leucocnytesconfirmatory evidence for an eventual improvement. Key words: rheumatoid arthritis, redox status, superoxide dismutase, caeruloplasmin. The gold compound auranofin (1-thio-,B-D-glucopyranosato)

Annals of the Rheumatic Diseases, 1986; 45, 839-846

Early and late changes in sulphydryl group andcopper protein concentrations and activities duringdrug treatment with aurothiomalate and auranofinK J RAE,' C N N MACKAY,' C J McNEIL,' D H BROWN,' W E SMITH,'D LEWIS,2 AND H A CAPELL2

From the 'Department ofPure andApplied Chemistry, University ofStrathclyde, Glasgow; and the 2CentreforRheumatic Diseases, Baird Street, Glasgow

SUMMARY Superoxide dismutase activity (SOD), plasma and lysate thiol concentrations (PSHand LSH), and caeruloplasmin oxidase activity (CP) reflect the underlying reduction-oxidationimbalance associated with rheumatoid arthritis (RA), and are believed to be involved in theprotection of the cell against free radical activity. The early and late changes in these parametershave been observed and compared with standard clinical and biochemical assessments of diseaseactivity in 90 patients with active RA, randomly assigned to receive either sodium aurothioma-late, auranofin, or auranofin placebo. An index based on clinical criteria was used to identifypatients as responders or non-responders after 24 weeks of therapy. In the first six weeks oftreatment a change in SOD activity and LSH concentration in a direction away from controls wasfollowed by a return towards control levels in responders only. This suggests that in RA evidenceof clinical improvement induced by gold drugs is preceded by an initial biochemical response inan inflammatory direction. The extracellular parameters PSH and CP did not show the sameearly response, but PSH levels in responders showed a slower change towards normal values,though at no time were values obtained that might suggest a complete remission. Thus theintracellular parameters appear to reflect an early effect of the drugs on cells which may possiblybe of use in predicting the outcome of therapy, whereas the extracellular parameters provideconfirmatory evidence for an eventual improvement.

Key words: rheumatoid arthritis, redox status, superoxide dismutase, caeruloplasmin.

The gold compound auranofin (1-thio-,B-D-glucopyranosato) (triethylphosphine) gold 2, 3, 4,6-tetra-acetate) (AF) is orally absorbed and appearsto be effective in the treatment of rheumatoidarthritis.1 2 It is at present being comprehensivelyassessed in controlled clinical trials and comparedwith the injectable drug, sodium aurothiomalate(GST or Myocrisin); it produces a different profileof clinical and biochemical activities.' Both GSTand AF metabolise rapidly in vivo to produce anumber of new gold compounds distributed inserum, cells, and most tissues. In the main, gold invivo is complexed as gold(I) through the formation

Accepted for publication 21 April 1986.Correspondence to Dr W E Smith, Department of Pure andApplied Chemistry, University of Strathclyde, Glasgow Gl 1XL.

of gold-sulphur bonds with accessible sulphydrylgroups of proteins, peptides, and amino acids. 4

In rheumatoid arthritis the plasma sulphydrylgroup concentration is low and the erythrocytelysate sulphydryl group concentration high com-pared with those of normal controls. This indicates achange in the degree of oxidation of the fluids oneach side of the cell membrane.5 The change in thereduction-oxidation (redox) status of the sulphydrylsystem is reinforced by the fact that the concentra-tion of the main plasma oxidase, the copper contain-ing protein caeruloplasmin, is raised in the diseasestate6 7 and the activity of superoxide dismutase, anintracellular copper protein involved in the controlof the concentration of superoxide ion8 (a reducedform of oxygen), is decreased.9 10The activities of the copper proteins and the

839

copyright. on M

arch 20, 2020 by guest. Protected by

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.45.10.839 on 1 O

ctober 1986. Dow

nloaded from

http://ard.bmj.com/

840 Rae, Mackay, McNeil, Brown, Smith, Lewis, Capell

sulphydryl group concentrations are believed to belinked. Caeruloplasmin is a specific oxidase forsmall molecules containing the sulphydryl group,such as cysteine and glutathione, producing thecorresponding disulphides."1 By exchange with pro-tein sulphydryl groups these disulphides are be-lieved to influence plasma sulphydryl groupconcentrations.5 A close correlation between super-oxide dismutase activity and glutathione concentra-tion has also been shown in vivo and ex vivo usinghaemolysate.12As part of the phagocytic process a burst of free

radical activity is produced by the phagocytosingcell. Among other effects this activity causes therelease of oxygen derived species such as superoxideions, hydroxyl radicals, and hydrogen peroxide.13Protection against these active oxygen species isafforded by glutathione14 and possibly caerulo-plasmin15 as well as superoxide dismutase. It hasbeen suggested that the ability of patients withRA to metabolise these oxygen species is impairedand that the extent to which this occurs is related tothe severity of the disease. Furthermore, SODactivity in polymorphonuclear leucocytes of childrenwith juvenile rheumatoid arthritis is low9 and thelevels in leucocnytes of adults with rheumatoidarthritis is high.Thus both because of the possible chemical action

of the gold drugs on sulphydryl group activities andbecause of the role of these groups and copperproteins in defining redox status and protecting thecell against free radical attack it seemed of value tostudy and compare the changes in the sulphydrylgroup concentration and copper enzyme activitycaused by AF and GST.

Patients and methods

Ninety patients (28 male, 62 female) with classical ordefinite RA according to the criteria of the Amer-ican Rheumatism Association were studied during24 weeks of chrysotherapy. All patients werereceiving optimal doses of non-steroidal anti-inflammatory drugs (NSAIDs) and none weretaking steroids or had previously been treated withgold in any form. In addition, none had receivedpenicillamine, levamisole, or immunosuppressivedrugs in the six months preceding the trial.

Initially, patients were randomly allocated to oneof three treatment groups. Thirty patients (ninemale, 21 female, median age 47 years, range 25-69years, median duration of disease six years) weregiven an initial 10 mg injection of GST followed by50 mg weekly until clinical response. Thereafter, thefrequency of injection was reduced to fortnightly,three weekly, and, ultimately, four weekly. Thirty

patients (eight male, 22 female, median age 55years, range 22-71 years, median duration of disease6-5 years) received auranofin tablets (3 mg twicedaily) throughout the study. Thirty patients ( 11male, 19 female, median age 55 years, range 31-72years, median duration of disease 5-3 years) weregiven placebo tablets identical with auranofin twicedaily.

Patients were initially monitored weekly, andassessments of disease activity were performed atweeks 0, 3, 6, 12, and 24 of treatment. Theparameters studied were: duration of morning stiff-ness recorded in minutes, pain score measured incentimetres on a 10 cm visual analogue scale, gripstrength measured with an aneroid manometerattached to a small cuff and inflated to 20 mm ofmercury, Ritchie articular index,17 haemoglobinconcentration (Hb), and Westergren erythrocytesedimentation rate (ESR).Blood samples from 22 healthy volunteers (17

male, five female) were used to determine controlvalues for comparison with values found for thepatient population. None were taking NSAIDs,steroids, or had previously been treated with gold inany form.The clinical indices of disease activity were

compared with laboratory measurements of ex-tracellular redox status-namely, caeruloplasminoxidase activity and plasma thiol and intracellularredox status in erythrocytes-namely, superoxidedismutase activity and lysate thiol concentrations.All examinations were made on fresh bloodsamples, all were separated within four hours ofsampling, and all estimations were carried out atstandard times after separation and within 48 hoursof sampling.The methods of estimation have been described

previously.5 10 Thiol concentrations in lysate and inplasma were estimated by the method of Ellman18using the thiol-disulphide interchange reaction be-tween 5,5'-dithiobis(2-nitrobenzoic acid) and biolo-gical thiols. Caeruloplasmin oxidase activity wasmeasured by a modification of the method ofMenden et a119 based on the caeruloplasmin cat-alysed oxidation of p-phenylenediamine to Band-rowski's base. Superoxide dismutase activity wasmeasured by the method of Misra and Fridovich20based on the increase in the rate ofphoto-oxidation ofo-dianisidine. The overall disease activity wasassessed with the method of Mallya and Mace."1 Thisindex defined four grades of disease activity frominactive (grade 1) to very active (grade 4) using acombination of morning stiffness, pain score, gripstrength, articular index, Hb concentration, andESR. Clinical improvement was defined as a de-crease in activity grade by at least one grade after 24

copyright. on M


http://ard.bmj.com

/A

nn Rheum


ctober 1986. Dow

nloaded from

http://ard.bmj.com/

Sulphydryl group and copper protein changes during drug treatment 841

weeks of treatment. The significance of the differ-ence between normal values and patient populationswas calculated by the Mann-Whitney U test and thedifference between the treatment groups by theKruskal-Wallis test. The significance of the changein all parameters at various stages of treatment wasobtained with the Wilcoxon matched pairs signedrank test and the time dependent changes in SODand LSH by a Spearman rank correlation.22

Results

The three treatment groups were similar in terms ofage, duration of disease, sex ratio, and initial clinicaland laboratory variables. The method of assessingdisease activity as described by Mallya and Maceshowed that the majority of patients (70/90) had

Table 1 Change in disease activity grade during the first 24weeks of therapy with GST, AF, or placebo in patients withrheumatoid arthritis

Treatment Responders Non-respondersgroup

Number Improved Deteriorated Unchangedstill on one grade one gradetherapy n % n % n %

GST 23 14 61 0 0 9 39AF 26 8 31 2 8 16 62Placebo 17 3 18 3 18 11 65

moderately active disease, 14/90 had slightly active,and 6/90 very active disease.

After 24 weeks of therapy 23/30 patients (77%)continued GST, 26/30 (87%) continued AF, and17/30 (57%) continued placebo. Ten of the with-drawals from the placebo group were for persistentdisease activity and lack of drug effect. Adverseeffects or intercurrent illness accounted for all otherpatients who discontinued.

CLINICAL ASSESSMENTComparison of 0 to 24 week results (Wilcoxon)showed significant improvements in duration ofmorning stiffness, pain score, grip strength, Ritchiearticular index, and ESR for both GST and AFgroups. Patients who remained on placebo showedno significant improvement in any parameter. Table1 shows the proportion of patients in each groupwho either improved or deteriorated by at least onegrade of the activity index. When an improvementof one grade was used as a definition of change indisease activity there were 25 responders (R) and 41non-responders (NR) in the first six months oftherapy.

REDOX ASSESSMENTRedox parameters were all significantly differentfrom normal at the start of the trial (p<0-001Mann-Whitney). Table 2 shows the median andinterquartile range of redox parameters for all

Table 2 Medians and interquartile (IQ) ranges of redox parameters for patients with rheumatoid arthritis and normalcontrols at the beginning of the trial

Extracellular Intracellular

CP (mg/l00 ml)* PSH (pmol/l) SOD (ig/ml) LSH (tunolll)

Median 60-1 311 62-0 423IQ range 47-0-74-3 266-5-361 44-0-85-0 264-5-564-5Normal range 38-0-52-8 457-543 78-2-92-5 123-363

*SI conversion: caeruloplasmin mg/100 mlxlO=mg/l.

Table 3 Medians (M) and interquartile ranges (IQ) for the redox parameters during the first 24 weeks of therapy for eachof the treatment groups*

GST AF Placebo

Week 0 (n=29) Week 24 (n=21) Week 0 (n=29) Week 24 (n=25) Week 0 (n=29) Week 24 (n=17)

CP M 64-0 50-5 54-8 45-1 62-1 52-2IQ 51-1-73-8 41-9-67-7 46-0-75-0 32-3-64-4 48-9-74-9 38-6-67-4

PSH M 320 357 326 338 287 330IQ 272-362 325-416 269-372 296-396 250-349 288-380

SOD M 50-3 71-4 66-9 67-7 66-9 74-4IQ 36-7-69-9 43-5-91-1 39-7-85-0 51-1-85-0 50-3-92-6 63-9-96-3

LSH M 465 460 507 462 340 395IQ 339-638 307-618 257-584 280-533 232-442 310-417

*Differences in numbers between Tables 1 and 2 and Table 3 are due to instrumental problems with one batch of samples.

copyright. on M


http://ard.bmj.com

/A

nn Rheum


ctober 1986. Dow

nloaded from

http://ard.bmj.com/


patients compared with those for normal controls. Itcan be seen that CP and LSH are raised and SODand PSH are low. In addition, there were nosignificant differences in any of these parametersbetween the three treatment groups (Kruskal-Wallis) and none which correlated with age, sex, orduration of disease.Table 3 shows the changes in redox parameters

between 0 and 24 weeks for the three treatments.The wide interquartile ranges in these groups meanthat many apparent differences were not significanteven at the 5% level, although if the direction ofchange in all groups was the same, a significantchange could be observed if the groups werecombined. CP activity decreased towards normalcontrol values in all groups but did not reachsignificance at the 5% level in any. PSH rosetowards normal control values in all groups,reaching significance in the GST and AF respondergroups (p<0-02 and p<0-025 respectively). SODactivity also appeared to rise towards control levelsin all groups but was significant only in the GSTresponder group (p<0-01). LSH rose away fromnormal control values in the placebo group (p<002)and also in the GST responder and AF non-responder groups; LSH fell in both the AF respon-der (p<0-01) and GST non-responder groups. Withthe exception of LSH all the changes were asexpected from the analysis of complete treatmentgroups, with the largest changes being in the GSTand AF improvers groups. The significant changesare listed in Table 4. Neither the GST nor AFnon-responder groups showed any significantchange and the only significant change in theplacebo group was in the opposite sense, i.e. awayfrom normal control values.

It was noted that the intracellular redox para-meters (LSH and SOD) altered in the first fewweeks of therapy, and consequently, all four para-meters were examined graphically and the resultsfor SOD and LSH are illustrated for the GST group(Figs la and lb). The significance of these changeswas examined (Wilcoxon) by comparing the results

at each time point with those at week 0 (Table 5).The improvement in PSH at 12 and 24 weeks in theresponder group and the changes in SOD and LSHare the most significant changes, together with anunexpected and unexplained observation of a signi-ficant change in CP in the non-responder group atsix weeks. As can be seen in Figs la and lb thechanges in SOD and LSH at about six weeks are in

' ' ''^100- GST NR

87O-j40J0

0 3 6 12 24

Fig. 1 Changes in redox parameters for GST (a)responders and (b) non-responders. Thefigure showsmedian and interquartile ranges at time pointsfrom zero to24 weeks.

Table 4 Significant changes in redox parameters between 0 and 24 weeks. All changes are towards normal values unlessmarked*

GST AF Placebo

Rt (n=14) NRt (n=8) R (n=8) NR (n=17) R (n=4) NR (n=13)

CP - - <0-10 - -PSH <0-02 - <0-025 - -SOD <0-01 - - - -

LSH - - <0-01 - - <0-02*

tR=responder; NR=non-responder.

copyright. on M


http://ard.bmj.com

/A

nn Rheum


ctober 1986. Dow

nloaded from

http://ard.bmj.com/


opposite senses for the responder and non-responder groups, with the responders showing achange away from the control values. Similarchanges in the auranofin group, though consistentwith these findings, were not statistically significant,and no trend was observed in the placebo group,except that LSH rose continually.We have previously shown a relation between

SOD and LSH. This relation was found to bepresent at both 0 and 24 weeks (Table 6), butgraphical presentation of the results indicated somechange in slope, which suggested that both golddrugs did affect the nature of the correlation. Thechange in SOD v the change in LSH was plcted(Fig. 2), and it was evident that a ASOD v a ALSH

correlation was present in the placebo group(p<O-005). This correlation, however, was not asclearly defined in the AF or GST groups.

Discussion

The criterion for inclusion of a patient in the trialwas a clinical assessment of active disease and, asindicated by the large variation at the start of thetrial, there was no attempt to preselect specificgroups of patients on the basis of any individualparameter. The Mallya and Mace index proved auseful measure of the clinical efficacy of the drugs,which could then be used to assess the copperprotein and sulphydryl group measurements. It

Table 5 Probability values from an analysis using a Wilcoxon test of the change in CP, PSH, SOD, and LSH between 0and 3, 6, 12, and 24 weeks for responders and non-responders after treatment with GST

Week Responders (n= 14) Non-responders (n=9)

CP PSH SOD LSH CP PSH SOD LSH

0-3 03 0.1 08 0.5 0-4 0-2 02 070-6 07 0.1 004 008 0o05 0-6 0-06 0-040-12 0-6 0-06 0-7 0-3 0-8 0-4 0-5 0-40-24 0o5 0-03 0-1 0-1 0-2 0.1 0-7 0-9

Table 6 Correlation between SOD and LSH for each of the three treatment groups

Week GST AF Placebo

n Corr Sig n Corr Sig n Corr Sig

0 27 -0-705 0-0005 26 -0-762 0-0005 27 -0-697 0o000524 20 -0-827 0o005 23 -0-598 0-0025 17 -0-758 0-0005

Placebo

d*aS

0 0*.

0

0S

-25 0 +25ASOD (Mg/ml)

AF

* 0S

OS0

0

.0

a aI I

-25 0 +25A SOD (,ug /ml)

GST

0

0

*O 00

0 .0

0

0

00

0I

-25 0 +25

ASOD (jug/ml)Fig. 2 Plot ofthe change in SOD against the change in LSH over 24 weeks. The gold drugs appear to affect the correlationbetween SOD and LSH.

+250-

o-

Itn-j<-250-

.

.

copyright. on M


http://ard.bmj.com

/A

nn Rheum


ctober 1986. Dow

nloaded from

http://ard.bmj.com/


should be stressed that the patients in theplacebo treated group who remained on the therapyfor six months are atypical of the group as a whole,as those with the most active disease had droppedout and are therefore not included in any statisticalanalysis of change with time. In addition, the highdrop out rate of patients receiving placebo waslargely because of worsening disease, and thetendency for some improvement to occur in theremaining placebo group is not unexpected. Thedeterioration in LSH even in this selected group,however, indicates that there is a difference in thebiochemistry of this improvement compared withthat produced by GST and AF. All patients in thisstudy continued their regular daily intake of non-steroidal anti-inflammatory drugs and as far aspossible this was not varied during the trial, thoughit is possible that those who improved may havereduced the dosage. It is also possible that somenon-responders are patients whom without treat-ment with GST or AF would have deterioratedrather than not improved. Both these points wouldreduce the significance of differences observedduring therapy and may mean that the changesdiscussed are more conservatively estimated thanshould be the case. Neither possibility shouldproduce any fake positive result. Furthermore,there was more flexibility in the GST regimen thanwith the new drug, in that at a later stage in therapythe dosage was where possible reduced. Since GSTstill shows the larger effects on the time scale studiedthe comparisons made remain valid, particularlysince there is no accurate way of comparing dosageschedules between such different drugs.

Division of patients on each treatment intoresponders and non-responders allowed assessmentof the biochemical parameters, not only longitudi-nally on the basis of the assigned treatment groupbut also on the basis of the response to treatment. Inthe three treatment groups the significant changes inGST and AF groups after 24 weeks were confined tothe responder groups and were in a directiontowards normal control values, whereas the onlysignificant change in the placebo group was awayfrom normal (Table 4). Despite clinical improve-ment in some patients, however, the inorganicbiochemical parameters did not return to normallevels. This suggests that the effect of therapy onthat aspect of the underlying biochemistry investi-gated is to create changes in the in vivo activitywhich produce a more quiescent state withoutreturning the underlying biochemistry to a normalnon-rheumatoid situation or producing a completeremission.During the early weeks of therapy and before any

clinical response could be discerned the intracellular

parameters SOD and LSH showed a definite patternof deterioration, particularly in the GST group, inthose patients who later showed a clinical response.The pattern was most clearly seen when the GSTgroup was subdivided into responders and non-responders. No other clinical or biochemical para-meters showed a clear pattern of early deteriorationin the responder group. This result, if borne out infuture studies, might mean that the likely responseof a patient to a particular therapy could bepredicted early in the course of therapy, even beforethe stage at which clinically assessed improvementbecomes apparent.Sulphydryl group concentrations on both sides of

the cell membrane are higher than CP and SOD(three groups per molecule of CP and 103 groups permolecule of SOD respectively). Consequently, in anassessment of the action of these reactive materialsit seems better to consider the sulphydryl groupsfirst and to view the action of CP and SOD as theyaffect the sulphydryl group-disulphide equilibrium.There has recently been increasing interest in therole of oxygen derived free radicals in inflammatorydisease and tissue destruction, and a large amount ofcircumstantial evidence now exists which implicatesthese species in the pathogenesis of rheumatoidarthritis. One method of assessing the free radicalactivity is to monitor changes occurring in para-meters relatively stable with time involved in thedefence of the cell against free radical attack. Twosuch parameters are the intracellular compoundssuperoxide dismutase and glutathione. Both havebeen separately discussed and postulated as theprime defence mechanism, but there is no consensusas to their relative importance. In vitro there is nodoubt that, molecule for molecule, SOD is moreeffective at removing superoxide ions than arethiols, but we have found that in the relativeconcentrations of SOD and thiol found in either redor white cells it is the thiol concentration which ispredominant. 13 Furthermore, the superoxide ion isshort lived, lasting approximately 10-5 S8 and thuswill react close to the environment in which it isgenerated. This suggests that in many cases themore ubiquitous sulphydryl system will be present atthe reactive site in higher relative concentrationthan that suggested above. In addition, the super-oxide ion is only one of the free radicals believed tobe produced by the phagocytosing cell. For theothers the specificity of SOD for superoxide ionwould suggest that it would be less effective in thesecases, whereas the more general radical scavengingactivity of sulphydryl groups would indicate a rolefor them in these processes also.The method of analysis used for thiol estimation

does not detect haemoglobin thiol and consequently,

copyright. on M


http://ard.bmj.com

/A

nn Rheum


ctober 1986. Dow

nloaded from

http://ard.bmj.com/


glutathione is the largest single component ofLSH. As well as being a key element in the defenceof the cell against attack by radicals in general, it is akey species in protein synthesis, a major source ofnon-protein bound sulphydryl groups within thecell, and, through the glutathione to diglutathioneratio, is used to control key enzyme processes incellular metabolism.23 During inflammation cellulardegradation of erythrocytes in the joint will releaseglutathione into the extracellular environment andhence into the plasma. The major individual compo-nent of the plasma sulphydryl group concentration(PSH) is the sulphydryl group on albumin. Thisgroup is protected from reaction with other proteinsby being situated in a pocket in the proteinstructure. Exchange reactions between other pro-teins and small molecule disulphides are likely,however, and consequently an indirect equilibriumbetween each plasma protein and membrane sul-phydryl groups is to be expected (Fig. 3). In this waysulphydryl reacting compounds, including smallmolecule disulphides (RSSR'), can act to potentiatethe various protein-SH concentrations in plasma.Caeruloplasmin is a specific oxidase for smallmolecule sulphydryl groups (RSH), producing smallmolecule disulphides (RSSR'). It would, therefore,be expected to have a direct effect on this equilib-rium. Thus the effect of additional extracellularglutathione produced by inflammation would be toincrease the concentration of plasma protein presentin the disulphide form and cause a correspondingreduction in PSH.The inorganic biochemical parameters chosen are

thus one measure of the oxidative status across thecell membrane and of the free radical protectionsystem in erythrocytes. Since both erythrocytes andleucocytes are derived from the same stem cells, andsince the mature erythrocyte can synthesise glu-tathione in much the same way as the leucocyte, itmay be that the intracellular parameters-althoughdetermined in this study by erythrocyte

RSH x

A RS=CP /7H 'RSSR HA

PROTEIN CELL MEMBRANE

Fig. 3 Exchange reactions between plasma protein, smallmolecule disulphide, and cell membrane. CP iscaeruloplasmin, an oxidase with many postulatedfunctions,one of which is the oxidation ofthe sulphydryl group insmall molecules.

concentration-also reflect analogous changes inleucocytes, though the exact nature of such changesseems to depend on the disease and cellular type.One practical point arises from this work-

namely, that sulphydryl groups interfere with mostSOD assays in vitro, so that determination of SODin lysates from red or white cells, for which thesulphydryl group concentration is not known, mustbe regarded as suspect. In the present study,however, the lysate sulphydryl group concentrationsare low since the sulphydryl groups have beenallowed to oxidise and consequently, with theexception of one or two of the highest results, arebelow the level at which added glutathione inter-feres with the assay. Thus the correlation betweenSOD activity and lysate thiol concentration appearsto have a biochemical rather than an analyticalmeaning. In all cases at 0 and 24 weeks a significantcorrelation between SOD and LSH at the 0.1%level is obtained (Table 6). When the change inSOD over the 24 week period is plotted against thechange in the sulphydryl group concentration overthe 24 week period, however, a very differentpattern emerges (Fig. 2). The placebo group shows asignificant correlation, but for auranofin this cor-relation is less obvious, and no correlation is evidentfor GST. Thus although the drugs over 24 weeksproduce little change in mean concentration oflysate thiol or SOD, there is evidence that they dopotentiate the free radical defence system in a rathermore subtle way.

In conclusion, in agreement with previousstudies'2 24 25 this work has further confirmed thatplasma thiol levels are an indicator, albeit a ratherslow one, of eventual therapeutic outcome. Theintracellular parameters give an early indication oftherapeutic response analogous to that found forpenicillamine.26 For penicillamine Munthe sug-gested that a rise of about 20% was required if thepatient was to show response, and it is clear fromour choice of responders that a similar rise isobserved here (Fig. la). It is too soon to recommendthis approach for use in clinical assessment but,taken together with other studies on the thiol systemand bearing in mind the confirmatory value of twocorrelated analyses, it would seem to offer potentialin this direction if confirmed in further studies. Inany event the fundamental nature of the process andthe nature of the relation between sulphydryl groupsand superoxide dismutase activity would seem to begood subjects for further investigation.

D L is supported by a grant from Smith, Kline, and French. C J Mis a Medical Research Council assistant. K J R is supported by theInternational Copper Research Association. C N N M is supportedby the Arthritis and Rheumatism Council.

copyright. on M


http://ard.bmj.com

/A

nn Rheum


ctober 1986. Dow

nloaded from

http://ard.bmj.com/


References

1 Finkelstein A E, Walz D T, Batista V, Mizraji M, Roisman F,Misher A. Auranofin: new oral gold compound for treatment ofrheumatoid arthritis. Ann Rheum Dis 1976; 35: 251-7.

2 Berglof F E, Berglof K, Walz D T. Auranofin: an oralchrysotherapeutic agent for treatment of rheumatoid arthritis. JRheumatol 1978; 5: 68-74.

3 Proceedings therapeutic innovation in rheumatoid arthritis,worldwide auranofin symposium. J Rheumatol 1982; suppl. 8,9:1-209.

4 Brown D H, Smith W E. The chemistry of the gold drugs usedin the treatment of rheumatoid arthritis. Chem Soc Rev 1980; 9:217-40.

5 Banford J C, Brown D H, Hazelton R A, McNeil C J, Smith WE, Sturrock R D. Altered thiol status in patients withrheumatoid arthritis. Rheumatol Int 1982; 2: 107-11.

6 Scudder P R, Al-Timini D, McMurray W, White A G, Zoob BC, Dormandy T L. Serum copper and related variables inrheumatoid arthritis. Ann Rheum Dis 1978; 37: 67-70.

7 Brown D H, Buchan W W, El-Ghobary A, Smith W E, TeapeJ. Serum copper and its relationship to clinical symptoms inrheumatoid arthritis. Ann Rheum Dis 1979; 38: 174-6.

8 Johnston R B, Lehmeyer J E. The involvement of oxygenmetabolites from phagocytic cells in bactericidal activity andinflammation. In: Michelson AM, McCord J M, Fridovich I,eds. Superoxide and superoxide dismutases. New York:Academic Press, 1977: 291-305.

9 Rister M, Bauermeister K, Gravert U, Gladtke E. Superoxidedismutase deficiency in rheumatoid arthritis. Lancet 1978; i:1094.

10 Banford J C, Brown D H, Hazelton R A, McNeil C J, SturrockR D, Smith W E. Serum copper and erythrocyte superoxidedismutase in rheumatoid arthritis. Ann Rheum Dis 1982; 41:458-62.

11 Zgirski A F, Chidambaram V, Freiden E. Comparison of thecatalytic activities of mammalian caeruloplasmin. In: Sorenson JR J, ed. Inflammatory diseases and copper. Clifton, NJ:Humana, 1982: 171-81.

12 McNeil C J, Banford J C, Brown D H, Smith W E. A

relationship between thiols and the superoxide ion. FEBS Lett1981; 133: 175-8.

13 Salin M L, McCord J M. Free radicals and inflammation:protection of the phagocytosing leukocytes by superoxidedismutase. J Clin Invest 1975; 59: 1319-23.

14 Fee J A. Is superoxide toxic? In: Bannister W H, Bannister J V,eds. Biological and clinical aspects ofsuperoxide and superoxidedismutase. New York: Elsevier. 1980: 41-8.

15 Goldstein I M, Kaplan H B, Edelson H S, Weissman G.Caeruloplasmin, a scavenger of superoxide anion radicals. JBiol Chem 1979; 254: 4040-5.

16 Youssef AR, Baron DN. Leucocyte superoxide dismutase inrheumatoid arthritis. Ann Rheum Dis 1983; 42: 558-62.

17 Ritchie D M, Boyle J A, McInnes J M, et al. Clinical studieswith an articular index for the assessment of joint tenderness inpatients with rheumatoid arthritis. Q J Med 1968; 37: 393-406.

18 Ellman G L. Tissue sulphydryl groups. Arch Biochem Biophys1959; 82: 70-7.

19 Menden E E, Boiano J M, Murthy L. Petering H G.Modification of a p-phenylenediamine oxidase method topermit ceruloplasmin determinations. Anal Lett 1977; 10:197-204.

20 Misra H P, Fridovich I. Superoxide dismutase, a photochemicalaugmentation assay. Arch Biochem Biophys 1977; 181: 308-12.

21 Mallya R K, Mace B E W. The assessment of disease activity inrheumatoid arthritis using a multivariate analysis. RheumatolRehabil 1981; 20: 14-6.

22 Siegel S. Non-parametric statistics: for the behavioural sciences.Intermational student ed. Tokyo: McGraw-Hill, 1956.

23 Kosower N S, Kosower E M. The glutathione-glutathionedisulphide system. In: Pryor W A. ed. Free radicals in biology.Vol. II. New York: Academic Press, 1976.

24 Pickup M E, Dixon J S, Bird H A. On the effects ofanti-rheumatic drugs on protein sulphydryl reactivity in humanserum. J Pharm Pharmacol 1980; 32: 301-2.

25 Hall N D, Gillan A H. Effects of anti-rheumatic drugs onprotein SH reactivity of human serum. J Pharm Pharmacol1979; 31: 676-80.

26 Munthe E, Kass E, Jellum E. D-Penicillamine induced increasein intracellular GSH correlated to clinical response in RA. JRheumatol [Suppl] 1981; 7: 14-9.

copyright. on M


http://ard.bmj.com

/A

nn Rheum


ctober 1986. Dow

nloaded from

http://ard.bmj.com/

Documents

leucocnytesconfirmatory evidence for an eventual improvement. Key words: rheumatoid arthritis, redox status, superoxide dismutase, caeruloplasmin. The gold compound auranofin (1-thio-,B-D-glucopyranosato)