Motec. Aspects IVied. Vol. 12, pp. 93-105, 1991 Printed in Great Britain. All rights reserved.
0098-2997/91 $0.00 + .50 © 1991 Pergamon Press plc.
REACTIVE OXYGEN SPECIES DAMAGE TO DNA AND ITS ROLE IN SYSTEMIC LUPUS ERYTHEMATOSUS S. Blount, H. R. Griffiths and J. Lunec Wolfson Research Laboratories, Department of Clinical Chemistry, Queen Elizabeth Medical Centre, Birmingham, B15 2 TH, U.K.
Systemic Lupus Erythematosus Systemic lupus erythematosus (SLE) is an autoimmune disease of unknown aetiology. It is characterised by a wide variety of clinical presentations, the most common features being arthritis, vasculitis, skin rashes and renal manifestations (Aitcheson and Tan, 1982). There is evidence of a marked disturbance of both cellular and humoral immunity, and the disease is notable for the wide variety of antibodies, especially antinuclear antibodies, found in the serum. SLE patients also have characteristically low complement levels which may play a role in the build up of circulating immune complexes found in the sera. Persistent production of excess autoantibodies and the subsequent formation of immune complexes are attributed to the severe organ lesions seen in this disease. In particular, deposition of immune complexes in the kidney basement membrane gives rise to the glomerulonephritis suffered by 50% of SLE patients (Bruneau and Benveniste, 1979). Diagnosis of SLE currently relies on the serological estimation of antibodies binding to doublestranded DNA in a radioimmunoassay using labelled DNA (Tan et al., 1982). Antibodies which bind to DNA are found in the sera of patients with Rheumatoid Athritis (RA) and other rheumatological disorders such as Sjogren's syndrome (Notman e t a l . , 1975). These bind preferentially to single-stranded DNA whereas antibodies binding to double-stranded DNA are most uniquely correlated with SLE and its periods of clinical activity, especially with its renal complications (Aitcheson and Tan, 1982).
Generation of Free Radicals Arthritis is a component of disease of both SLE and RA. It is characterised by a perpetuated inflammatory response and resultant tissue damage. During inflammation neutrophils, macrophages and eosinophils become activated by a variety of stimuli, such as immune complexes, and undergo a cyanide-independent respiratory burst and degranulation. During the respiratory burst oxygen consumption is increased and the membrane-bound NADPH oxidase is activated (Lunec et al., 1987a). This enzyme catalyses the reduction of molecular oxygen to superoxide by electron transfer from NADPH. Reduction of oxygen at the cell surface results in the superoxide
93
94
S. Blount et al.
radical being released into the extracellular space where it can dismutate to form hydrogen peroxide. The reaction is catalysed by superoxide dismutase - an enzyme found in eukaryotes to protect against the toxic effects of oxygen (McCord and Fridovich, 1969). Hydrogen peroxide, although itself not a radical, can react in the presence of metal ions to produce the highly reactive hydroxyl radical (McCord and Day, 1978). Hydrogen peroxide is an important metabolite in the conversion of oxygen to water via this route because it is lipophillic and thus has the ability to traverse cell membranes (Allan et al., 1988). The hydroxyl radical diffuses only 15/~ before reacting so the location of the hydrogen peroxide determines where the radical attack will take place (Ward et al., 1987). The hydroxyl radical can react with all cellular macromolecules including lipids (Logani and Davies, 1980), proteins (Wolff et al., 1986; Lunec et al., 1987b) and DNA (Filho and Meneghini, 1984). Reaction with lipid results in the formation of lipid peroxides, hydroperoxides and diene conjugates. Measurement of free radical damage to lipids can be assessed by thiobarbituric acid (TBA) reactive products (Gutteridge and Quinlan, 1983). Many different lipid breakdown products react with TBA and it is difficult to define and quantify one specific product of peroxidation.
Biological End Points of Free Radical Damage Hydroxyl radical damage to proteins can be measured by high performance liquid chromatography (HPLC) which characterises oxidation products of particular amino acids. Those most susceptible to this kind of damage include the aromatics (tryptophan, tyrosine and phenylalanine), histidine and the disulphide bond formed between cysteine molecules (Griffiths et al., 1988). Damage to DNA can be measured at the level of the individual bases (Teebor et al., 1988) or sugar molecules (Gutteridge, 1984) or by detecting changes in the overall conformational structure (Curtis Johnson, Jr, 1978). Several HPLC assays to measure altered DNA bases exist such as those for thymine glycol (Frenkel etal., 1981) and 5-hydroxymethyluracil (Frenkel etal., 1985). Recently, an assay for the measurement of 8-hydroxydeoxyguanosine (8OHDG) has been described which links the use of HPLC with electrochemical detection, thus increasing the sensitivity of the assay several fold (Kasai and Nishimura, 1984a). This product is formed in a hydroxyl radicalmediated reaction and is a definitive end point for measurement of oxidative damage to DNA (Aruoma et al., 1989).
Free Radicals in Inflammation As described above a perpetuated inflammatory response gives rise to release of highly reactive species into the extracellular milieu which may then enter cells via the hydrogen peroxide intermediate. In such a case as arthritis where inflammation is localised to a joint free radical products may exist in I.tM quantities and thus exert a damaging effect on surrounding macromolecules (Root et al., 1975). A detailed study of the effects of reactive oxygen species (ROS) on the protein IgG have been investigated and are described in detail elsewhere in this issue. Briefly, in Rheumatoid Arthritis antibodies are raised to the "self" molecule IgG. It was suggested that IgG may have been altered by free radical damage and that antibodies were in fact raised against an altered form of IgG (Lunec et al., 1985). Results of the experiments showed that both structural and effector functions were significantly altered as a result of free radical attack (Griffiths and Lunec, 1988). Extrapolating from this it was decided to investigate the effects of ROS on the macromolecule DNA. SLE is a chronic inflammatory disease in which antibodies binding to "self" DNA are found. Although the measurement of these antibodies has diagnostic value their role in the disease process is unclear. Certainly they are involved in the formation of immune complexes, responsible for tissue damage particularly in the kidney basement membranes. Since it had been shown that ROS,
Reactive Oxygen Species Damage to DNA
95
in physiologically generated concentrations appertaining to the inflamed joint, were able to induce changes to the IgG molecule, studies of the effects of ROS, as generated during inflammation, on DNA were initiated. Polyclonal antibodies found in SLE sera served as a diagnostic tool to investigate these effects. Antibodies in SLE sera bind both to double and single-stranded DNA but as previously mentioned titres of binding to double-stranded DNA are most uniquely correlated with disease activity. In view of this it is important to obtain a DNA sample for use as antigen which is in a pure form. Commercially bought double-stranded DNA contains some single-stranded contaminants which can be digested with the enzyme S1 nuclease which selectively cleaves bases from an unpaired strand. If this is then used as the immobilised antigen in an enzyme-linked immunosorbent assay (ELISA) antibodies which recognise single-stranded DNA will not interfere.
Methods for Measuring Free Radical-Induced DNA Damage ELISA To investigate the effects of binding of polyclonal antibodies found in SLE sera as a marker of ROS-mediated damage to DNA, purif.,ed double-stranded DNA was immobilised onto an ELISA plate firstly in its native form and after sufficient blocking of unreactive sites polyclonal SLE sera was added and binding was measured using peroxidase-conjugated goat anti-human antibodies to label serum IgG, IgA and IgM antibodies binding to the DNA. The level of binding was then compared with levels of binding achieved in a similar assay using double-stranded DNA that had been reacted with ROS. The DNA used was incubated with physiological concentrations of hydrogen peroxide in the presence of ascorbic acid, a reducing agent. No metal ions were added to the system and all reagents were treated with chelex resin. This was to ensure that the only metal ions available to catalyse the formation of the highly reactive hydroxyl radical were located on the DNA structure itself thereby creating site-specific damage rather than producing the radical species in free solution. A comparison of the binding of SLE antibodies to native DNA compared to the DNA incubated with reactive oxygen species showed that hydrogen peroxide mediated damage caused an increase in antibody binding (Blount etal., 1989). This effect was inhibited if the compounds desferrioxamine or thiourea were added to the DNA prior to its incubation with hydrogen peroxide and ascorbic acid. Since thiourea is a hydroxyl radical scavenger and desferrioxamine a metal ion chelator, the results would suggest that the hydroxyl radical can react with and alter the DNA structure. Also, that antibodies found in SLE sera bind in greater number only if metal ions found on the surface are available to react with the hydrogen peroxide and are not chelated, with desferrioxamine for example. During these studies the nature of the antibodies binding to DNA, both in its native and radicaldenatured form, was investigated. It was found that the predominant isotype binding to both types of DNA was IgG. However, when the changes induced by the hydroxyl radical were analysed it was clear that antibodies of the IgA isotype were able to detect hydrogen peroxide mediated changes to the DNA most sensitively (Fig. 1). This is of particular interest since antibodies of the IgA isotype have been described'to be sensitive to changes in disease activity in SLE, particularly relating to renal manifestations (Gripenberg and Helve, 1986). The changes induced by hydrogen peroxide and ascorbic acid in the DNA molecule have several implications. Firstly, the increase in binding of antibodies to a denatured form of DNA might suggest that in SLE some form of alteration of the DNA molecule does occur and it is this DNA that is the antigen against which anti-DNA antibodies are raised. These results are also supportive of the hypothesis of a perpetuated inflammatory response initiating macromolecular damage.
96
S. Blount et al.
Secondly, SLE sera antibodies have increased binding to a DNA that has been incubated with hydrogen peroxide and ascorbic acid in the presence of catalytic metal ions. This observation is relevant to the development of a better assay in which to measure anti-DNA antibodies.
0.40 -
IgA
0.39 0.36 -
\
0.37 E 0.36 _ E
._T_~ " ,.X,
0.35 0.34 t~ 0.33 O 0.32
\,
0.31 0.30 DNA+PBS control
2.0
E c-
DNA+ H202/AA
DNA + H202/AA + desferrioxarnine
2.0
IgM
1.8
1.8
1.6
1.6
1.4
q
IgG
T
/
1.4
T
O4 1.2 O~ 1 O
-
DNA + H202/AA + thiourea
e~ 1.2 o3 ,~r
q 0.6
0.8
0
0.6
0.6
0.4
0.4
0.2
0.2 DNA + PBS DNA + DNA + H202/AA H202/AA + desferrioxamine
DNA + H 202/AA + thiourea
)NA+PBS
..L..,._
DNA+ DNA + H202/AA H202/AA + desferrioxamine
S,"
DNA + H 202/AA + thiourea
Fig. 1. Differential binding of anti-DNA isotypes against reactive oxygen species denatured DNA. Binding of anti-IgG, IgM and IgA isotypes to double ( ~ ) and single (ff3) stranded DNA. At present the assays available for measuring anti-DNA antibodies use nauve double-stranded DNA in a radioimmunoassay (Stokes et al., 1982). Results obtained from this method are useful in the diagnosis of SLE. However, other rheumatoid disorders have antibodies which bind to DNA, particularly single-stranded DNA and these may create false positive results. Because of this it is essential to use a purified antigen of the single or double stranded form only. Even in doing this patients with RA have serum antibodies which bind to give a similar titre to mildly active SLE patients (Swaak et al., 1981). From the initial observations that antibodies in SLE sera bind better to D N A that had been reacted with reactive oxygen species it was decided to investigate the ability of antibodies found in rheumatoid sera to bind to DNA reacted with reactive oxygen species when
Reactive Oxygen Species Damage to DNA
97
compared to the native double-stranded DNA used in standard immunochemical and radioimmunoassay. It was found that whilst in SLE patients there was a significant increase in binding to the ROS-denatured DNA this trend was not observed in the rheumatoid patients. In the rheumatoid group, antibodies in individual sera were shown to either increase or decrease in binding to the ROS-denatured DNA (Blount et al., 1989). An ELISA assay was established using double-stranded DNA (reacted with hydrogen peroxide and ascorbic acid) as the immobilised antigen and it has been found to compare favourably with the standard radioimmunoassay used, with the advantage that it can assign an optical density value to each patient's sera analysed. This is most important for patients believed to be mildly active since fluctuations in antibody binding can easily be quantified to see if they relate to clinical activity. The radioimmunoassays used conventionally defined values below a defined cut-off point simply as negative titres. Within this group patients with mild SLE may go undetected. Using the ELISA assay with ROS-denatured-DNA results obtained may be more discriminating towards this group. The reaction of hydrogen peroxide and ascorbic acid with DNA produces changes in the macromolecule which can be detected using polyclonal SLE sera antibodies. The levels of reactive oxygen species generated in these experiments are consistent with those produced during an inflammatory response. Further investigation is needed to determine whether or not DNA damage plays a role in anti-DNA antibody production and also whether or not DNA from patients with SLE shows any signs of degradation by reactive oxygen species in vivo. This latter point will be discussed below. Activated granulocytes are not the only source of oxygen free radicals produced during normal oxidative metabolism. The superoxide radical, for example can also be produced by autoxidation of the drug detoxifying system, cytochrome P450, and the mitochondrial respiratory chain (Fridovich, 1978). In the presence of superoxide dismutase and metal ions this will react to form hydrogen peroxide and the hydroxyl radical as previously described. The hydroxyl radical can cause double- and single-strand breaks (Rhaese and Freese, 1968; Ward et al., 1987), base alterations (Massie et al., 1972) and damage to the sugar-phosphate moieties (Gutteridge, 1984). All these changes could affect germ-line genes and affect the highly accurate transcription of the DNA into proteins. It is thus important to monitor changes in DNA structure at several levels.
Cell Culture Experiments which investigated the effects of hydrogen peroxide on cell viability using a lymphocyte model were able to show that hydrogen peroxide added to cell culture medium caused a significant increase in cell death when compared to cells incubated in hydrogen peroxide-free medium (Allan et al., 1988). It was also shown that changes in nuclear DNA were involved in this process. This reaffirms the idea that hydrogen peroxide is able to gain access to nuclear DNA and, as in this case, cause structural changes. Nuclear DNA was analysed using a fluorescence activated cell sorter to study DNA nucleoids -.nuclear DNA stripped of histone proteins (Milner et al., 1987). After incubation with hydrogen peroxide nucleoid DNA showed significant expansion, which was inhibited by hydroxyl radical scavengers or metal ion chelators. It was also found that within the lymphocyte population, cell subsets were differentially susceptible to damage; CD8 ÷ lymphocytes were highly susceptible, B cells highly resistant, and CD4 ÷ lymphocytes were of intermediate susceptibility (Allan et al., 1987). Experiments using cells cultured with medium containing ROS have not only provided information about changes in cellular DNA, they have also investigated the hypothesis that during the radical chain reactions clastogenic factors may be produced which are able to cause secondary damage to
98
$. Blount et al.
compounds as a result of the initial free radical insult. Results of these studies have shown that ultrafillrates of cells in culture medium containing ROS can induce chromosomal aberrations and sister chromatid exchanges in lymphocytes from normal healthy donors (Emerit et al., 1985).
Monoelonal Antibodies Having established that gross structural damage to DNA occurs following exposure to hydrogen peroxide, either as a direct attack from hydroxyl radicals or from secondary clastogenic factors with stabilities significantly greater than those of the oxygen radicals, a more detailed investigation to identify these changes was carried out using a panel of monoclonal antibodies produced spontaneously from (NZB/NZW)F1 and MRL/MPlpr/lpr mice. Results of in vitro experiments showed that DNA previously incubated with hydrogen peroxide in the presence of a reducing agent and metal ions had: increased exposure of DNA bases, changes in binding to the sugar-phosphate moiety, and increased expression of conformational determinants when compared to native doublestranded DNA.
HPLC A further, and yet more specific way, of determining DNA damage is to measure a particular defined end product of oxidative damage such as an alteration in a base structure. As briefly mentioned earlier, assays to measure DNA bases commonly use HPLC methods. In the detection of 8OHDG, a product of deoxyguanosine formed in a hydroxyl radical mediated reaction (Fig. 2a), HLPC has been linked with electrochemical detection thereby increasing the sensitivity of the method. Femtomoles of base adduct can be detected (Floyd etal., 1988). Levels of 8OHDG in DNA can be analysed by this method following a simple DNA digestion (Beland et al., 1979) and detection limits mean that starting concentrations of DNA can be kept to a minimum, an important consideration when dealing with biological samples. The assay can also be applied to measuring urinary output of the adduct because during normal cellular metabolism faulty DNA bases are excised from DNA and are excreted in the urine (Cundy et al., 1988). 8OHDG has been used as an end point to characterise, in vitro, oxidative damage to DNA during exposure to ultraviolet light (Floyd et al., 1988), X-rays (Kasai et al., 1984), gamma irradiation (Kasai et al., 1986) and hydrogen peroxide (Kasai and Nishimura, 1984a). Effects of chemicals, such as asbestos, on DNA have also been investigated and shown to produce this altered base adduct (Kasai and Nishimura, 1984b). Excretion of 8OHDG in urine has been monitored in patients with chronic gra-nulocytic leukemia who undergo total body irradiation as part of the disease therapy. Fig. 2b shows urinary output of the altered base adduct immediately after a splenic boost of irradiation and at intervals during and after the total body irradiation. Levels of 8OHDG are significantly increased following the splenic boost and also after treatment of the whole body. The level does not remain elevated, however, and falls towards the normal range twelve hours after treatment. This study provides evidence of a repair mechanism for the altered base adduct in an in vivo situation. If inflammatory disorders do give rise to DNA damage in vivo, as might be predicted from the observations that ROS are produced in elevated quantities and are known from in vitro studies to alter DNA in such a way that polyclonal and monoclonal antibody binding is altered, it might be expected that 8OHDG would be produced in greater amounts in patients with inflammatory conditions.
Reactive Oxygen Species Damage to DNA
99
251
i
20
UI
~'~ e~ .c-e 0
"~"
/ /
• Human DNA + thiourea (50mM) ta Human DNA + ethanol (10mM)
/ -
~cI I " 100 200 300 400 Dose of gamma radiation (Gray)
--I
I 500
Fig. 2a. Dose-dependent increase in 8-hydroxydeoxyguanosine formation in response to gamma irradiation ( O ). Inhibition by hydroxyl radical scavengers, 50 mM thiourea ( • ) and 10 mM ethanol (13).
20 18 c
le 14 12 .-r o
16
10 Gy Tota,
lOGy
/
\
body /
\
E e
0
1/ 1; 2/ 2/ /4 /
/
I
/ 8'~10
I
I
I
I
I
210 1 14 ~ 4I
I
/
I
8I
Time (hrs) Fig. 2b. Serial detection of 8-hydroxydeoxyguanosine (8OHDG) in the urine of a patient with chronic granulocytic leukemia undergoing a splenic boost of 10 Gray and total body irradiation therapy of a further 10 Gray over 200 minutes.
I
I
12
100
S. Blount et al.
o
LLI -J
--
t,cr~
.=. ~"
J
.==o~
< hr.
"~= r
J
°~
_~°--= oo
0 Z
io
AU..J - e~UO(:ISeJ G~)3
Reactive Oxygen Species Damage to DNA
101
Activatedphagocyte
(
~
Superoxide anion
Hydrogenperoxide ymphocyte ell Susceptibilityrange CD8+ > CD4+> B
Hydrogenperoxide +
DNA
\ Altered
,~~ucleoi~d Clastogenic expansion factors
/
utation ) Excretitionof faulty bases
Circulating immune complex~ (antigen) SLErenaldeposition
Cell death- DNArelease I Enzymaticbreakdown
Urinary "--~-----~---~excretition
Antibodiesraised
Fig. 4. Schematic diagram of the role of ROS, produced in inflammatory conditions, in the disease state SLE. Production of 8OHDG was measured in patients suffering from SLE and RA. Urine samples were analysed by HPLC with electrochemical detection. After establishing a normal range of excretion for 8OHDG, levels in the two patient groups were analysed. It was found that in the rheumatoid group, levels of the DNA base adduct were elevated when compared with the control group but in contrast, in the SLE group there was no detectable excretion of 8OHDG (Fig. 3). There are several explanations why 8OHDG is not excreted in detectable quantities in SLE patients. It may be that the excision repair process which would normally remove this adduct from DNA is faulty or absent. There is evidence that SLE patients have a defective repair mechanism for removing the mutagenic
102
S. Blount et al.
adduct O6-methylguanine (Harris et al., 1982). Alternatively, DNA excised from cells may not be cleared effectively because of low complement levels and persist in the circulation in the form of immune complexes rather than being cleared and excreted in urine. Analysis of the content of circulating immune complexes isolated from SLE sera would tend to suggest that both mechanisms may be involved. Firstly, DNA from the immune complexes was digested and the level of 8OHDG was measured. It was found to be significantly elevated when compared to the levels produced in normal urine and also in excess of the levels found in the rheumatoid group. However, DNA isolated from immune complexes is about 20 kilobases in size which would imply that once 8OHDG is formed, it remains in the DNA.
Role of Free Radical Damage to DNA in Disease Processes Development of altered DNA bases as a result of inflammatory derived ROS may have important consequences. It has been suggested that oxygen radicals released during the respiratory burst are responsible for the development of cancer in inflamed tissues (Cerutti et al., 1985). 8OHDG is known to be mutagenic causing faulty base pairing not only at its location but its presence also affects the pairing of neighbouring pyrimidine residues both upstream and downstream of the adduct (Kuchino et al., 1987): Concentration of this adduct in DNA may result in cell death due to faulty functioning of mis-read genes and the DNA found in immune complexes may arise from cell breakdown. It is also possible that these fragments of DNA that have been exposed to ROS and have high levels of altered bases may form the immunogen against which antibodies to DNA are raised in SLE. Much of the work described here has related the effects of ROS on DNA to the disease state SLE (Fig. 4). The findings have supported some earlier theories concerning radical involvement in the denaturation of macromolecules and in the ability of antibodies to recognise and respond to these changes. It has also been possible to use effectively the measurement of 8OHDG as a marker of oxidative stress in this and related inflammatory disorders. The relationship of production of a mutagenic adduct such as 8OHDG by ROS with the induction of malignancy is clearly an important observation. The use of a sensitive assay to measure levels of this adduct should be useful for analysing DNA, isolated from cells, to investigate whether or not 8OHDG is repaired in cells of patients both with inflammatory and malignant conditions, and whether or not a build up of DNA adducts is sufficient to alter the state of a normal cell or if this is only one of many necessary injuries the cell must suffer.
References Aitcheson, C. and Tan, E. (1982). Antinuclear Antibodies. In: $cigntifi¢ Basis of Rh~umat01o~v (G.S. Panayi, ed.), Chap 6. Churchill, Livingstone. Allan, I.M., Lunec, J., Salmon, M. and Bacon, P.A. (1987). Reactive oxygen species selectively deplete normal T lymphocytes via a hydroxyl radical dependent mechanism. Scand. J. Immunol, 26, 47-53. Allan, I.M., Vaughan, A.T.M., Milner, A.E., Lunec, J. and Bacon, P.A. (1988). Structural damage to lymphocyte nuclei by H20 2 or irradiation is dependent on the mechanism of OH radical production. Br. J, Cancer 58.34-37. Amoma, O.I., Halliwell, B. and Dizdaroglu, M. (1989). Iron ion-dependent modification of bases in DNA by the superoxide radical-generating system hypoxanthine/xanthine oxidase. J. Biol. Chem. 264, 1302413028.
Reactive Oxygen Species Damage to DNA
103
Beland, F.A., Dooley, K.L. and Casciano, D.N. (1979). Rapid isolation of carcinogen-bound DNA and RNA by hydroxyapatite chromatography. J. Chromatogr, 174, 177-186. Blount, S., Griffiths, H.R. and Lunec, J. (1989). Reactive oxygen species induce antigenic changes in DNA. ~ , 100-104. Bruneau, C. and Benveniste, J. (1979). Circulating DNA : anti-DNA complexes in Systemic Lupus Erythematosus. J. Clin. Invest. 64, 191-198. Cundy, K.C., Kohen, R. and Ames, B.N. (1988). Determination of 8-hydroxydeoxyguanosine in human urine: a possible assay for in vivo oxidative DNA damage. In: Oxwen radicals in Biology and Medicine (M.G. Simic, K.A.Taylor, J.F. Ward and C. Von Sonntag, ed.), pp. z179-482. Plenum Press, New York. Cerutti, P.A. (1985). Prooxidant states and tumor promotion. Science 227, 375-381. Emerit, I., Khan, S.H. and Cerutti, P.A. (1985). Treatment of lymphocyte cultures with a hypoxanthinexanthine oxidase system induces the formation of transferable clastogenic material J. Free Radical Biol. Mgd, 1, 51-57. Filho, A.C.M. and Meneghini, R. (1984). In vivo formation of single strand breaks in DNA by H202 is mediated by the Haber Weiss reaction. Bi0qhim, Bioohvs. Acta 781, 56-63. Floyd, R.A., West, M.S., Eneff, K.L., Hogsett, W.E. and Tingey,D.T. (1988). Hydroxyl radical mediated formation of 8-hydroxyguanine in isolated DNA. Arch. Biochem. Biophys. 262, 266-272. Frenkel, K., Goldstein, M. and Teebor, G.W. (1981). Identification of the cis-thymine glycol moiety in chemically oxidised and gamma-irradiated DNA by HPLC analysis. BiQqhemistrv 20. 7566-7571. Frenkel, K., Cummings, A., Solomon, J., Cadet, J., Steinberg, J.J. and Teebor, J.W. (1985). Quantitative determination of 5-(hydroxymethyl)uracil moiety in the DNA of gamma-irradiated cells. Biochemistrv 2_4,4527-4533. Fridovich, I. (1978). The biology of oxygen radicals. Science 201,875-880. Griffiths, H.R., Unsworth, J., Blake, D.R. and Lunec, J. (1988). Oxidation of amino acids within serum proteins. In: Fr¢¢ R~diqals: Chemistry. Pathology and Medicine (C. Rice-Evans and T. Dormandy, ed.), pp. 439-454. Richelieu Press, London. Griffiths, H.R. and Lunec, J. (1988). Effect of polymorph derived oxidants on IgG in relation to rheumatoid factor binding. Scand. J. Rheum. Suppl. 75, 148-156. Gripenberg, M. and Helve, T. (1986). Anti-DNA antibodies of IgA class in patients with systemic lupus erythematosus. Rheumatol. Int. 6, 53-55. Gutteridge, J.M.C. (1984). Reactivity of hydroxyl and hydroxyl-like radicals discriminated 0by release of thiobarbituric acid-reactive material from deoxy sugars, nucleosides and benzoate. BiQchem. J, 224, 761767. Gutteridge, J.M.C. and Quinlan, G.J. (1983). Malondialdehyde formation from lipid peroxides in the thiobarbituric acid test: the role of lipid radicals, iron salts, and metal chelators. J. Appl. Biochem. 5, 293299. Harris, G., Asbery, L., Lawley, P.D., Denman, A.M. and Hytton, W. (1982). Defective repair of 06methylguanine in autoimmune diseases. Lance~ ii, 952-956.
104
S. Blount et aL
Johnson, W.C. Jr. (1978). Circular dichroism spectroscopy and the vacuum ultraviolet region. Ann. Rev. Phys. Chem. 29, 93-114. Kasai, H. and Nishimura, S. (1984a). Hydroxylation of deoxyguanosine at the C-8 position by ascorbic acid and other reducing agents. Nucleic Acids Res. 12.2137 -2145. Kasai, H. and Nishimura, S. (1984b). DNA damage induced by asbestos in the presence of hydrogen peroxide. Gann 75, 841-844. Kasai, H., Tanooka, H. and Nishimura, S. (1984). Formation of 8-hydroxyguanine residues in DNA by Xirradiation. Garm 75, 1037-1039. Kasai, H., Crain, P.F., Kuchino, Y., Nishimura, S., Ootsuyama, A. and Tanooka, H. (1986). Formation of 8hydroxyguanine moiety in cellular DNA by agents producing oxygen radicals and evidence for its repair. ~rcinog~nesis 7, 1849-1851. Kuchino, Y., Moil, F., Kasai, H., Inoue, H., Iwai, S., Miura, K., Ohtsuka, E. and Nishimura, S. (1987). Misreading of DNA templates containing 8-hydroxydeoxyguanosine at the modified base and at adjacent residues. Natur~ 327, 77-79. Logani, M.K. and Davies, R.E. (1980). Lipid oxidation : Biologic effects and antioxidants- a review. Science 15,485-495. Lunec, J., Blake, D.R., McCleary, S.J., Brailsford, S. and Bacon, P.A. (1985). Self perpetuating mechanism of immunoglobulin G aggregation in rheumatoid inflammation. J. Clin. Inve~t. 76, 2084-2089. Lunec, J., Griffiths, H.R. and Blake, D.R. (1987a). Oxygen radicals in inflammation. ISI Atlas Sci. 1,45-48. Lunec, J., Griffiths, H.R., Jones, A.F. and Blake, D.R. (1987b). Protein fluorescence and its relationship to free radical activity. In: F r ~ radicals. Oxidant S~ress and Drug Action (C. Rice-Evans, ed.), pp. 151-167. Richelieu Press, London. Massie, H.R., Samis, H.V. and Baird, M.B. (1972). The kinetics of degradation of DNA and RNA by H20 2. ]~iochim, Bioohys. AqIa 272,539-548. McCord, J.M. and Day~ E.D. (1978). Superoxide dependent production of hydroxyl radical catalysed by iron EDTA complex. FEBS letts. 86, 139-142. McCord, J.M. and Fildovich, I. (1969). Superoxide dismutase. An enzymic function for erythrocuprein (hemocuprein). J. Biol. Chem. 244. 604%6055. Milner, A.E., Vaughan, A.T.M. and Clark, I.P. (1987). Measurement of DNA damage in mammalian cells using flow cytometry. Radiat. Res. 110, 108-117. Notman, D.D., Kurata, M.D. and Tan, E.M. (1975). Profiles of antinuclear antibodies in systemic rheumatic diseases. Ann. Intern. Med. 8~, 464-469. Rhaese, H.J. and Freese, E. (1968). Chemical analysis of DNA alterations 1. Base liberation and backbone breakage of DNA and oligodeoxyadenylic acid by hydrogen peroxide and hydroxylamine. Bigqhim, Bionhvs. Acta 155,476-490. Root, R.K., Metcalf, J., Oshino, N. and Chance, B. (1975). H20 2 release from human phagocytes during phagocytosis. I. Documentation, quantification and some regulatory factors. J, ~lin. Invest. 55, 945-955.
Reactive Oxygen Species Damage to DNA
105
Stokes, R.P., Cordwell, A and Thompson, R.A. (1982). A simple, rapid ELISA method for the detection of DNA antibodies. J. Clin. Pathol. 35.566-573. Swaak, A.J.G., Groenwold, J., Aarden, L.A. and Feltkamp, T.E.W. (1981). Detection of anti-dsDNA as a diagnostic tool. Ann. Rheum. Dis. 40, 45-49. Tan, E.M., Cohen, A.S., Fries, J.F. et al. (1982). The 1982 revised criteria for the classification of Systemic Lupus Erythematosus. Arthritis Rheum. 2~, 1271-1277. Teebor, G.W., Boorstein, R.J. and Cadet, J. (1988). The repairability of oxidative free radical mediated damage to DNA: a review. Int, J. Radiat. Biol. 54. 131-150. Ward, J.F., Evans, J.W. Limoli, C.L. and Calabro-Jones, P.M. (1987). Radiation and hydrogen peroxide induced damage to DNA. Br. J. Cancer 55. Suppl. VIII, 105-112. Wolff, S.P., Gamer, A. and Dean, R.T. (1986). Free Radicals, lipids and protein degradation. TIBS 11, 27-31.
0098-2997/91 $0.00 + .50 © 1991 Pergamon Press plc.
REACTIVE OXYGEN SPECIES DAMAGE TO DNA AND ITS ROLE IN SYSTEMIC LUPUS ERYTHEMATOSUS S. Blount, H. R. Griffiths and J. Lunec Wolfson Research Laboratories, Department of Clinical Chemistry, Queen Elizabeth Medical Centre, Birmingham, B15 2 TH, U.K.
Systemic Lupus Erythematosus Systemic lupus erythematosus (SLE) is an autoimmune disease of unknown aetiology. It is characterised by a wide variety of clinical presentations, the most common features being arthritis, vasculitis, skin rashes and renal manifestations (Aitcheson and Tan, 1982). There is evidence of a marked disturbance of both cellular and humoral immunity, and the disease is notable for the wide variety of antibodies, especially antinuclear antibodies, found in the serum. SLE patients also have characteristically low complement levels which may play a role in the build up of circulating immune complexes found in the sera. Persistent production of excess autoantibodies and the subsequent formation of immune complexes are attributed to the severe organ lesions seen in this disease. In particular, deposition of immune complexes in the kidney basement membrane gives rise to the glomerulonephritis suffered by 50% of SLE patients (Bruneau and Benveniste, 1979). Diagnosis of SLE currently relies on the serological estimation of antibodies binding to doublestranded DNA in a radioimmunoassay using labelled DNA (Tan et al., 1982). Antibodies which bind to DNA are found in the sera of patients with Rheumatoid Athritis (RA) and other rheumatological disorders such as Sjogren's syndrome (Notman e t a l . , 1975). These bind preferentially to single-stranded DNA whereas antibodies binding to double-stranded DNA are most uniquely correlated with SLE and its periods of clinical activity, especially with its renal complications (Aitcheson and Tan, 1982).
Generation of Free Radicals Arthritis is a component of disease of both SLE and RA. It is characterised by a perpetuated inflammatory response and resultant tissue damage. During inflammation neutrophils, macrophages and eosinophils become activated by a variety of stimuli, such as immune complexes, and undergo a cyanide-independent respiratory burst and degranulation. During the respiratory burst oxygen consumption is increased and the membrane-bound NADPH oxidase is activated (Lunec et al., 1987a). This enzyme catalyses the reduction of molecular oxygen to superoxide by electron transfer from NADPH. Reduction of oxygen at the cell surface results in the superoxide
93
94
S. Blount et al.
radical being released into the extracellular space where it can dismutate to form hydrogen peroxide. The reaction is catalysed by superoxide dismutase - an enzyme found in eukaryotes to protect against the toxic effects of oxygen (McCord and Fridovich, 1969). Hydrogen peroxide, although itself not a radical, can react in the presence of metal ions to produce the highly reactive hydroxyl radical (McCord and Day, 1978). Hydrogen peroxide is an important metabolite in the conversion of oxygen to water via this route because it is lipophillic and thus has the ability to traverse cell membranes (Allan et al., 1988). The hydroxyl radical diffuses only 15/~ before reacting so the location of the hydrogen peroxide determines where the radical attack will take place (Ward et al., 1987). The hydroxyl radical can react with all cellular macromolecules including lipids (Logani and Davies, 1980), proteins (Wolff et al., 1986; Lunec et al., 1987b) and DNA (Filho and Meneghini, 1984). Reaction with lipid results in the formation of lipid peroxides, hydroperoxides and diene conjugates. Measurement of free radical damage to lipids can be assessed by thiobarbituric acid (TBA) reactive products (Gutteridge and Quinlan, 1983). Many different lipid breakdown products react with TBA and it is difficult to define and quantify one specific product of peroxidation.
Biological End Points of Free Radical Damage Hydroxyl radical damage to proteins can be measured by high performance liquid chromatography (HPLC) which characterises oxidation products of particular amino acids. Those most susceptible to this kind of damage include the aromatics (tryptophan, tyrosine and phenylalanine), histidine and the disulphide bond formed between cysteine molecules (Griffiths et al., 1988). Damage to DNA can be measured at the level of the individual bases (Teebor et al., 1988) or sugar molecules (Gutteridge, 1984) or by detecting changes in the overall conformational structure (Curtis Johnson, Jr, 1978). Several HPLC assays to measure altered DNA bases exist such as those for thymine glycol (Frenkel etal., 1981) and 5-hydroxymethyluracil (Frenkel etal., 1985). Recently, an assay for the measurement of 8-hydroxydeoxyguanosine (8OHDG) has been described which links the use of HPLC with electrochemical detection, thus increasing the sensitivity of the assay several fold (Kasai and Nishimura, 1984a). This product is formed in a hydroxyl radicalmediated reaction and is a definitive end point for measurement of oxidative damage to DNA (Aruoma et al., 1989).
Free Radicals in Inflammation As described above a perpetuated inflammatory response gives rise to release of highly reactive species into the extracellular milieu which may then enter cells via the hydrogen peroxide intermediate. In such a case as arthritis where inflammation is localised to a joint free radical products may exist in I.tM quantities and thus exert a damaging effect on surrounding macromolecules (Root et al., 1975). A detailed study of the effects of reactive oxygen species (ROS) on the protein IgG have been investigated and are described in detail elsewhere in this issue. Briefly, in Rheumatoid Arthritis antibodies are raised to the "self" molecule IgG. It was suggested that IgG may have been altered by free radical damage and that antibodies were in fact raised against an altered form of IgG (Lunec et al., 1985). Results of the experiments showed that both structural and effector functions were significantly altered as a result of free radical attack (Griffiths and Lunec, 1988). Extrapolating from this it was decided to investigate the effects of ROS on the macromolecule DNA. SLE is a chronic inflammatory disease in which antibodies binding to "self" DNA are found. Although the measurement of these antibodies has diagnostic value their role in the disease process is unclear. Certainly they are involved in the formation of immune complexes, responsible for tissue damage particularly in the kidney basement membranes. Since it had been shown that ROS,
Reactive Oxygen Species Damage to DNA
95
in physiologically generated concentrations appertaining to the inflamed joint, were able to induce changes to the IgG molecule, studies of the effects of ROS, as generated during inflammation, on DNA were initiated. Polyclonal antibodies found in SLE sera served as a diagnostic tool to investigate these effects. Antibodies in SLE sera bind both to double and single-stranded DNA but as previously mentioned titres of binding to double-stranded DNA are most uniquely correlated with disease activity. In view of this it is important to obtain a DNA sample for use as antigen which is in a pure form. Commercially bought double-stranded DNA contains some single-stranded contaminants which can be digested with the enzyme S1 nuclease which selectively cleaves bases from an unpaired strand. If this is then used as the immobilised antigen in an enzyme-linked immunosorbent assay (ELISA) antibodies which recognise single-stranded DNA will not interfere.
Methods for Measuring Free Radical-Induced DNA Damage ELISA To investigate the effects of binding of polyclonal antibodies found in SLE sera as a marker of ROS-mediated damage to DNA, purif.,ed double-stranded DNA was immobilised onto an ELISA plate firstly in its native form and after sufficient blocking of unreactive sites polyclonal SLE sera was added and binding was measured using peroxidase-conjugated goat anti-human antibodies to label serum IgG, IgA and IgM antibodies binding to the DNA. The level of binding was then compared with levels of binding achieved in a similar assay using double-stranded DNA that had been reacted with ROS. The DNA used was incubated with physiological concentrations of hydrogen peroxide in the presence of ascorbic acid, a reducing agent. No metal ions were added to the system and all reagents were treated with chelex resin. This was to ensure that the only metal ions available to catalyse the formation of the highly reactive hydroxyl radical were located on the DNA structure itself thereby creating site-specific damage rather than producing the radical species in free solution. A comparison of the binding of SLE antibodies to native DNA compared to the DNA incubated with reactive oxygen species showed that hydrogen peroxide mediated damage caused an increase in antibody binding (Blount etal., 1989). This effect was inhibited if the compounds desferrioxamine or thiourea were added to the DNA prior to its incubation with hydrogen peroxide and ascorbic acid. Since thiourea is a hydroxyl radical scavenger and desferrioxamine a metal ion chelator, the results would suggest that the hydroxyl radical can react with and alter the DNA structure. Also, that antibodies found in SLE sera bind in greater number only if metal ions found on the surface are available to react with the hydrogen peroxide and are not chelated, with desferrioxamine for example. During these studies the nature of the antibodies binding to DNA, both in its native and radicaldenatured form, was investigated. It was found that the predominant isotype binding to both types of DNA was IgG. However, when the changes induced by the hydroxyl radical were analysed it was clear that antibodies of the IgA isotype were able to detect hydrogen peroxide mediated changes to the DNA most sensitively (Fig. 1). This is of particular interest since antibodies of the IgA isotype have been described'to be sensitive to changes in disease activity in SLE, particularly relating to renal manifestations (Gripenberg and Helve, 1986). The changes induced by hydrogen peroxide and ascorbic acid in the DNA molecule have several implications. Firstly, the increase in binding of antibodies to a denatured form of DNA might suggest that in SLE some form of alteration of the DNA molecule does occur and it is this DNA that is the antigen against which anti-DNA antibodies are raised. These results are also supportive of the hypothesis of a perpetuated inflammatory response initiating macromolecular damage.
96
S. Blount et al.
Secondly, SLE sera antibodies have increased binding to a DNA that has been incubated with hydrogen peroxide and ascorbic acid in the presence of catalytic metal ions. This observation is relevant to the development of a better assay in which to measure anti-DNA antibodies.
0.40 -
IgA
0.39 0.36 -
\
0.37 E 0.36 _ E
._T_~ " ,.X,
0.35 0.34 t~ 0.33 O 0.32
\,
0.31 0.30 DNA+PBS control
2.0
E c-
DNA+ H202/AA
DNA + H202/AA + desferrioxarnine
2.0
IgM
1.8
1.8
1.6
1.6
1.4
q
IgG
T
/
1.4
T
O4 1.2 O~ 1 O
-
DNA + H202/AA + thiourea
e~ 1.2 o3 ,~r
q 0.6
0.8
0
0.6
0.6
0.4
0.4
0.2
0.2 DNA + PBS DNA + DNA + H202/AA H202/AA + desferrioxamine
DNA + H 202/AA + thiourea
)NA+PBS
..L..,._
DNA+ DNA + H202/AA H202/AA + desferrioxamine
S,"
DNA + H 202/AA + thiourea
Fig. 1. Differential binding of anti-DNA isotypes against reactive oxygen species denatured DNA. Binding of anti-IgG, IgM and IgA isotypes to double ( ~ ) and single (ff3) stranded DNA. At present the assays available for measuring anti-DNA antibodies use nauve double-stranded DNA in a radioimmunoassay (Stokes et al., 1982). Results obtained from this method are useful in the diagnosis of SLE. However, other rheumatoid disorders have antibodies which bind to DNA, particularly single-stranded DNA and these may create false positive results. Because of this it is essential to use a purified antigen of the single or double stranded form only. Even in doing this patients with RA have serum antibodies which bind to give a similar titre to mildly active SLE patients (Swaak et al., 1981). From the initial observations that antibodies in SLE sera bind better to D N A that had been reacted with reactive oxygen species it was decided to investigate the ability of antibodies found in rheumatoid sera to bind to DNA reacted with reactive oxygen species when
Reactive Oxygen Species Damage to DNA
97
compared to the native double-stranded DNA used in standard immunochemical and radioimmunoassay. It was found that whilst in SLE patients there was a significant increase in binding to the ROS-denatured DNA this trend was not observed in the rheumatoid patients. In the rheumatoid group, antibodies in individual sera were shown to either increase or decrease in binding to the ROS-denatured DNA (Blount et al., 1989). An ELISA assay was established using double-stranded DNA (reacted with hydrogen peroxide and ascorbic acid) as the immobilised antigen and it has been found to compare favourably with the standard radioimmunoassay used, with the advantage that it can assign an optical density value to each patient's sera analysed. This is most important for patients believed to be mildly active since fluctuations in antibody binding can easily be quantified to see if they relate to clinical activity. The radioimmunoassays used conventionally defined values below a defined cut-off point simply as negative titres. Within this group patients with mild SLE may go undetected. Using the ELISA assay with ROS-denatured-DNA results obtained may be more discriminating towards this group. The reaction of hydrogen peroxide and ascorbic acid with DNA produces changes in the macromolecule which can be detected using polyclonal SLE sera antibodies. The levels of reactive oxygen species generated in these experiments are consistent with those produced during an inflammatory response. Further investigation is needed to determine whether or not DNA damage plays a role in anti-DNA antibody production and also whether or not DNA from patients with SLE shows any signs of degradation by reactive oxygen species in vivo. This latter point will be discussed below. Activated granulocytes are not the only source of oxygen free radicals produced during normal oxidative metabolism. The superoxide radical, for example can also be produced by autoxidation of the drug detoxifying system, cytochrome P450, and the mitochondrial respiratory chain (Fridovich, 1978). In the presence of superoxide dismutase and metal ions this will react to form hydrogen peroxide and the hydroxyl radical as previously described. The hydroxyl radical can cause double- and single-strand breaks (Rhaese and Freese, 1968; Ward et al., 1987), base alterations (Massie et al., 1972) and damage to the sugar-phosphate moieties (Gutteridge, 1984). All these changes could affect germ-line genes and affect the highly accurate transcription of the DNA into proteins. It is thus important to monitor changes in DNA structure at several levels.
Cell Culture Experiments which investigated the effects of hydrogen peroxide on cell viability using a lymphocyte model were able to show that hydrogen peroxide added to cell culture medium caused a significant increase in cell death when compared to cells incubated in hydrogen peroxide-free medium (Allan et al., 1988). It was also shown that changes in nuclear DNA were involved in this process. This reaffirms the idea that hydrogen peroxide is able to gain access to nuclear DNA and, as in this case, cause structural changes. Nuclear DNA was analysed using a fluorescence activated cell sorter to study DNA nucleoids -.nuclear DNA stripped of histone proteins (Milner et al., 1987). After incubation with hydrogen peroxide nucleoid DNA showed significant expansion, which was inhibited by hydroxyl radical scavengers or metal ion chelators. It was also found that within the lymphocyte population, cell subsets were differentially susceptible to damage; CD8 ÷ lymphocytes were highly susceptible, B cells highly resistant, and CD4 ÷ lymphocytes were of intermediate susceptibility (Allan et al., 1987). Experiments using cells cultured with medium containing ROS have not only provided information about changes in cellular DNA, they have also investigated the hypothesis that during the radical chain reactions clastogenic factors may be produced which are able to cause secondary damage to
98
$. Blount et al.
compounds as a result of the initial free radical insult. Results of these studies have shown that ultrafillrates of cells in culture medium containing ROS can induce chromosomal aberrations and sister chromatid exchanges in lymphocytes from normal healthy donors (Emerit et al., 1985).
Monoelonal Antibodies Having established that gross structural damage to DNA occurs following exposure to hydrogen peroxide, either as a direct attack from hydroxyl radicals or from secondary clastogenic factors with stabilities significantly greater than those of the oxygen radicals, a more detailed investigation to identify these changes was carried out using a panel of monoclonal antibodies produced spontaneously from (NZB/NZW)F1 and MRL/MPlpr/lpr mice. Results of in vitro experiments showed that DNA previously incubated with hydrogen peroxide in the presence of a reducing agent and metal ions had: increased exposure of DNA bases, changes in binding to the sugar-phosphate moiety, and increased expression of conformational determinants when compared to native doublestranded DNA.
HPLC A further, and yet more specific way, of determining DNA damage is to measure a particular defined end product of oxidative damage such as an alteration in a base structure. As briefly mentioned earlier, assays to measure DNA bases commonly use HPLC methods. In the detection of 8OHDG, a product of deoxyguanosine formed in a hydroxyl radical mediated reaction (Fig. 2a), HLPC has been linked with electrochemical detection thereby increasing the sensitivity of the method. Femtomoles of base adduct can be detected (Floyd etal., 1988). Levels of 8OHDG in DNA can be analysed by this method following a simple DNA digestion (Beland et al., 1979) and detection limits mean that starting concentrations of DNA can be kept to a minimum, an important consideration when dealing with biological samples. The assay can also be applied to measuring urinary output of the adduct because during normal cellular metabolism faulty DNA bases are excised from DNA and are excreted in the urine (Cundy et al., 1988). 8OHDG has been used as an end point to characterise, in vitro, oxidative damage to DNA during exposure to ultraviolet light (Floyd et al., 1988), X-rays (Kasai et al., 1984), gamma irradiation (Kasai et al., 1986) and hydrogen peroxide (Kasai and Nishimura, 1984a). Effects of chemicals, such as asbestos, on DNA have also been investigated and shown to produce this altered base adduct (Kasai and Nishimura, 1984b). Excretion of 8OHDG in urine has been monitored in patients with chronic gra-nulocytic leukemia who undergo total body irradiation as part of the disease therapy. Fig. 2b shows urinary output of the altered base adduct immediately after a splenic boost of irradiation and at intervals during and after the total body irradiation. Levels of 8OHDG are significantly increased following the splenic boost and also after treatment of the whole body. The level does not remain elevated, however, and falls towards the normal range twelve hours after treatment. This study provides evidence of a repair mechanism for the altered base adduct in an in vivo situation. If inflammatory disorders do give rise to DNA damage in vivo, as might be predicted from the observations that ROS are produced in elevated quantities and are known from in vitro studies to alter DNA in such a way that polyclonal and monoclonal antibody binding is altered, it might be expected that 8OHDG would be produced in greater amounts in patients with inflammatory conditions.
Reactive Oxygen Species Damage to DNA
99
251
i
20
UI
~'~ e~ .c-e 0
"~"
/ /
• Human DNA + thiourea (50mM) ta Human DNA + ethanol (10mM)
/ -
~cI I " 100 200 300 400 Dose of gamma radiation (Gray)
--I
I 500
Fig. 2a. Dose-dependent increase in 8-hydroxydeoxyguanosine formation in response to gamma irradiation ( O ). Inhibition by hydroxyl radical scavengers, 50 mM thiourea ( • ) and 10 mM ethanol (13).
20 18 c
le 14 12 .-r o
16
10 Gy Tota,
lOGy
/
\
body /
\
E e
0
1/ 1; 2/ 2/ /4 /
/
I
/ 8'~10
I
I
I
I
I
210 1 14 ~ 4I
I
/
I
8I
Time (hrs) Fig. 2b. Serial detection of 8-hydroxydeoxyguanosine (8OHDG) in the urine of a patient with chronic granulocytic leukemia undergoing a splenic boost of 10 Gray and total body irradiation therapy of a further 10 Gray over 200 minutes.
I
I
12
100
S. Blount et al.
o
LLI -J
--
t,cr~
.=. ~"
J
.==o~
< hr.
"~= r
J
°~
_~°--= oo
0 Z
io
AU..J - e~UO(:ISeJ G~)3
Reactive Oxygen Species Damage to DNA
101
Activatedphagocyte
(
~
Superoxide anion
Hydrogenperoxide ymphocyte ell Susceptibilityrange CD8+ > CD4+> B
Hydrogenperoxide +
DNA
\ Altered
,~~ucleoi~d Clastogenic expansion factors
/
utation ) Excretitionof faulty bases
Circulating immune complex~ (antigen) SLErenaldeposition
Cell death- DNArelease I Enzymaticbreakdown
Urinary "--~-----~---~excretition
Antibodiesraised
Fig. 4. Schematic diagram of the role of ROS, produced in inflammatory conditions, in the disease state SLE. Production of 8OHDG was measured in patients suffering from SLE and RA. Urine samples were analysed by HPLC with electrochemical detection. After establishing a normal range of excretion for 8OHDG, levels in the two patient groups were analysed. It was found that in the rheumatoid group, levels of the DNA base adduct were elevated when compared with the control group but in contrast, in the SLE group there was no detectable excretion of 8OHDG (Fig. 3). There are several explanations why 8OHDG is not excreted in detectable quantities in SLE patients. It may be that the excision repair process which would normally remove this adduct from DNA is faulty or absent. There is evidence that SLE patients have a defective repair mechanism for removing the mutagenic
102
S. Blount et al.
adduct O6-methylguanine (Harris et al., 1982). Alternatively, DNA excised from cells may not be cleared effectively because of low complement levels and persist in the circulation in the form of immune complexes rather than being cleared and excreted in urine. Analysis of the content of circulating immune complexes isolated from SLE sera would tend to suggest that both mechanisms may be involved. Firstly, DNA from the immune complexes was digested and the level of 8OHDG was measured. It was found to be significantly elevated when compared to the levels produced in normal urine and also in excess of the levels found in the rheumatoid group. However, DNA isolated from immune complexes is about 20 kilobases in size which would imply that once 8OHDG is formed, it remains in the DNA.
Role of Free Radical Damage to DNA in Disease Processes Development of altered DNA bases as a result of inflammatory derived ROS may have important consequences. It has been suggested that oxygen radicals released during the respiratory burst are responsible for the development of cancer in inflamed tissues (Cerutti et al., 1985). 8OHDG is known to be mutagenic causing faulty base pairing not only at its location but its presence also affects the pairing of neighbouring pyrimidine residues both upstream and downstream of the adduct (Kuchino et al., 1987): Concentration of this adduct in DNA may result in cell death due to faulty functioning of mis-read genes and the DNA found in immune complexes may arise from cell breakdown. It is also possible that these fragments of DNA that have been exposed to ROS and have high levels of altered bases may form the immunogen against which antibodies to DNA are raised in SLE. Much of the work described here has related the effects of ROS on DNA to the disease state SLE (Fig. 4). The findings have supported some earlier theories concerning radical involvement in the denaturation of macromolecules and in the ability of antibodies to recognise and respond to these changes. It has also been possible to use effectively the measurement of 8OHDG as a marker of oxidative stress in this and related inflammatory disorders. The relationship of production of a mutagenic adduct such as 8OHDG by ROS with the induction of malignancy is clearly an important observation. The use of a sensitive assay to measure levels of this adduct should be useful for analysing DNA, isolated from cells, to investigate whether or not 8OHDG is repaired in cells of patients both with inflammatory and malignant conditions, and whether or not a build up of DNA adducts is sufficient to alter the state of a normal cell or if this is only one of many necessary injuries the cell must suffer.
References Aitcheson, C. and Tan, E. (1982). Antinuclear Antibodies. In: $cigntifi¢ Basis of Rh~umat01o~v (G.S. Panayi, ed.), Chap 6. Churchill, Livingstone. Allan, I.M., Lunec, J., Salmon, M. and Bacon, P.A. (1987). Reactive oxygen species selectively deplete normal T lymphocytes via a hydroxyl radical dependent mechanism. Scand. J. Immunol, 26, 47-53. Allan, I.M., Vaughan, A.T.M., Milner, A.E., Lunec, J. and Bacon, P.A. (1988). Structural damage to lymphocyte nuclei by H20 2 or irradiation is dependent on the mechanism of OH radical production. Br. J, Cancer 58.34-37. Amoma, O.I., Halliwell, B. and Dizdaroglu, M. (1989). Iron ion-dependent modification of bases in DNA by the superoxide radical-generating system hypoxanthine/xanthine oxidase. J. Biol. Chem. 264, 1302413028.
Reactive Oxygen Species Damage to DNA
103
Beland, F.A., Dooley, K.L. and Casciano, D.N. (1979). Rapid isolation of carcinogen-bound DNA and RNA by hydroxyapatite chromatography. J. Chromatogr, 174, 177-186. Blount, S., Griffiths, H.R. and Lunec, J. (1989). Reactive oxygen species induce antigenic changes in DNA. ~ , 100-104. Bruneau, C. and Benveniste, J. (1979). Circulating DNA : anti-DNA complexes in Systemic Lupus Erythematosus. J. Clin. Invest. 64, 191-198. Cundy, K.C., Kohen, R. and Ames, B.N. (1988). Determination of 8-hydroxydeoxyguanosine in human urine: a possible assay for in vivo oxidative DNA damage. In: Oxwen radicals in Biology and Medicine (M.G. Simic, K.A.Taylor, J.F. Ward and C. Von Sonntag, ed.), pp. z179-482. Plenum Press, New York. Cerutti, P.A. (1985). Prooxidant states and tumor promotion. Science 227, 375-381. Emerit, I., Khan, S.H. and Cerutti, P.A. (1985). Treatment of lymphocyte cultures with a hypoxanthinexanthine oxidase system induces the formation of transferable clastogenic material J. Free Radical Biol. Mgd, 1, 51-57. Filho, A.C.M. and Meneghini, R. (1984). In vivo formation of single strand breaks in DNA by H202 is mediated by the Haber Weiss reaction. Bi0qhim, Bioohvs. Acta 781, 56-63. Floyd, R.A., West, M.S., Eneff, K.L., Hogsett, W.E. and Tingey,D.T. (1988). Hydroxyl radical mediated formation of 8-hydroxyguanine in isolated DNA. Arch. Biochem. Biophys. 262, 266-272. Frenkel, K., Goldstein, M. and Teebor, G.W. (1981). Identification of the cis-thymine glycol moiety in chemically oxidised and gamma-irradiated DNA by HPLC analysis. BiQqhemistrv 20. 7566-7571. Frenkel, K., Cummings, A., Solomon, J., Cadet, J., Steinberg, J.J. and Teebor, J.W. (1985). Quantitative determination of 5-(hydroxymethyl)uracil moiety in the DNA of gamma-irradiated cells. Biochemistrv 2_4,4527-4533. Fridovich, I. (1978). The biology of oxygen radicals. Science 201,875-880. Griffiths, H.R., Unsworth, J., Blake, D.R. and Lunec, J. (1988). Oxidation of amino acids within serum proteins. In: Fr¢¢ R~diqals: Chemistry. Pathology and Medicine (C. Rice-Evans and T. Dormandy, ed.), pp. 439-454. Richelieu Press, London. Griffiths, H.R. and Lunec, J. (1988). Effect of polymorph derived oxidants on IgG in relation to rheumatoid factor binding. Scand. J. Rheum. Suppl. 75, 148-156. Gripenberg, M. and Helve, T. (1986). Anti-DNA antibodies of IgA class in patients with systemic lupus erythematosus. Rheumatol. Int. 6, 53-55. Gutteridge, J.M.C. (1984). Reactivity of hydroxyl and hydroxyl-like radicals discriminated 0by release of thiobarbituric acid-reactive material from deoxy sugars, nucleosides and benzoate. BiQchem. J, 224, 761767. Gutteridge, J.M.C. and Quinlan, G.J. (1983). Malondialdehyde formation from lipid peroxides in the thiobarbituric acid test: the role of lipid radicals, iron salts, and metal chelators. J. Appl. Biochem. 5, 293299. Harris, G., Asbery, L., Lawley, P.D., Denman, A.M. and Hytton, W. (1982). Defective repair of 06methylguanine in autoimmune diseases. Lance~ ii, 952-956.
104
S. Blount et aL
Johnson, W.C. Jr. (1978). Circular dichroism spectroscopy and the vacuum ultraviolet region. Ann. Rev. Phys. Chem. 29, 93-114. Kasai, H. and Nishimura, S. (1984a). Hydroxylation of deoxyguanosine at the C-8 position by ascorbic acid and other reducing agents. Nucleic Acids Res. 12.2137 -2145. Kasai, H. and Nishimura, S. (1984b). DNA damage induced by asbestos in the presence of hydrogen peroxide. Gann 75, 841-844. Kasai, H., Tanooka, H. and Nishimura, S. (1984). Formation of 8-hydroxyguanine residues in DNA by Xirradiation. Garm 75, 1037-1039. Kasai, H., Crain, P.F., Kuchino, Y., Nishimura, S., Ootsuyama, A. and Tanooka, H. (1986). Formation of 8hydroxyguanine moiety in cellular DNA by agents producing oxygen radicals and evidence for its repair. ~rcinog~nesis 7, 1849-1851. Kuchino, Y., Moil, F., Kasai, H., Inoue, H., Iwai, S., Miura, K., Ohtsuka, E. and Nishimura, S. (1987). Misreading of DNA templates containing 8-hydroxydeoxyguanosine at the modified base and at adjacent residues. Natur~ 327, 77-79. Logani, M.K. and Davies, R.E. (1980). Lipid oxidation : Biologic effects and antioxidants- a review. Science 15,485-495. Lunec, J., Blake, D.R., McCleary, S.J., Brailsford, S. and Bacon, P.A. (1985). Self perpetuating mechanism of immunoglobulin G aggregation in rheumatoid inflammation. J. Clin. Inve~t. 76, 2084-2089. Lunec, J., Griffiths, H.R. and Blake, D.R. (1987a). Oxygen radicals in inflammation. ISI Atlas Sci. 1,45-48. Lunec, J., Griffiths, H.R., Jones, A.F. and Blake, D.R. (1987b). Protein fluorescence and its relationship to free radical activity. In: F r ~ radicals. Oxidant S~ress and Drug Action (C. Rice-Evans, ed.), pp. 151-167. Richelieu Press, London. Massie, H.R., Samis, H.V. and Baird, M.B. (1972). The kinetics of degradation of DNA and RNA by H20 2. ]~iochim, Bioohys. AqIa 272,539-548. McCord, J.M. and Day~ E.D. (1978). Superoxide dependent production of hydroxyl radical catalysed by iron EDTA complex. FEBS letts. 86, 139-142. McCord, J.M. and Fildovich, I. (1969). Superoxide dismutase. An enzymic function for erythrocuprein (hemocuprein). J. Biol. Chem. 244. 604%6055. Milner, A.E., Vaughan, A.T.M. and Clark, I.P. (1987). Measurement of DNA damage in mammalian cells using flow cytometry. Radiat. Res. 110, 108-117. Notman, D.D., Kurata, M.D. and Tan, E.M. (1975). Profiles of antinuclear antibodies in systemic rheumatic diseases. Ann. Intern. Med. 8~, 464-469. Rhaese, H.J. and Freese, E. (1968). Chemical analysis of DNA alterations 1. Base liberation and backbone breakage of DNA and oligodeoxyadenylic acid by hydrogen peroxide and hydroxylamine. Bigqhim, Bionhvs. Acta 155,476-490. Root, R.K., Metcalf, J., Oshino, N. and Chance, B. (1975). H20 2 release from human phagocytes during phagocytosis. I. Documentation, quantification and some regulatory factors. J, ~lin. Invest. 55, 945-955.
Reactive Oxygen Species Damage to DNA
105
Stokes, R.P., Cordwell, A and Thompson, R.A. (1982). A simple, rapid ELISA method for the detection of DNA antibodies. J. Clin. Pathol. 35.566-573. Swaak, A.J.G., Groenwold, J., Aarden, L.A. and Feltkamp, T.E.W. (1981). Detection of anti-dsDNA as a diagnostic tool. Ann. Rheum. Dis. 40, 45-49. Tan, E.M., Cohen, A.S., Fries, J.F. et al. (1982). The 1982 revised criteria for the classification of Systemic Lupus Erythematosus. Arthritis Rheum. 2~, 1271-1277. Teebor, G.W., Boorstein, R.J. and Cadet, J. (1988). The repairability of oxidative free radical mediated damage to DNA: a review. Int, J. Radiat. Biol. 54. 131-150. Ward, J.F., Evans, J.W. Limoli, C.L. and Calabro-Jones, P.M. (1987). Radiation and hydrogen peroxide induced damage to DNA. Br. J. Cancer 55. Suppl. VIII, 105-112. Wolff, S.P., Gamer, A. and Dean, R.T. (1986). Free Radicals, lipids and protein degradation. TIBS 11, 27-31.
