec-Spectrcphotometer (Helena Laboratories, Beaumont, TX). Atomic absoprtion measurements were carried out using a Perkin-Elmer Model 603, Atomic Absorption Spectrophotometer and nephelometric assays were performed on an Array Protein System Analyzer (Beckmann Instruments, Brea, CA). An Ektachem 700 Analyzer (Eastman Kodak, Rochester, NY) was used for albumin and total protein determinations. The HPLC system used for the assay of ascorbate included a WISP (Waters Intelligent Sample Processor), Waters Model 510 pump, Waters Model 441 Absorbance Detector and Waters Data Module 710 (Waters Corporation, Medford, MA ). The Gas Chromatography/Mass Spectrometric analysis of fatty acid methyl esters were done by a Model 5890 Gas Chromatograph and Model 5970 Mass Selective Detector (Hewlett Packard, Palo Alto, CA). Stock solutions of cupric sulfate (44.4 mmol/L) and ascorbic acid (26.6 mmol/L) were prepared in HPLC grade water while that of (x-tocopherol (26.6 mmol/L) was prepared in absolute ethanol. Aliquots from serum pooled from normal volunteers were supplemented with ascorbate and copper sulfate, or c¢-tocopherol and copper sulfate, or copper sulfate alone. We added 150 ILL of stock solutions containing the chemicals to 3mL aliquots of pooled serum. Volume differences were supplemented by 0.9% saline solution. The final concentration of copper sulfate was 2 mmol/L, while that of ascorbate or c¢-tocopherol was 1.2 mmol/L. The samples were incubated at 37 ° C and at specific times a small portion (800 p.L) was removed for analysis. Four different serum pools were studied to verify the reproducibility of our results. The quantity of lipid peroxidation products in serum was determined as described by Yagi with minor modifications (14). Briefly, lipoproteins were coprecipitated with other serum proteins using phosphotungstic and sulfuric acids prior to assay in order to remove soluble interferences present in serum. Then the precipitate was allowed to react with 2-thiobarbituric acid in glacial acetic acid in the presence of butylated hydroxytoluene (BHT). BHT was added to the reaction mixture to prevent further formation of lipid peroxides during incubation, thus allowing only those lipid hydroperoxides already formed to decompose to malonaldehyde. In the
Vol. 50, No. 12, 1992
Lipid Peroxidation and Ascorbate
877
final reaction, malonaldehyde reacts with 2-thiobarbituric acid. Following incubation, chromogens were extracted into n-butanol and absorption was measured at 530nm. Malonaldehyde bis-(dimethyl acetal) was used as a standard and the lipid peroxidation products expressed as pmol/L of maionaldehyde equivalents. The total serum lipids were extracted with chloroform/methanol. Fatty acid methyl esters were prepared from lipids by trans-esterification using 14% boron trifluoride in methanol and were purified by silica gel column chromatography prior to analysis as described earlier (15). The concentration of ascorbic acid in serum or in phosphate buffer (pH 7.4) was measured with HPLC using an octadecyltrichlorosilane column (C18 colum, Waters Corporation) and a mobile phase containing 8 mmol/L phosphate buffer at pH 5.3 The elution of peaks was monitored at 254 nm using a U.V. detector (16). Serum albumin concentrations were measured using chemistry slides containing bromocresol green. Total protein determinations were performed with chemistry slides using the biuret reaction. Both assays were done on an Ektachem 700 Analyzer which operates on the principle of reflectance photometry. Transferrin and ceruloplasmin were determined using nephelometric immunoassay techniques (Beckman, Brea, CA). Protein-free ultrafiltrates for measurement of non-protein bound copper were prepared from serum using the Amicon Micropartition System (Danvers, MA). The Centrifree Micropartition System contains an anisotropic hydrophilic YMT membrane with a molecular weight cut off of 30,000. The partition efficiency of the YMT membrane had been demonstrated by retention of >99.9% serum proteins and < 5% of L-thyroxine (31). The sera were centrifuged at 1162Xg for 30min at 25 ° C and protein free ultrafiltrates were analyzed for copper using atomic absorption spectrophotometry. RESULTS AND DISCUSSION
LiDid Peroxidation in Human Serum The concentrations of albumin, total proteins, transferrin and ceruloplasmin in all serum pools we studied were within normal reference range. Lipid peroxidation in serum was initiated with 2 mmol/L cupric sulfate. Under such strong oxidative stress, thiobarbituric acid reactive substances (TBARS) in serum were increased 500-630% over controls following 48h of incubation at 37 ° C. On the other hand, with sera incubated without cupric ions, TBARS increased by only 10-20%, indicating a good antioxidant property of human serum. Serum incubated with 1.2 mmol/L ascorbate or ¢¢-tocopherol, in the absence of cupric ion, resisted any further lipid peroxidation as demonstrated by marginal increase in TBARS. More interestingly, serum incubated with cupric ions in the presence of high concentration of ascorbate (10 times higher than the upper limit of physiological range), demonstrated an antioxidant role of ascorbate instead of its expected prcoxidant effect. After 6h of incubation, we observed little increase in TBARS in serum supplemented with ascorbate whereas TBARS increased 40-80% in samples incubated with copper sulfate only. The concentrations of TBARS were significantly lower by independent t-test, two-tailed, in sera supplemented with ascorbate and copper sulfate compared to sera supplemented with copper sulfate alone. After 24h of incubation, TBARS increased 200-330% in sera supplemented with copper sulfate alone as compared to controls. The increase of TBARS was significantly lower (80-160%) in sera supplemented with both copper sulfate and ascorbate, indicating a substantial antioxidant effect of ascorbate even in the presence of high concentrations of cupric ions (Table 1). Also ~-tocopheroi, a well known antioxidant, provided protection similar to that of ascorbate at an equimolar concentration for the first 24h of incubation. At the end of a 48h of incubation, TBARS were still significantly lower (by independent t-test, two-tailed) in serum supplemented with ascorbate and even more in sera supplemented with a-tocopherol.
878
Lipid Peroxidation and Ascorbate
Vol. 50, No. 12, 1992
Unsaturated Fattv Acid Profile in Serum: Protective Role of Ascorbate The antioxidant effects of ascorbate were further studied in 3rd and 4th serum pools by analyzing the fatty acid profiles. Usually the proportion of unsaturated to saturated fatty acids decreases following lipid peroxidation (17-18). The major unsaturated fatty acids we identified
Thiobarbituric
TABLE 1 acid reactive substances (TBARS) in Serum Pools 0 h
Serum Pool 1 Mean(SD) ^ Serum 2.2(0.10) Serum+ Cu ++ * 2.3(0.20) Serum+ Cu ++ + Ascorbate ** 2.0(0.10) Serum + Cu ++ +o¢-Tocopherol # 2.1(0.15) Serum Pool 2 Serum Serum + Cu ++ Serum + Cu ++ + Ascorbate Serum + Cu ++ + c¢-Tocopherol Serum Pool 3 Serum Serum + Cu ++ Serum + Cu ++ + Ascorbate Serum + Cu ++ + ¢¢- Tocopherol Serum Pool 4 Serum Serum + Cu ++ Serum + Cu ++ + Ascorbate Serum +Cu ++ +¢¢-Tocopherol
48h
6 h 24 h wnol/L malonaldehyde equivalent
2.2(0.14) 3.3(0.12)
2.4(0.14) 8.1(0.10)
2.7(0.10) 15.9(0.23)
2.3(0.18)
5.1(0.09)
13.1(0.37)
2.4(0.11)
5.9(0.10)
10.2(0.56)
3.0(0.13) 2.8(0.20)
2.9(0.11) 4.2(0.18)
3.2(0.10) 9.7(0.15)
3.2(0.11) 16.5(0.90)
2.8(0.15)
3.1(0.10)
5.9(0.16)
11.9(0.40)
3.0(0.15)
2.8(0.18)
5.5(0.05)
8,9(0.50)
2.6(0.10) 2.7(0.15)
2.8(0.20) 3,9(0.12)
2.8(0.10) 8.7(0.11)
2.9(0.20) 14.2(0.12)
2.6(0.20)
3.1 (0.07)
5.6(0.19)
12.7(0.25)
2.5(0.12)
3.3(0.08)
5.2(0.05)
11.6(0.25)
2.6(0.12) 2.5(0.12)
2.5(0.10) 4.5(0.20)
2.6(0.15) 10.8(0.23)
2.8(0.13) 18.4(0.41)
2.4(0.20)
3.2(0.12)
6.3(0.15)
13.2(0.16)
2.5(0.14)
3.5(0.07)
5.6(0.10)
9.8(0.07)
*Cu 2+ : 2.0 mmol/L, ** Ascorbate: 1.2 mmol/L, # ¢¢-tocopheroh 1.2 mmol/L, ^ Mean of three replicates were palmitoleic (16:1), linoleic (18:2), oleic (18:1), arachidonic (20:4) and decosahexaenoic (22:6). The major saturated fatty acids were myristic (14:0), palmitic (16:0) and stearic (18:0). These fatty acids were identified based on their mass spectral characteristics and retention times.
Vol. 50, No. 12, 1992
Lipid Peroxldation and Ascorbate
879
Little change was observed in the relative composition of fatty acids in serum incubated in the absence of copper, demonstrating the antioxidant properties of normal human serum. In serum pool 3, the proportion of unsaturated fatty acids dropped from 72.8% to 69.8°/. following 48h of incubation at 37o C whereas in the presence of cupric ions, the proportion of unsaturated fatty acids dropped from 71.3% to 64.8% and then to 57.9% following 24 and 48h of incubation. As expected, the composition of polyunsaturated fatty acids (20:4, 20:3 and 22:6) dropped from an initial value of 16.2% to 10.1% and finally to 6.9%, indicating that polyunsaturated fatty acids are most susceptable to oxidative damage. However, in the presence of ascorbate, after 24h of incubation, the proportion of unsaturated fatty acids decreased from an initial value of 72.0% to only 70.7%. This indicates a substantial antioxidant defense provided by ascorbate against copper-induced lipid peroxidation in human serum. The relative proportion of polyunsaturated fatty acids dropped from 16.4% to only 14.4%. Interestingly, (z-tocopherol provided similar protection as the composition of unsaturated fatty acids dropped from 71.6% to 70.4% (Table 2). TABLE II Effect of peroxldation on fatty acid profile in serum pool 3. FATrY ACIDS % Saturated (14:0, 16:0, 18:0)
% Unsaturated (18:1, 18:2, 20:3, 20:4.,22:6)
%Polyunsaturated only (20:3, 20:4, 22:6)
Serum, Oh
27.8
72.8
16.5
Serum, 24h
29.0
71.0
15.9
Serum, 48h
30.8
69.8
15.7
Serum + Copper, Oh
28.7
71.3
16.2
Serum + Copper, 24h
35,2
64.8
10.1
Serum +Copper, 48h
41.2
57.9
6.9
Serum + Copper + Ascorbate, Oh
28.0
72.0
16.4
Serum + Copper + Ascorbate, 24h
29.4
70.7
14.4
Serum + Copper+ Ascorbate, 48h
39.3
60.6
8.3
Serum + Copper+ u-Tocopherol, Oh
28.4
71.6
16.1
Serum + Copper + a-Tocopherol, 24h
29.6
70.4
13.4
Serum + Copper + u-Tocopherol, 48h
34.6
65.4
9.9
880
Lipid Peroxidation
and Ascorbate
Vol.
50, No.
12, 1992
After 48h of incubation the proportion of unsaturated fatty acids were higher in serum supplemented with ~-tocopherol (65.4%) compared to ascorbate (60.6%). However, both were higher than that observed in serum containing only copper (57.9%). This observation indicates that some antioxidant protection due to ascorbate is still present following such a long incubation time and under strong oxidative stress. Mechanism of Antioxidant DroDertv of Ascorbate It has been generally accepted that ascorbate can act as an antioxidant by scavenging free radicals as it undergoes oxidation to dehydroascorbic acid. In the presence of transition metal ions, however, ascorbate is oxidized very rapidly and generates free radicals rather than scavenging free radicals (19). Auto-oxidation of ascorbate in the presence of copper generates hydrogen peroxide which subsequently reacts with both ascorbate and dehydroascorbate, generating free radicals. Recently, Smith et al reported that 2-imidazolethiones protect ascorbic acid from oxidation induced by copper. Such protective effects were ascribed to the formation of copper complexes (20). The oxidation of ascorbate can also be prevented by chelating agents such as EDTA, cystein, purines, and bathocuproine (21-23). Chelation of metal ions due to protein binding can also prevent auto-oxidation of ascorbate. For example, ceruloplasmin prevents the copper catalyzed oxidation of ascorbate (20). We studied the concentration-dependent serum protein binding of copper by supplementing normal serum pools with 0.1 mmol/L to 2 mmol/L cupric ions. The concentration of copper in protein-free ultrafiltrates (representing non-protein bound copper) is very low when sera were supplemented with up to 0.25 mmol/L copper (ten times higher than the upper limit of the physiological range). Surprisingly, the concentration of copper in the ultrafiltrate was only 0.33 mmoVL even when the serum was supplemented with 2 mmol/L copper (Fig.l), indicating that 84.5% of the copper was still bound to serum proteins. Moreover, uric acid, present in serum may provide additional antioxidant protection because uric acid protects polyunsaturated fatty acids against copper-induced peroxidation (24). We also studied the decay of supplemented ascorbate in human serum. When ascorbate was added to the serum in absence of copper and incubated at 37° C, 69.2% of original ascorbate concentration was present after 24h, and 28.6% after 48h. Phosphate buffer, with equimolar ascorbate showed almost a complete decay of ascorbate in 24h, indicating again that serum prevents oxidation of ascorbate. When ascorbate in phosphate buffer was incubated in presence of 2 mmol/L copper, the expected rapid decrease of ascorbate was observed with almost no ascorbate (
Pergamon Press
IN VITRO LIPID PEROXIDATION OF HUMAN SERUM CATALYZED BY CUPRIC ION: ANTIOXlDANT RATHER THAN PROOXlDANT ROLE OF ASCORBATE Amitava Dasgupta and Tereslta Zdunek Department of Pathology, The University of Chicago Pritzker School of Medicine (Received in final form January 17, 1992)
The dual role of ascorbate as an antioxidant and a prooxidant has been clearly documented in the literature. Ascorbate acts as an antioxidant by protecting human serum from lipid peroxidation induced by azo dye-generated free radicals. On the other hand, ascorbate is readily oxidized in the presence of transition metal ions, (especially cupric ion) and accelerates lipid peroxidation in tissue homogenates by producing free radicals. Interestingly, we observed an antioxidant rather than an expected prooxidant role of ascorbate when human serum supplemented with 1.2mmol/L ascorbate underwent lipid peroxidations initiated by 2mmol/L copper sulfate. The antioxidant role of ascorbate was confirmed by studying the conventional thiobarbituric acid reactive substances (TBARS) as well as by observing the protective effect of ascorbate on the copper-induced peroxidation of unsaturated and polyunsaturated fatty acids. The antioxidation protection provided by ascorbate was comparable to that of equimolar (x-tocopherol when incubated for 24h. However, lipid peroxidation products were lower in serum supplemented with r,-tocopherol after 48h of incubation. This effect may be attributed to the binding of copper by serum proteins, thus preventing direct interaction between cupric ions and ascorbate. This proposed mechanism is based on the observation that the concentration of ascorbate decreased more slowly in serum than in phosphate buffer at physiological pH. Our results also indicate an outstanding anti-oxidant property of human serum due to the chelation of transition metal ions (even at high concentrations) by various serum proteins. While oxygen is essential to life, it can be potentially toxic to living organisms due to its participation in the formation of free radicals. Free radicals can damage all biological compounds; lipide, proteins, nucleic acids, carbohydrates and the macromolecules of connective tissues. Recently, free radical-mediated lipid peroxidations and high concentrations of lipid peroxidation products have been reported in several human diseases: atherosclerosis (1), multiple sclerosis (2), diabetes (3), cancer and aging (4). Knight et al also reported increased urinary lipoperoxides in drug abusers (5). It has been proposed that eliminating lipid hydroperoxides in plasma before they are taken up by peripheral tissue may improve the prognosis of such diseases (6). Ames and his coinvestigators described an important antioxidant role for ascorbate in human plasma when free radicals were generated by the decomposition of azo dyes (6,7). On the other hand, the literature reports clearly indicate that ascorbate can generate free radicals in the presence of transition metal ions, especially cupric ions, due to autooxidation. This prooxidant activity results in significant cytotoxicity (8-10). It is also known that ascorbate Address correspondence to Dr. Amitava Dasgupta, Clinical Laboratories, Box 146, The University of Chicago Hospitals, Chicago, IL 60637 0024-3205/92 $5.00 + .00 Copyright © 1992 Pergamon Press plc All rights reserved.
876
Lipid Peroxidation
and Ascorbate
Vol.
50, No.
12, 1992
can stimulate collagen synthesis in cultured fibroblast cells due to increased lipid peroxidation and that such effect can be inhibited by antioxidants (11). High concentrations of ascorbate also cause genetic damage. This effect is enhanced in the presence of cupric ions due to catalytic generation of free radicals (12). The U.S. recommended daily allowance (RDA) for ascorbate is based solely on its effect on collagen synthesis (13). The reluctance to raise the RDA for ascorbate is due to the reported prooxidation properties of ascorbate in tissue homogenates when catalyzed by transition metal ions (7), In order to understand the dual role of ascorbic acid which seems to be dependent on concentration and the presence or absence of high concentrations of transition metal ions, we studied the in vitro effects of high concentrations of ascorbate on cupric ion catalyzed lipid peroxidation of human serum. The physiological range of ascorbate in human serum is 0.034 0.11mmol/L, while that of copper is 3.14 - 10.99 Izmol/L. Subbiah et al recently described a protocol to test susceptibility of peroxidation of lipoproteins in human plasma and used cupric ions to initiate lipid peroxidation (29,30). Piriou and his coinvestigators also used high concentrations of copper sulfate (up to 2mmol/L) to initiate lipid peroxidation of human plasma in order to study antioxidant effect of hemolysate (17). The concentration of copper sulfate we selected for our study was also 2mmol/L. We decided to study copper-ascorbate system instead of the iron-ascorbate system because Winkler indicated that ascorbic acid oxidation is principally a cupric ion-dependent process at physiological pH (23). Now we would like to report an antioxidant role rather than the expected prooxidant effect of ascorbate in the cupric ion-catalyzed lipid peroxidation of human sera. MATERIALS AND METHODS
c~-Tocopherol, ascorbic acid, cupric sulfate and 2-thiobarbituric acid were purchased from Sigma Chemical Co (St. Louis, MO). Phosphotungstic acid was obtained from Merck and Company (Rahway, NJ), glacial acetic acid from Mallinckrodt (Paris, KY) and malonaldehyde bis(dimethyl acetal) from Aldrich (Milwaukee, WI). Spectroscopic measurements were done using a Digispec-Spectrcphotometer (Helena Laboratories, Beaumont, TX). Atomic absoprtion measurements were carried out using a Perkin-Elmer Model 603, Atomic Absorption Spectrophotometer and nephelometric assays were performed on an Array Protein System Analyzer (Beckmann Instruments, Brea, CA). An Ektachem 700 Analyzer (Eastman Kodak, Rochester, NY) was used for albumin and total protein determinations. The HPLC system used for the assay of ascorbate included a WISP (Waters Intelligent Sample Processor), Waters Model 510 pump, Waters Model 441 Absorbance Detector and Waters Data Module 710 (Waters Corporation, Medford, MA ). The Gas Chromatography/Mass Spectrometric analysis of fatty acid methyl esters were done by a Model 5890 Gas Chromatograph and Model 5970 Mass Selective Detector (Hewlett Packard, Palo Alto, CA). Stock solutions of cupric sulfate (44.4 mmol/L) and ascorbic acid (26.6 mmol/L) were prepared in HPLC grade water while that of (x-tocopherol (26.6 mmol/L) was prepared in absolute ethanol. Aliquots from serum pooled from normal volunteers were supplemented with ascorbate and copper sulfate, or c¢-tocopherol and copper sulfate, or copper sulfate alone. We added 150 ILL of stock solutions containing the chemicals to 3mL aliquots of pooled serum. Volume differences were supplemented by 0.9% saline solution. The final concentration of copper sulfate was 2 mmol/L, while that of ascorbate or c¢-tocopherol was 1.2 mmol/L. The samples were incubated at 37 ° C and at specific times a small portion (800 p.L) was removed for analysis. Four different serum pools were studied to verify the reproducibility of our results. The quantity of lipid peroxidation products in serum was determined as described by Yagi with minor modifications (14). Briefly, lipoproteins were coprecipitated with other serum proteins using phosphotungstic and sulfuric acids prior to assay in order to remove soluble interferences present in serum. Then the precipitate was allowed to react with 2-thiobarbituric acid in glacial acetic acid in the presence of butylated hydroxytoluene (BHT). BHT was added to the reaction mixture to prevent further formation of lipid peroxides during incubation, thus allowing only those lipid hydroperoxides already formed to decompose to malonaldehyde. In the
Vol. 50, No. 12, 1992
Lipid Peroxidation and Ascorbate
877
final reaction, malonaldehyde reacts with 2-thiobarbituric acid. Following incubation, chromogens were extracted into n-butanol and absorption was measured at 530nm. Malonaldehyde bis-(dimethyl acetal) was used as a standard and the lipid peroxidation products expressed as pmol/L of maionaldehyde equivalents. The total serum lipids were extracted with chloroform/methanol. Fatty acid methyl esters were prepared from lipids by trans-esterification using 14% boron trifluoride in methanol and were purified by silica gel column chromatography prior to analysis as described earlier (15). The concentration of ascorbic acid in serum or in phosphate buffer (pH 7.4) was measured with HPLC using an octadecyltrichlorosilane column (C18 colum, Waters Corporation) and a mobile phase containing 8 mmol/L phosphate buffer at pH 5.3 The elution of peaks was monitored at 254 nm using a U.V. detector (16). Serum albumin concentrations were measured using chemistry slides containing bromocresol green. Total protein determinations were performed with chemistry slides using the biuret reaction. Both assays were done on an Ektachem 700 Analyzer which operates on the principle of reflectance photometry. Transferrin and ceruloplasmin were determined using nephelometric immunoassay techniques (Beckman, Brea, CA). Protein-free ultrafiltrates for measurement of non-protein bound copper were prepared from serum using the Amicon Micropartition System (Danvers, MA). The Centrifree Micropartition System contains an anisotropic hydrophilic YMT membrane with a molecular weight cut off of 30,000. The partition efficiency of the YMT membrane had been demonstrated by retention of >99.9% serum proteins and < 5% of L-thyroxine (31). The sera were centrifuged at 1162Xg for 30min at 25 ° C and protein free ultrafiltrates were analyzed for copper using atomic absorption spectrophotometry. RESULTS AND DISCUSSION
LiDid Peroxidation in Human Serum The concentrations of albumin, total proteins, transferrin and ceruloplasmin in all serum pools we studied were within normal reference range. Lipid peroxidation in serum was initiated with 2 mmol/L cupric sulfate. Under such strong oxidative stress, thiobarbituric acid reactive substances (TBARS) in serum were increased 500-630% over controls following 48h of incubation at 37 ° C. On the other hand, with sera incubated without cupric ions, TBARS increased by only 10-20%, indicating a good antioxidant property of human serum. Serum incubated with 1.2 mmol/L ascorbate or ¢¢-tocopherol, in the absence of cupric ion, resisted any further lipid peroxidation as demonstrated by marginal increase in TBARS. More interestingly, serum incubated with cupric ions in the presence of high concentration of ascorbate (10 times higher than the upper limit of physiological range), demonstrated an antioxidant role of ascorbate instead of its expected prcoxidant effect. After 6h of incubation, we observed little increase in TBARS in serum supplemented with ascorbate whereas TBARS increased 40-80% in samples incubated with copper sulfate only. The concentrations of TBARS were significantly lower by independent t-test, two-tailed, in sera supplemented with ascorbate and copper sulfate compared to sera supplemented with copper sulfate alone. After 24h of incubation, TBARS increased 200-330% in sera supplemented with copper sulfate alone as compared to controls. The increase of TBARS was significantly lower (80-160%) in sera supplemented with both copper sulfate and ascorbate, indicating a substantial antioxidant effect of ascorbate even in the presence of high concentrations of cupric ions (Table 1). Also ~-tocopheroi, a well known antioxidant, provided protection similar to that of ascorbate at an equimolar concentration for the first 24h of incubation. At the end of a 48h of incubation, TBARS were still significantly lower (by independent t-test, two-tailed) in serum supplemented with ascorbate and even more in sera supplemented with a-tocopherol.
878
Lipid Peroxidation and Ascorbate
Vol. 50, No. 12, 1992
Unsaturated Fattv Acid Profile in Serum: Protective Role of Ascorbate The antioxidant effects of ascorbate were further studied in 3rd and 4th serum pools by analyzing the fatty acid profiles. Usually the proportion of unsaturated to saturated fatty acids decreases following lipid peroxidation (17-18). The major unsaturated fatty acids we identified
Thiobarbituric
TABLE 1 acid reactive substances (TBARS) in Serum Pools 0 h
Serum Pool 1 Mean(SD) ^ Serum 2.2(0.10) Serum+ Cu ++ * 2.3(0.20) Serum+ Cu ++ + Ascorbate ** 2.0(0.10) Serum + Cu ++ +o¢-Tocopherol # 2.1(0.15) Serum Pool 2 Serum Serum + Cu ++ Serum + Cu ++ + Ascorbate Serum + Cu ++ + c¢-Tocopherol Serum Pool 3 Serum Serum + Cu ++ Serum + Cu ++ + Ascorbate Serum + Cu ++ + ¢¢- Tocopherol Serum Pool 4 Serum Serum + Cu ++ Serum + Cu ++ + Ascorbate Serum +Cu ++ +¢¢-Tocopherol
48h
6 h 24 h wnol/L malonaldehyde equivalent
2.2(0.14) 3.3(0.12)
2.4(0.14) 8.1(0.10)
2.7(0.10) 15.9(0.23)
2.3(0.18)
5.1(0.09)
13.1(0.37)
2.4(0.11)
5.9(0.10)
10.2(0.56)
3.0(0.13) 2.8(0.20)
2.9(0.11) 4.2(0.18)
3.2(0.10) 9.7(0.15)
3.2(0.11) 16.5(0.90)
2.8(0.15)
3.1(0.10)
5.9(0.16)
11.9(0.40)
3.0(0.15)
2.8(0.18)
5.5(0.05)
8,9(0.50)
2.6(0.10) 2.7(0.15)
2.8(0.20) 3,9(0.12)
2.8(0.10) 8.7(0.11)
2.9(0.20) 14.2(0.12)
2.6(0.20)
3.1 (0.07)
5.6(0.19)
12.7(0.25)
2.5(0.12)
3.3(0.08)
5.2(0.05)
11.6(0.25)
2.6(0.12) 2.5(0.12)
2.5(0.10) 4.5(0.20)
2.6(0.15) 10.8(0.23)
2.8(0.13) 18.4(0.41)
2.4(0.20)
3.2(0.12)
6.3(0.15)
13.2(0.16)
2.5(0.14)
3.5(0.07)
5.6(0.10)
9.8(0.07)
*Cu 2+ : 2.0 mmol/L, ** Ascorbate: 1.2 mmol/L, # ¢¢-tocopheroh 1.2 mmol/L, ^ Mean of three replicates were palmitoleic (16:1), linoleic (18:2), oleic (18:1), arachidonic (20:4) and decosahexaenoic (22:6). The major saturated fatty acids were myristic (14:0), palmitic (16:0) and stearic (18:0). These fatty acids were identified based on their mass spectral characteristics and retention times.
Vol. 50, No. 12, 1992
Lipid Peroxldation and Ascorbate
879
Little change was observed in the relative composition of fatty acids in serum incubated in the absence of copper, demonstrating the antioxidant properties of normal human serum. In serum pool 3, the proportion of unsaturated fatty acids dropped from 72.8% to 69.8°/. following 48h of incubation at 37o C whereas in the presence of cupric ions, the proportion of unsaturated fatty acids dropped from 71.3% to 64.8% and then to 57.9% following 24 and 48h of incubation. As expected, the composition of polyunsaturated fatty acids (20:4, 20:3 and 22:6) dropped from an initial value of 16.2% to 10.1% and finally to 6.9%, indicating that polyunsaturated fatty acids are most susceptable to oxidative damage. However, in the presence of ascorbate, after 24h of incubation, the proportion of unsaturated fatty acids decreased from an initial value of 72.0% to only 70.7%. This indicates a substantial antioxidant defense provided by ascorbate against copper-induced lipid peroxidation in human serum. The relative proportion of polyunsaturated fatty acids dropped from 16.4% to only 14.4%. Interestingly, (z-tocopherol provided similar protection as the composition of unsaturated fatty acids dropped from 71.6% to 70.4% (Table 2). TABLE II Effect of peroxldation on fatty acid profile in serum pool 3. FATrY ACIDS % Saturated (14:0, 16:0, 18:0)
% Unsaturated (18:1, 18:2, 20:3, 20:4.,22:6)
%Polyunsaturated only (20:3, 20:4, 22:6)
Serum, Oh
27.8
72.8
16.5
Serum, 24h
29.0
71.0
15.9
Serum, 48h
30.8
69.8
15.7
Serum + Copper, Oh
28.7
71.3
16.2
Serum + Copper, 24h
35,2
64.8
10.1
Serum +Copper, 48h
41.2
57.9
6.9
Serum + Copper + Ascorbate, Oh
28.0
72.0
16.4
Serum + Copper + Ascorbate, 24h
29.4
70.7
14.4
Serum + Copper+ Ascorbate, 48h
39.3
60.6
8.3
Serum + Copper+ u-Tocopherol, Oh
28.4
71.6
16.1
Serum + Copper + a-Tocopherol, 24h
29.6
70.4
13.4
Serum + Copper + u-Tocopherol, 48h
34.6
65.4
9.9
880
Lipid Peroxidation
and Ascorbate
Vol.
50, No.
12, 1992
After 48h of incubation the proportion of unsaturated fatty acids were higher in serum supplemented with ~-tocopherol (65.4%) compared to ascorbate (60.6%). However, both were higher than that observed in serum containing only copper (57.9%). This observation indicates that some antioxidant protection due to ascorbate is still present following such a long incubation time and under strong oxidative stress. Mechanism of Antioxidant DroDertv of Ascorbate It has been generally accepted that ascorbate can act as an antioxidant by scavenging free radicals as it undergoes oxidation to dehydroascorbic acid. In the presence of transition metal ions, however, ascorbate is oxidized very rapidly and generates free radicals rather than scavenging free radicals (19). Auto-oxidation of ascorbate in the presence of copper generates hydrogen peroxide which subsequently reacts with both ascorbate and dehydroascorbate, generating free radicals. Recently, Smith et al reported that 2-imidazolethiones protect ascorbic acid from oxidation induced by copper. Such protective effects were ascribed to the formation of copper complexes (20). The oxidation of ascorbate can also be prevented by chelating agents such as EDTA, cystein, purines, and bathocuproine (21-23). Chelation of metal ions due to protein binding can also prevent auto-oxidation of ascorbate. For example, ceruloplasmin prevents the copper catalyzed oxidation of ascorbate (20). We studied the concentration-dependent serum protein binding of copper by supplementing normal serum pools with 0.1 mmol/L to 2 mmol/L cupric ions. The concentration of copper in protein-free ultrafiltrates (representing non-protein bound copper) is very low when sera were supplemented with up to 0.25 mmol/L copper (ten times higher than the upper limit of the physiological range). Surprisingly, the concentration of copper in the ultrafiltrate was only 0.33 mmoVL even when the serum was supplemented with 2 mmol/L copper (Fig.l), indicating that 84.5% of the copper was still bound to serum proteins. Moreover, uric acid, present in serum may provide additional antioxidant protection because uric acid protects polyunsaturated fatty acids against copper-induced peroxidation (24). We also studied the decay of supplemented ascorbate in human serum. When ascorbate was added to the serum in absence of copper and incubated at 37° C, 69.2% of original ascorbate concentration was present after 24h, and 28.6% after 48h. Phosphate buffer, with equimolar ascorbate showed almost a complete decay of ascorbate in 24h, indicating again that serum prevents oxidation of ascorbate. When ascorbate in phosphate buffer was incubated in presence of 2 mmol/L copper, the expected rapid decrease of ascorbate was observed with almost no ascorbate (
