Follow-up study of lipid peroxides, superoxide dismutase and glutathione peroxidase in the synovial membrane, serum and liver of young and old mice with collagen-induced arthritis

Life Sciences, Vol. 43, pp. 1887-1896 Printed in the U.S.A.

Pergamon Press

FOLLOW-UP STUDY OF LIPID PEROXIDES, SUPEROXIDE DISMUTASE AND GLUTATHIONE PEROXIDASE IN THE SYNOVIAL MEMBRANE, SERUM AND LIVER OF YOUNG AND OLD MICE WITH COLLAGEN-INDUCED ARTHRITIS Tsuyoshi Kasama, Kazuo Kobayashi, Fusao Sekine, Masao Negishl, Hirotsugu Ide, Teruml Takahashl, and Yukle Niwa* Department of Internal Medicine, School of Medicine, Showa University, and Niwa Institute for Immunology* (Received in final form October 6, 1988) Summary Because reactive oxygen species (ROS) are generally believed to play an important role in tissue injury in rheumatoid arthritis, we examined the levels of lipid peroxides, superoxide dismutase (SOD), and glutathione peroxidase (GSH-Px) in the synovial membrane, serum and liver of young (8 wk) and old (12 mo) mice with collagen-induced arthritis. In the synovial membrane, serum and liver, lipid peroxide levels of both young and old mice were increased beginning on the 3rd day after the onset of arthritis. SOD activity, which scavenges Oand inhibits lipid peroxidation, rose markedly in the synovial 2 membrane of young mice in parallel with the increase in lipid peroxide levels, but not so markedly in old mice. Liver GSH-Px activity, which metabolizes already formed lipid peroxides, also rose in young arthritic mice to a greater degree than in old mice. This study suggests that in inflammatory synovial lesions, lipid peroxides are generated due to an increase in ROS concentration, with resultant cytotoxicity, and that younger animals or humans can prevent this unfavorable reaction more effectively than aged ones by enzyme induction. The hypothesis that lipid peroxides formed in the oxidative lesions of the primary organ are released into the serum, trapped by the liver and metabolized there is further supported by the present study. Reactive oxygen species (ROS) and lipid peroxides, which are produced by a free radical chain reaction that can also be initiated by ROS, 'have been implicated in the induction of tissue injury and in the prolongation of severely inflamed and erosive skin lesions including burns and wounds (I-3). Whereas the action of R0S is potent but transient, tissue damage due to lipid peroxides is, although somewhat milder, persistent compared with that due to ROS. One of the ROS scavenging enzymes, superoxide dismutase (SOD), which can also inhibit lipid peroxidation, is able to prevent oxidative damage, and its enzymatic activity is induced under oxygen stress (2, 4-7). Another enzyme, glutathione peroxidase (GSH-Px), is capable of degrading lipid peroxides that have been formed (8, 9). It has been speculated that excess lipid peroxides Correspondence should be addressed to Yukie Niwa, M.D., Ph.D., Niwa Institute for Immunology, 4-4 Asahimachi, Tosashimizu, 787-03, Japan 0024-3205/88 $3.00 + .00 Copyright (c) 1988 Pergamon Press plc

1888

Lipid Peroxides, SOD and GSH-Px in RA Mice

Vol. 43, No. 23, 1988

formed in tissues are released into the serum (3) and eventually metabolized into alcohol and water by liver GSH-Px. Serum lipid peroxide levels are markedly increased in fulminant hepatitis (I0), eclampsia (Ii), diabetes (12, 13) and extensive burns (1-3, 7). Nishigaki et al. (3) found that lipid peroxides generated by skin cell damage at the site of burn lesions are released into the serum and diffuse into other organs, causing subsequent organ injury. In rheumatoid arthritis (RA), it has been reported that excessive ROS exert a cytotoxic effect in the joints (14, 15) and also contribute to the decrease in lymphocytes or lymphocyte subsets and reduced lymphocyte responses (14). Yoshikawa et al. (16) recently documented an increase in lipid peroxides in the synovial membrane of the rats with adjuvant-induced arthritis; administration of SOD was effective in decreasing lipid peroxide levels in situ, with resultant amelioration of the arthritis. In the present study, young and old mice with collagen-induced arthritis were examined over a 6 month period to assess the changes in lipid peroxide levels and the activity of SOD and GSH-Px in the synovial membrane, serum and liver. We have recently reported a correlation between changes in the lipid peroxide levels and SOD and GSH-Px activities in the skin lesions, serum and liver of burned mice (17); moreover, we have demonstrated that in aged animals and humans, the level of SOD induction in burned skin is lower than in nonaged controls (7). In this manuscript, we discuss the correlation between the changes in lipid peroxide levels and the protective mechanism of enzymes in the joint, serum and liver, and the comparison of the capacity for SOD and GSH-Px induction between young and old R A m i c e . An hypothesis regarding the metabolic pathway of lipid peroxides is proposed. Materials and Methods Animals. DBA/IJ male mice, age 8 wk or 12 mo, were purchased from Jackson Laboratories (Bar Harbor, ME). Immunization with type II collagen. I00 ~g chicken type II collagen (Genzyme Corporation, 75 Kneeland Street, Boston, MA), dissolved in i00 ~i of 0.I M acetic acid was emulsified with I00 ~i complete Freund's adjuvant (Difco Laboratories, Inc., Detroit, MI). The emulsion was injected at time 0 in four to six sites on the back of mouse. Three weeks later, a booster injection of I00 pg of collagen in I00 ~i acetic acid was given i.p. As a control, complete Freund's adjuvant with acetic acid only was injected at time 0, and acetic acid alone was injected i.p. at 3 wk. Evaluation of arthritis. The onset of arthritis was identified by the appearance of paw redness and swelling; foot pad thickness was then measured daily with a micrometer (MITUTOYO MFG Co LTD). The degree of arthritis was expressed as the percentage increase in foot pad thickness. Sample preparation. Mice were sacrificed at 0, I, 2 and 3 wk after initial immunization, and twice to four times a week after the onset of arthritis, through wks 7 or 8. Each fresh organ sample was homogenized in a teflon homogenizer for about four min under aerobic conditions, then suspended in 0.9% NaCI for the lipid peroxide assay or in 125 mM phosphate buffer for the SOD and GSH-Px assays. Thereafter, the samples of synovial membrane, serum and liver were assessed in these assays.

Vol. 43, No. 23, 1988

Lipid Peroxides,

SOD and GSH-Px in RAM_ice

1889

Lipid peroxide and enzyme activity assays. Lipid peroxide levels were assessed by the methods of Yagi (18) and Ohkawa et al. (19) with some modifications (i, 2). This assay measures lipid peroxidation as thiobarbituric acid (TBA)-reactive substances. The standard assay mixture contained 0.I ml of 10% tissue homogenates, 0.4 ml H20 , and 0.2 ml of 7% sodium dodecylsulfate. This mixture was gently stirred and 2 ml of 0.I N HCI was added. Next, 1 ml of a mixture of 0.67% TBA and acetic acid (I:I) was added, and the mixture was heated for 45 min at 95°C. After cooling in water, 5 ml of n-butanol was added to extract the lipid peroxides, and the sample was vortexed and centrifuged for 15 min at 1,250 g. The lipid peroxide concentration in the butanol layer was then determined on a fluorescence spectrophotometer (Hitachi, MPF4, Japan) at 515 nm for excitation and 553 nm for emission. To determine the lipid peroxide levels in the serum, 20 ul serum was added to 4 ml of 1/12 N H2SO 4 and 0.5 ml of 10% phosphotungstic acid. The mixture was stirred, left for 5 min at room temperature, and centrifuged for i0 min at 1,250 g. The sediment was then resuspended in 4 ml distilled water and assayed for lipid peroxides as described above, beginning with addition of i ml of 0.67% TBA : acetic acid (I:I). The SOD activity was assayed by the method previously described for blood cells or skin tissues (2, 20, 21). Briefly, 5% tissue homogenates were made 0.08% in Triton-x-100, kept on ice for 1 hr and centrifuged for I0 min at 7,000 g; 0.5 ml of the supernatant was added to 2 ml of the assay mixture containing I00 p M xanthine and 66 ~ M ferricytochrome ~ (type III) in 125 mM potassium phosphate, pH 7.8 (20, 21). For serum, i00 p l serum was directly added to the assay mixture. The formation of 02 on addition of xanthine oxidase was determined by cytochrome £ reduction following by the initial rate of absorbance increase (30-60 sec) at 550 nm by a spectrophotometer (Beckman, UV5260, USA). The SOD activity was estimated by the inhibition of the cytochrome c reduction after correction for the reduction due to samples before the addition of xanthine oxidase (2). The amount of SOD in the sample to inhibit the rate of cytochrome c reduction by 50% was defined as I unit of activity and calculated using the Asada's formula (22). GSH-Px was measured by the method of Lawrence and Burk (23), in which GSH-Px activity was coupled to the oxidation of NADPH by glutathione reductase. The oxidation of NADPH was followed spectrophotometrically at 340 nm at 37°C. The reaction mixture consisted of 50 mM potassium phosphate buffer (pH 7.Q), 1 mM EDTA, i mM NAN_, 0.2 mM NADPH, 1 mM glutathione, 2 U of glutathione reductase, and i.~ mM cumene hydroperoxide or I0 mM tert-butyl hydroperoxide. With the samples of each organ the total volume was made 2.0 ml. Enzymatic activity was expressed as nmol NADPH oxidized per min. Protein was measured by the method of Lowry et al. (24). Specific activity was expressed as units per mg protein. Five to seven mice were sacrificed for each experiment; the results are expressed as the mean + SEM of replicate assays. Statistical significance was ascertained by StudentTs t-test. Results Incidence of arthritis. As shown in Figs. 1 and 2, arthritis appeared 3 to 5 wk after immunization, both in young and old mice. Control mice showed no swelling through the 7th wk after immunization. Lipid peroxides and the activity of SOD and GSH-Px in synovial membrane. Before the occurrence of arthritis, the lipid peroxide levels and the activity of SOD and GSH-Px in synovial membrane in all groups of mice did not significantly

1890

Lipid Peroxides, SOD and GSH-Px in R A M i c e

Vol. 43, No. 23, 1988

change from those at wk 0 (Fig. I). In addition, there was no significant difference in lipid peroxide levels between collagen-immunized and control mice or between young and old immunized mice or control mice. On the 3rd day after the onset of arthritis, both young and old arthritic mice showed a significant rise in synovial membrane lipid peroxide levels (P< 0.01), which remained elevated for 3 wk (Fig. 3). There was no significant difference in lipid peroxide levels between young and old mice in either the diseased or control groups (P>0.05) (Fig. 3). Before the onset of arthritis, there was no significant difference among the synovial SOD and GSH-Px activities of all the four groups of mice (Figs. 4 and 5). However, the synovial SOD activity of the young mice increased on the 3rd day after the onset of arthritis (P<0.01) and remained elevated for 2 wk (0.01
4o

~z =

=.

0

!

t

...... ii t)oo~loI

I11111111111/ It [qHl

FIG. 1 % swelling of foot pad of collageninduced arthritis mice plotted on the date all of which were calculated from immunization. Open circle (O) denotes diseased, 8 wk age mice, closed circle ($) diseased, 12 mo mice, open triangle (~) nondiseased, 8 wk age control mice, and closed triangle (A) non-diseased, 12 mo mice.

60,

I

FIG. 2 % swelling of foot pad of collageninduced arthritis mice plotted on the date calculated from immunization and from actual occurrence of arthritis.

40,

(o), (.), (~), CA): od

T

I n l l l l U n l Zll | i o n

iw

~w

3w

~, ~ 3d

~d

7d

2w

,c u r i ' c n t , o

Tot q[

brJo,~lor

IIItllllll,.,

,ll|ui

~lllhllll~,

~w

see legends for Fig. I.

Vol. 43, No. 23, 1988

Lipid Peroxides, SOD and GSH-Px in RA Mice

1891

50

4O

30

FIG. 3 The changes in lipid peroxide levels in the synovial membrane of collagen-induced arthritis mice.

=

= 2o

lO'

(o), ( i ) , •

(}d

~lw

T

-~tv ,ll|l'l

I',v

IIIIIIIIIII I/*ll

41". I¢lll

4'

_

o

(~), CA):

see legends for

Fig. I.

ill :llll'l

[

.il lhllll-

I I l ' t I I I I I'111 I b

I l l l l l l l l II I / , I | I I ) l l

OI

.irl

hrlll,,

30 =

FIG. 4 The changes in SOD activity in the synovial membrane of collagen-induced arthritis mice.

(o), (e), Ca), (-): see legends for lid

l;~

Ji'

~V,' .ll I('l

J

|',v

4W ~'

i i Iiilll.i i i i / Ii t l i J l l

ii1.11111111 / i i l l l l l

'1{I

/

7tl

.>iv

.1[11.1.

ill Ihllll%

I',v

Fig. I.

o l t t l l ' l ( ' l i t l" lii" .11 t i l l ' t i P ,

0.2

Fig. 5

& _

/"

The changes in GSH-Px activity in the

--'= 0 I

synovial membrane of collagen-induced arthritis mice.

Co), (e), (~), (-): see legends for Otl I

~lw

2'," ;ift(~l

v

3w

IIIIIIItllll/;ItlOll

4w, ~ u _ h l I -/ /

IIlllllillki/.it

|Lilt

Ll( I It1 i , . l i t L lie . i t t h l i | l ~

7¢I

~ ,

.llh'l

.illhllll~

h~

Fig. I.

1892

Lipid Peroxides, SOD and GSH-Px in RA Mice

Vol. 43, No. 23, 1988

The GSH-Px activity of the young mice showed only a slight increase from the 3rd day to 3 wk after the onset of arthritis, compared with both that before the booster and that of control mice (0.025O.05). SOD and GSH-Px activities in the serum were almost negligible compared with those obtained in other tissues (data not shown). The serum SOD activity determined in this study was almost comparable with extracellular-SOD (EC-SOD) reported by Marklund (25). However, although he demonstrated specific EC-SOD (25), we (26) previously reported a remarkable increase in SOD activity in the serum from the patients with hemolytic anemia at their exacerbated stages and the decrease at their controlled stages. Our data is also supported by Michelson (27, 28). In this connection, there are two possibilities regarding the sources of serum SOD; one is EC-SOD by Marklund and the other may be the result from the SOD derived from a damaged organs and tissues. Lipid peroxides and the two enzyme activities in liver. Before the onset of arthritis there was no statistically significant difference in lipid peroxide levels in the liver among all of the groups of mice tested. Those of the young mice showed a significant increase at the 3rd day after the onset of arthritis, compared with those before the booster and those of control mice and diseased old mice (P<0.O01 vs control mice, P
Vol. 43, No. 23, 1988

Lipid Peroxides, SOD and GSH-Px in R A M i c e

1893

10 =

FIG. 6 The changes in lipid peroxide levels in serum of collageninduced arthritis mice.

_=

(o), (e), (~), (A): see legends for Fig. I.

lOd Iw

2w • 3w hi|el

lllllnlll|l/

4w I

T,~ 3~1

Inlmlll'llTlll ion

2W

7(I

|W,

,tftez au thn Ill'-,

Ill(HI IlL'I" u l I~1' I I1" I~

OI ,11 I h l II1~

100,

r,

FIG. 7 The changes in lipid peroxide levels in liver of collageninduced arthritis mice.

E

=

50

(o), (e), (A), (A): see legends for Fig. I.

lW

Od

l

,

2W

.

3w

4w;

T'

3d

al'tel' IiiiriluR|zot|on

?d

81"1o1'

.

2w

3w.

Al'lhl'lllS

occurrence

lrlllll g l11zilt |OR

o[

n r t h z ltls

FIG. 8 The changes in GSH-Px in liver of collageninduced arthritis mice.

= | U

(o), (,), (~), (A): see legends for Fig. I.

0dl

,]w

2w ~ .If[el'

IIIIiilu ]117,1t ion

3w

4wl,._~3d

IIIIIIILIIIIZtI|IOII

occu r rorlco of art hr=tls

7d .

2w

at'tel' o|'(h|'||lS

3wt

b

1894

Lipid Peroxides, SOD and GS}I-Px in RA Mice

Vol. 43, No. 23, 1988

Discussion We found that the lipid peroxide levels in the synovial membrane of young and old mice with collagen-induced arthritis were significantly increased for at least 3 wk, beginning on the third day after the onset of arthritis. This increase in lipid peroxides in synovial membrane seems to be due to an increase in the in situ production of ROS (14, 15). Exposure to oxidative stress is reported to cause a I0 - 20 fold increase in SOD activity in some bacteria (4) and to enhance significantly the SOD activity in plants (5, 6). We have also recently documented significant induction of SOD activity in inflammatory skin lesions with erosion and ulcer formation (2, 7). Conflicting reports have appeared regarding the effect of age on SOD activity. While a decrease in SOD activity in elderly people has been reported (29-31), other studies in both humans and animals have found comparable levels of SOD in young and old individuals (7, 32). Before the onset of arthritis, we observed similar levels of SOD activity in synovial membrane in both young and old mice. However, the SOD activity in young mice with arthritis rose significantly, presumably induced by the increased levels of ROS in the arthritic synovial membrane, while the SOD activity in old mice did not increase, although lipid peroxide levels increased in both young and old mice. This is consistent with our recent findings (7), and those of Michelson (32), that SOD activity in aged animal or humans is not lowered but similar to that in non-aged ones unless they are under oxidative stress. In contrast to SOD, GSH-Px activity increased only slightly in the arthritic synovial membrane of both young and old mice. This may be explained by the fact that in the diseased joint, lipid peroxidation occurs as a result of excessive generation of ROS in situ; SOD is induced in response to elevated ROS levels, whereas GSH-Px, which only metabolizes lipid peroxides, is not induced in response to ROS. However, it is the liver that is the major site of lipid peroxide metabolism. In young mice with arthritis, liver GSH-Px activity was significantly elevated, apparently in response to the increase in lipid peroxides in the liver, whereas liver SOD activity was not so markedly elevated in either young or old mice with arthritis. This may be explained by the fact that lipid peroxides are metabolized in the liver by GSH-Px but not SOD, which cannot scavenge lipid peroxides once they are formed. Thus, in old arthritic mice, sufficient induction of activity was found neither for SOD in the joint nor for GSH-Px in the liver, furthermore, the induction that was observed, was sluggish in comparison with the increase in lipid peroxides (Fig. 4), supporting our recent investigation (17, Ito et al., submitted for publication). However, as shown in Fig. 5, the level of GSH-Px activity in the joint was induced comparably in arthritis old and young mice, although the degree was slight. This is speculated to be due to that joint is not the specific organ where GSH-Px specifically metabolized GSH-Px. From this study we may infer that SOD is more strongly induced in response to oxygen toxicity than GSH-Px; furthermore, there is a greater difference between old and young mice in the induction of SOD than in the induction of GSH-Px. It is well known that in the liver, turnover rate of GSH-Px is rapid and its concentration is high (5-10 mM). In our present experiment, in contrast to those in the joint and serum (Figs. 3 and 6), lipid peroxide levels of young mice began decreasing 7 days after the occurrence of arthritis and kept the trends to decrease thereafter (Fig. 7). Our recent study on non-aged,

Vol. 43, No. 23, 1988

Lipid Peroxides, SOD and GSH-Px in R A M ice

1895

burnt mice also showed that lipid peroxide levels in the liver decreased 8 hr after burn infliction and again began lowering from 6th day although they increased slightly 4 hr and 3 day after burn. These changes in lipid peroxide levels in the liver seems to reflect the rapid turnover rate and high concentration of GSH-Px in the liver. It is generally speculated that lipid peroxides, once formed in primary inflammatory lesions, are released into the serum and trapped by the liver, where they are metabolized to alcohol and water by GSH-Px (8, 9). The changes in lipid peroxide levels in the joint, serum and liver, as well as the SOD activity in the joint and GSH-Px activity in the liver observed in this study are all consistent with this hypothesis. Finally, this study raises the possibility that assaying synovial SOD activity in arthritis patients might be of diagnostic or prognostic value. Given the differences seen between young and old mice with respect to SOD induction, it would also be of interest to compare SOD levels in JRA and adult RA patients. References I. Y. NIWA, T. KANOH, T. SAKANE, H. SOH, S. KAWAI and Y. MIYACHI, J. Clin. Biochem. Nutr. 2 245-251 (1987). 2. Y. NIWA, T. KANOH, T. SAKANE, H. SOH, S. KAWAI and Y. MIYACHI, Life Sci. 40 921-927 (1987). 3, I. NISHIGAKI, M. HAGIHARA, M. HIRAMATSU, Y. IZAWA and K. YAGI, Biochem. Med. 24 185-189 (1980). 4. E.M. G-REGORY and I. FRIDOVICH, J. Bacteriol. 114 543-548 (1973). 5. K. TANAKA and K. SUGAHARA, Plant Cell Physiol. 21 601-611 (1980). 6. H.D. RABINOWITCH, D.A. CLARE, J.D. CRAPO and I. FRIDOVICH, Arch. Biochem. Biophys. 225 640-648 (1983). 7. Y. NIWA, T. KASAMA, S. KAWAI, J. KOMURA, T. SAKANE, T. KANOH and Y. MIYACHI, Life Sci. 42 351-356 (1988). 8. B.O. CHRISTOPHERSEN, Biochim. Biophys. Aeta 164 35-46 (1968). 9. C. LITTLE and P.J. O'BRIEN, Biochem. Biophys. Res. Commun. 31 145-150 (1968). I0. T. SUEMATSU, T. KAMADA, H. ABE, S. KIKUCHI and K. YAGI, Clin. Chim. Acta 79 267-270 (1977). Ii. M--?MASEKI, I. NISHIGAKI, M. HAGIHARA, Y. TOMODA and K. YAGI, Clin. Chim. Acta 115 155-161 (1981). 12. Y. SATO, N. HOTTA, N. SAKAMOTO, S. MATSUOKA, N. OHISHI and K. YAGI, Biochem. Med. 21 104-107 (1979). 13. I. NISHIGAKI, M--?HAGIHARA, H. TSUNEKAWA, M. MASEKI and K. YAGI, Biochem. Med. 25 373-378 (1981). 14. Y. NIW-A, T. SAKANE, M. SHINGU and M.M. YOKOYAMA, J. Clin. Immunol. 3 228240 (1983). 15. J.M. McCORD, Science 185 529-531 (1974). 16. T. YOSHIKAWA, H. TANAKA and M. KONDO, Biochem. Med. 33 320-326 (1985). 17. S. KAWAI, J. KOMURA, Y. ASADA and Y. NIWA, Arch. Dermatol. Res., in press (1988). 18. K. YAGI, Biochem. Med. 15 212-216 (1976). 19. H. OHKAWA, N. OHISHI and---K. YAGI, Anal. Biochem. 95 351-358 (1979). 20. Y. NIWA, T. SAKANE and Y. MIYACHI, Biochem. Pharmacol. 33 2355-2360 (1984). 21. Y. NIWA, T. SAKANE, Y. MIYACHI and M. OZAKI, J. Clin. Microbiol. 20 837842 (1984). 22. K. ASADA, M. TAKAHASHI and M. NAGATE, Agr. Biol. Chem. 38 471-473 (1974). 23. R.A. LAWRENCE and R.F. BURK, Biochem. Biophys. Res. Commun. 71 952-958 (1976).

1896

Lipid Peroxides, SOD and GSH-Px in RA Mice

Vol. 43, No. 23, 1988

24. 0.H. LOWRY, N.J. ROSEBROUGH, A.L. FARR and R.J. RANDALL, J. Biol. Chem. 193 265-275 (1951). 25. S. L. MARKLUND, Biochem. J. 222 649-655 (1984). 26. Y. NIWA, K. SOMIYA, A.M. MICHELSON and K. PUGET, Free Rad. Res. Commun. 1 137-153 (1985). 27. A. BARET, M.A. BAETEMAN, J.F. MATTEI, P. MICHEL, B. BROUSSOLLE and F. GIRAUD, Biochem. Biophys. Res. Commun. 98 1035-1043 (1981). 28. A. BARET, P. SCHIAVI, P. MICHEL, A.M. MICHELSON and K. PUGET, FEBS Lett. 112 25-29 (1980). 29. G.A. GLASS and D. GERSHON, Biochem. Biophys. Res. Commun. 103 1245-1253 (1981). 30. U. REISS and D. GERSHON, Biochem. Biophys. Res. Commun. 73 255-262 (1976). 31. M.J. IM and J.E. HOOPES, J. Invest. Dermatol. 82 437(A) (1984). 32. A.M. MICHELSON, K. PUGET, P. DUROSAY and J.C. BONNEAU, Superoxide and Superoxide Dismutases, A.M. Michelson, J.M. McCord and I. Fridovich (eds.), p. 467-499, Academic Press, London (1977).