Ann Rheum Dis: first published as 10.1136/ard.42.5.566 on 1 October 1983. Downloaded from

Annals ofthe Rheumatic Diseases, 1983, 42, 566-570

Gold levels produced by treatment with auranofin and sodium aurothiomalate

D. LEWIS,' H. A. CAPELL,1 C. J. McNEIL,2 M. S. IQBAL,2 D. H. BROWN, 2 AND W. E. SMITH2 From the 1CentreforRheumatic Diseases, University DepartmentofMedicine, Baird Street, Glasgow G4 OEH, and the 2Department ofPure and Applied Chemistry, University ofStrathclyde, Glasgow GJ IXL

SUMMARY Sixty-three patients with rheumatoid arthritis were randomly divided into 3 groups, and treated with either sodium aurothiomalate (Myocrisin), auranofin, or placebo. Gold levels in whole blood, plasma, and haemolysate were measured serially along with clinical and laboratory para- meters of efficacy. Auranofin produced a higher ratio of haemolysate to plasma gold than Myocrisin, and it appears that the affinity of the red cell for gold is reduced during therapy with auranofin. Gold levels did not correlate with changes in the pain score, erythrocyte sedimentation rate, and C-reactive protein, nor with the development of toxicity. In the Myocrisin group the haemolysate gold level achieved was dependent on the number of cigarettes smoked. In the auranofin group there was no such correlation, but the haemolysate gold level was higher for

smokers than non-smokers. The likely action of gold is discussed. copyright.

Gold compounds have been used in the treatment of The purpose of the present study was to measure rheumatoid arthritis for over 50 years, and theirplace the distribution ofgold in the blood of patients receiv- in clinical practice is firmly established.' 2 However, ing Myocrisin or auranofin, both in order to deter- there remains a need to monitor the use of these mine whether or not parameters such as haemolysate compounds very carefully to establish efficacy and to gold or the ratio of haemolysate to plasma gold are anticipate the onset of any toxic reaction which may correlated with efficacy or toxicity and to improve http://ard.bmj.com/ occur. At present the mechanism of action of these our understanding of the in-vivo chemistry of gold compounds is unknown, and no specific parameter, drugs. such as the plasma gold level, has been found to provide an early indication of either efficacy or Materials and methods toxicity. Gold is usually administered parenterally as Myo- Sixty-three patients with definite or classical to the criteria of the crisin (sodium aurothiomalate), but d new orally rheumatoid arthritis, according on September 26, 2021 by guest. Protected administered and chemically different compound, American Rheumatology Association, were included auranofin (triethylphosphine gold thioglucose tetra- in the study. All had active synovitis unresponsive to acetate or triethylphosphine g6ld-2,3,4,6-tetra- nonsteroidal anti-inflammatory drugs. They had not o-acetyl-l-thio-,-D-glucopyranoside) has become previously been treated with gold in any form and available for testing and controlled clinical trials.3 In had not received penicillamine, levamisole, immuno- animal studies the distribution of gold produced by the suppressive drugs, or corticosteroids in the 3 months 2 compounds is different, with auranofin producing preceding this trial. lower tissue gold levels and a higher erythrocyte gold The patients were randomly allocated to 3 groups. level.3 4 It has also been suggested that the erythro- Twenty-one patients (4 male and 17 female, median cyte gold level is a better indicator of tissue gold levels age 46 years, range 28-69) were given weekly intra- during Myocrisin therapy5 and that the erythro- muscular injections of 50 mg of Myocrisin for 12 cyte gold level is a function of cigarette smoking.6 weeks following an initial test injection of 10 mg. Thereafter the frequency of injection was varied Accepted for publication 13 August 1982. Correspondence to Dr W. E. Smith, Department of Pure and from 1 to 4 weeks according to clinical response. Applied Chemistry, University of Strathclyde, Thomas Graham Twenty-one patients (5 male and 16 female, median Building, Cathedral Street, Glasgow Gl 1XL. age 57 years, range 28-71) received auranofin 3 mg 566 Ann Rheum Dis: first published as 10.1136/ard.42.5.566 on 1 October 1983. Downloaded from

Gold levels produced by treatment with auranofin and sodium aurothiomalate 567 b.d. throughout the study. Twenty-one patients (5 Table 2 Gold levels ofpatients withdrawing from the trial male and 16 female, median age 55 years, range 31-72) received placebo tablets identical with Treatment Week Reason for Plasma Au Haemolysate auranofin twice daily. withdrawal (,Ag/ml) Au (pgl/ml) Blood was taken by venesection prior to the next Myocrisin 1 Nitritoid gold or placebo dose at 0, 3, 6, 12, and 24 weeks. 5 ml reaction - - of heparinised blood was centrifuged at 3000 g 16 Rash 1-9 0-2 the 16 Rash 2-0 0-2 for 10 minutes, and the plasma was removed by suc- 20 Rash 2-7 1-0 tion. The cells were washed twice with isotonic saline 20 Proteinuria 5-2 0-1 solution and lysed with an equal volume of distilled Auranofin 5 Diarrhoea 0-2 0-5 water (2 h at 4°C). The sample was centrifuged again 6 Leucopenia 0-1 (3000 g for 10 min) and a sample of haemolysate 16 Haematuria 1-1 1-0 removed by suction. Whole blood, plasma, and blood Placebo 3 Diarrhoea - - levels of haemolysate gold were determined by dilut- 8 Haematuria - 12 7 patients- ing the samples 15 times with distilled water and lackof effect - analysing for gold by atomic absorption spectrometry using carbon furnace atomisation. The methodology SI conversion: ,g/mi = mg/l. has been published previously. Care should be taken to ensure that the background corrector is working satisfactorily and that all sources of smoke and saline interference have been successfully eliminated.7 Clinical and other biochemical assessments of dis- 4. ease activity were performed weekly or monthly. Pain score, visual analogy scale, erythrocyte I I sedimentation rate (ESR), and C-reactive protein Myocr i sin copyright. (CRP) values were selected as the most useful 3. measurements to compare with gold levels. The statistical comparison was by Wilcoxon's I matched-pairs signed-rank test or linear regression as 2. appropriate.

Results 1 http://ard.bmj.com/ Both Myocrisin and auranofin produced a clinical improvement as measured by falls in ESR and CRP a,E ., .- - U- _., _-nL between zero and 24 weeks. However, Myocrisin 3 6 1 2 24 week produced an earlier response which was significant at 12 weeks and produced an improvement in pain 1-2. score which appears to be sustained through the -i 24-week period (Table 1). Of the 63 patients 48 completed the 6-month study. The largest drop-out on September 26, 2021 by guest. Protected occurred in the placebo group, due to the lack of *8. Auranofin effect. Only a few patients discontinued active I Table 1 A statistical analysis ofthe change in pain score, ESR, and CRP measurements between week 0 and weeks 12 *4~ and24 ofthe trial. Wilcoxon's matched-pairs signed-rank test / was used Compound Week Pain score ESR CRP

. c) _ iL = .L- Auranofin 12 p<0-05 NS NS 3 6 12 A24 week-.- 24 NS p<0-05 p<0-01 Myocrisin 12 p<0-05 p<001 p<0-01 Fig. 1 Plasma and lysate gold levels following 24 p<0-01 p<0-01 p<0-01 administration ofMyocrisin or auranofin. Lysate values are 12 NS NS Placebo 24 p<0-05NS NS NS shaded. The heights ofthe bars are medians and the error bars refer to interquartile ranges. Ann Rheum Dis: first published as 10.1136/ard.42.5.566 on 1 October 1983. Downloaded from

568 Lewis, Capell, McNeil, Iqbal, Brown, Smith therapy. There was no obvious correlation between was a different affinity for the plasma gold in its plasma or lysate gold levels and drop-out due to rash auranofin-administered form than for Myocrisin, and or other toxic reaction with either Myocrisin or untreated cells appreared to be able to absorb gold auranofin. However, the number of drop-outs was not taken up by the original auranofin-treated cells. rather small (Table 2). No appreciable amount of gold was transferred from Plasma gold levels increased with time in both either auranofin- or Myocrisin-treated cells to Myocrisin and auranofin groups (Fig. 1). The Myo- placebo plasma. crisin group reached a plateau value more quickly at Although there was a continual increase in gold about 6 weeks, whereas with auranofin a plateau level on both treatments, there was no correlation value was reached after about 13 weeks. The between clinical improvement as measured by ESR 24-week result reflects maintenance therapy with or CRP and the gold level for whole blood, plasma, Myocrisin and continuing therapy with auranofin. haemolysate, or haemolysate-to-plasma ratio, sug- The median gold concentration between 6 and 12 gesting that none of these parameters may be used weeks with Myocrisin was 3.6 ug/ml and with per se as an indicator of clinical or toxicological auranofin was 09 ,ug/ml. The interquartile range is response. larger with auranofin. (SI conversion: ug/ml = mg/l.) Cigarette smoking affected the haemolysate-to- Haemolysate gold levels were detectable during plasma ratio with both auranofin and Myocrisin (Fig. both Myocrisin and auranofin treatment. Haemoly- 2), but the ratio for non-smoking patients treated sate gold concentrations were lower with Myocrisin with auranofin was sufficiently high to suggest that than auranofin (median concentrations after 12 there was a non-smoking-dependent mechanism of weeks of therapy were 066 ,g/ml and 095 ug/ml, absorption in that case. A plot of the number of respectively). Since plasma gold levels were much cigarettes against haemolysate gold for Myocrisin- higher with Myocrisin, the haemolysate-to-plasma treated patients who smoke indicated that there was a ratio was much lower. After 3 weeks of auranofin statistically significant (r = 0-86, p<0Q001) dose- therapy a very high haemolysate-to-plasma ratio response relationship (Fig. 3). There was no similar was recorded, but with that exception this ratio correlation with auranofin. copyright. remained approximately constant with time for each drug. To examine the reactivity of the gold administered orally, separated plasma samples from an auranofin- treated patient and from a Myocrisin-treated patient 2. after 12 weeks' therapy were incubated for 12 hours The cells were with separate aliquots of placebo cells. http://ard.bmj.com/ separated and a haemolysate prepared as before. About 10 to 20% of the Myocrisin gold was trans- ferred to the placebo cells and most of the auranofin plasma gold was transferred (Table 3). Thus, there

0 Table 3 Gold incubation study. Erythrocyte and plasma

gold levels measured on 3 individual patients after 12 weeks on September 26, 2021 by guest. Protected oftreatment. Plasma and cells were then exchanged, ; 1 21X incubated for 12 hours, reseparated, and gold levels CL. ++ 0 remeasured. (a) Gold levels in each component at the start of experiment; (b) gold-containing cells and placebo plasma; (c) placebo cells and gold-containing plasma 8 0~~~~~~~ * *0 Cells Plasma Gold levels (in pglml) CeUl Plasma * (a) Auranofin Auranofin 0-99 0-79 Myocrisin Myocrisin 0-3 3-36 Placebo Placebo 0-0 0.0 " AF AF M M (b) Auranofin Placebo 0-72 0 0 non - non- Myocrisin Placebo 0-28 0.0 smoker smoker smoker smoker (c) Placebo Auranofin 0-68 0-05 Placebo Myocrisin 0-05 3 0 Fig. 2 A comparison ofthe ratios ofhaemolysate-to-plasma goldfor auranofin smokers and non-smokers with those for SI conversion: ,sg/ml = mg/l. Myocrisin smokers and non-smokers. Ann Rheum Dis: first published as 10.1136/ard.42.5.566 on 1 October 1983. Downloaded from

Gold levels produced by treatment with auranofin and sodium aurothiomalate 569

0) 2- 0 0 0 0 0 Fig. 3 A relationship between the w number ofcigarettes smoked and 0 CI) a gold levelforpatientsafter 12 weeks >-J ofMyocrisin therapy. 0

z.4 0 5 10 15 NQ OF CIGARETTES per DAY

Discussion with auranofin, either because the membrane changes or because some saturation process occurs. Modern gold therapy uses a gold(i)-thiol compound With auranofin there appears to be a natural such as Myocrisin, Solganal (aurothioglucose), or mechanism of cellular uptake which is missing with auranofin. In each of these compounds the gold is Myocrisin. This could be due, for example, to extra copyright. tightly bound to the sulphur, and in the case of negative charges on Myocrisin metabolites, to their auranofin it is also bound to a phosphorus group.4 In- polymeric nature, or to their affinity for aqueous vivo it seems likely that there will be an exchange rather than lipid media. In smokers either the cell between the sulphur and phosphorus ligands and membrane is disturbed or, more likely, the extra naturally occurring thiols as thiocyanate or cyanide ligands added to the plasma react with the gold to form a compound that is readily -* AuSR + R' SH AuSR' + RSH transported across the membrane. Thiocyanate is and it is believed that the in-vitro forms of gold con- present in the larger amounts in the blood of smok- http://ard.bmj.com/ sist of a series of such compounds, with little or no ers9 but is the least likely of the 2 ligands to be active, free gold. It has been suggested that the thiol or since gold thiocyanate is not very stable and is unlikely sulphydryl group ligand is the active species,8 but the to exist in a thiol matrix such as blood and tissue. It quantities used in gold therapy are much smaller than seems more likely that cyanide will form a stable those used with active ligands containing thiols (for complex which can be transported across the mem- example, penicillamine), so that it seems more likely brane. The cyanide may then recycle and pick up con- more so as an trans- that it is the gold moiety which is the active gold and could act effective on September 26, 2021 by guest. Protected stituent. porting agent in the quantities available in-vivo. Auranofin is a rather different compound from Thus 2 quite different compounds appear to form Myocrisin in that it is more lipid and less water- different metabolites and give different gold distribu- soluble, and it is monomeric whereas Myocrisin is tions in blood. Nevertheless, both produce a thera- polymeric. One objective of the present study was to peutic response which is likely to be due to the gold try to discover whether a common form of gold was moiety. Plasma and lysate gold levels do not correlate produced in serum. The experiment in which placebo with therapeutic response and there seems no cells were incubated with the plasma confirmed what evidence that smoking enhances the action of Myocri- is clear from the distribution, that the forms of gold in sin, so that the measurement of intracellular or extra- the plasma after 12 weeks are chemically distinct. cellular gold levels does not seem to be useful in A higher proportion of the auranofin gold is in the monitoring gold therapy per se. The conventional erythrocytes 3 weeks after the first administration of explanation for the action of gold, namely, that it gold than at longer time intervals. Since virtually all stabilises lysosomal membranes, is based on the plasma gold after 12 weeks of therapy can still be detection of electron-dense deposits"0 round the removed by placebo erythrocytes, it would seem that membrane and on the discovery of aurosomes in the nature of the erythrocyte alters during therapy white cells." This form of gold accounts for very little Ann Rheum Dis: first published as 10.1136/ard.42.5.566 on 1 October 1983. Downloaded from

570 Lewis, Capell, McNeil, Iqbal, Brown, Smith of the total body gold, and the techniques used are 2 Sigler J W, Bluhm G B, Duncan H, Sharp J T, Ensign D C, insufficiently sensitive to pick up the reactive molecu- McCrum W R. A double-blind study of the effects of gold salts in the treatment of rheumatoid arthritis. Arthritis Rheum 1972; 15: lar gold in the system. In view of the apparent effect 125-6. of auranofin on cell membranes, and since intra- 3 Walz D T, DiMartino M J, Sutton B M. In: Scherrer R A, cellular parameters are not more, and are possibly Whitehouse M W, eds. Antiinflammatory agents-chemistry and less, significant than extracellular ones, it seems poss- pharmacology. New York: Academic Press, 1974: 1: 217-44. 4 Brown D H, Smith W E. The chemistry of the gold drugs used in ible that the true role of gold is to modify some the treatment of rheumatoid arthritis. Chem Soc Rev 1980; 9: sulphydryl-disulphide interchange processes by 217-40. reacting with them, possibly at the cell membrane. 5 Moller Pedersen S, Moller Graabaek P. Gold in erythrocytes, Thus a more germane approach to the problem of whole blood and plasma during long-term chrysotherapy. Ann Rheum Dis 1980; 39: 576-9. predicting efficacy and toxicity may be to investigate 6 Graham G G, Champion G D, Haavisto T M, McNaught P J. the effect of plasma gold on the cellular function of Gold binding to red blood cells. Ann Rheum Dis 1981; 40: leucocytes. In any event auranofin seems to be effec- 210-1. tive, albeit at a slower rate, at gold concentrations in 7 Kamel H, Brown D H, Ottaway J M, Smith W E. Determination of gold in blood fractions by atomic absorption spectrometry blood which are below those used in Myocrisin using carbon rod and carbon furnace atomisation. Analyst 1976; therapy, and it may be possible to maintain patients 101: 790-7. on this compound at a full dose for a much longer 8 Jeilum E. Gold and thiol compounds in rheumatoid arthritis. period than with Myocrisin. Scand J Rheumatol 1979; suppl 28: 28-36. 9 Pettigrew A R, Logan R W, Willocks J. Smoking in preg- D.L. acknowledge support from Smith Kline and French and nancy-effects on birth weight and on cyanide and thiocyanate C.J.McN. a Medical Research Council research assistantship. levels in mother and baby. BrJ Obstet Gynaecol 1977; 84: 31-4. 10 Norton W L, Ziff M. Electron microscopic observations on the rheumatoid synovial membrane. Arthritis Rheum 1966; 9: References 589-610. 1 Empire Rheumatism Council. Gold therapy in rheumatoid arth- 11 Ghadially F N. The aurosome. J Rheumatol 1979; 6 (suppl 5): ritis. Ann Rheum Dis 1961; 20: 315-34. 45-50. copyright. http://ard.bmj.com/ on September 26, 2021 by guest. Protected