Comparison of rac- and meso-2,3-DimercaptosuccinicAcids for Chelation of Mercury and Cadmium Using

Chemical Speciation Models

Xiaojun Fang,†,‡ Fengmei Hua,§ and Quintus Fernando*,†

Departments of Chemistry and Pharmaceutical Sciences, University of Arizona,Tucson, Arizona 85721

Received June 7, 1995X

The formation constants of various mercury and cadmium chelates of the stereoisomers of2,3-dimercaptosuccinic acid (DMSA)1 have been determined from potentiometric titrations inthe presence of the competing ligand EDTA. The mercury chelates formed at pH 7.4 are themonomeric HgL of the DMSA diastereoisomers and HHgL2 of rac-DMSA. Mercury is completelycomplexed at pH greater than 3.0 in solutions containing more than 1 equiv of either rac- ormeso-DMSA. At high concentrations (10 µM and above) mercury tends to bind to a greaterextent to rac- than tomeso-DMSA. At pH 7.4, the predominant cadmiummeso-DMSA chelatespecies in solution is CdL, and HCdL is present at a much smaller concentration. With rac-DMSA, however, the predominant cadmium chelate species is HCdL at a low concentration ofthe ligand, and at a high concentration of the ligand the species CdL2 predominates. Cadmiumis completely chelated at pH 7.4 in solutions containing more than 1 equiv of either rac- ormeso-DMSA. At pH around 5.5, which corresponds to the pH of the kidney, however, asignificant amount of free cadmium is present in solutions containing 1 equiv or less of eitherDMSA stereoisomer. From the results of an analysis of speciation models, probable kidneydamage, that may result from free cadmium ion release in the kidney during chelation therapy,is inferred when meso-DMSA is used for mobilizing cadmium. In contrast, the release of freecadmium ion is negligible in the pH range in the kidney when rac-DMSA is used. On thebasis of the speciation models, rac-DMSA is found to be far superior to meso-DMSA in thetreatment of acute cadmium poisoning.

Introduction

Effects of mercury toxicity manifest themselves pri-marily in the central nervous system and in kidneys,where mercury accumulates after exposure. Studies withexperimental animals showed that meso-2,3-dimercap-tosuccinic acid (meso-DMSA)1 (1, 2) and 2,3-dimercapto-propanesufonic acid (DMPS) (1) are very effective in thetreatment of mercury intoxication. The lack of knowl-edge of the chemistry of mercurymeso-DMSA chelation,however, still hampers our understanding of the ef-fectiveness of meso-DMSA in comparison with otherchelating agents in the treatment of acute and chronicmercury poisoning.Cadmium is highly nephrotoxic (3) and has an ex-

tremely long biological half-life (4). An assessment of anumber of chelating agents for their abilities in removingcadmium after injection of 115mCdCl2 revealed thatmeso-DMSA was most effective for the treatment of acutecadmium poisoning (5). Comparison ofmeso-DMSA withother chelating ligands by Jones et al. (6), using computermodeling of chelation equilibria, showed that meso-DMSA is likely to be one of the most useful ligands for

the treatment of acute cadmium intoxication, and thiscorrelates very well with the results of an in vivostudy (5).DMSA exists in two diastereoisomeric forms,meso and

racemo. Unlike its meso isomer, rac-2,3-dimercaptosuc-cinic acid (rac-DMSA), which is an equimolar mixture of(R,R)-2,3-dimercaptosuccinic acid and (S,S)-2,3-dimer-captosuccinic acid, is very soluble in water, in stronglyacidic solutions, and in organic solvents such as ethylacetate and ethyl ether. These striking differences in thehydrophilicities and lipophilicities of the two DMSAdiastereoisomers suggest that rac-DMSA is a more ef-fective antidote than meso-DMSA for the treatment ofheavy metal poisoning. A computer modeling of leadchelation equilibria at physiological pH as well as acalculation of the relative plasma mobilization index(RPMI) of DMSA stereoisomers for lead also indicatedthat rac-DMSA is more effective than itsmeso isomer inremoving in vivo lead (7).Egorova et al. (8) and Okonishnikova (9, 10) reported

in the early 1970’s, in their toxicological studies with theisotopes Hg203 and Cd115, that with identical doses of therac- andmeso-DMSA, the rac-DMSA invariably increasedthe elimination of mercury and cadmium in comparisonwith the meso-DMSA. Despite this important observa-tion that rac-DMSA was superior to meso-DMSA in thetreatment of mercury and acute cadmium poisoning, thechelating properties of rac-DMSA, especially with mer-cury and cadmium, have received little attention. Theformation constants of the mercury and cadmium che-lates of rac-DMSA have not been determined, probablybecause of the difficulties encountered in the synthesis

* To whom correspondence should be addressed.† Department of Chemistry.‡ Present address: Department of Pharmacology and Toxicology,

College of Pharmacy, University of Arizona, Tucson, AZ 85721.§ Department of Pharmaceutical Sciences.X Abstract published in Advance ACS Abstracts, December 15, 1995.1 Abbreviations: FDA, Food and Drug Administration;meso-DMSA,

meso-2,3-dimercaptosuccinic acid; DMPS, 2,3-dimercaptopropane-1-sulfonic acid; EDTA, ethylenediaminetetraacetic acid; rac-DMSA, rac-2,3-dimercaptosuccinic acid; RPMI, relative plasma mobilizing index;NIST, National Institute of Standards and Technology.

284 Chem. Res. Toxicol. 1996, 9, 284-290

0893-228x/96/2709-0284$12.00/0 © 1996 American Chemical Society

+ +

of rac-DMSA and the low solubilities of the mercury andcadmium chelates of rac-DMSA. It is important torationalize these observed differences of rac- and meso-DMSA in mobilizing in vivo mercury and cadmium ions.The objectives of this paper are (i) to postulate thecompositions of the various mercury and cadmium che-lates ofmeso- and rac-DMSA that are formed in aqueoussolution, (ii) to determine their formation constants, (iii)to calculate the distributions of the various mercury andcadmium species as a function of pH, (iv) to assess therelative effectiveness of meso- and rac-DMSA in thechelation of mercury and cadmium at physiological pH,and finally, (v) to provide a rationale for the therapeuticuse of DMSA stereoisomers in the treatment of mercuryand cadmium poisoning.

Experimental Section

Materials. rac-DMSA was synthesized in our laboratory aspreviously described (11);meso-2,3-dimercaptosuccinic acid wasa gift from Johnson & Johnson Baby Products Co. (Skillman,NJ); buffer solutions were purchased from Fisher Scientific (FairLawn, NJ); siliconizing fluid was purchased from PrincetonApplied Research (Princeton, NJ); all other inorganic compoundsused were purchased from Mallinckrodt, Inc. (Paris, KY), andwere of analytical reagent grade. Caution:Gloves are necessaryin the handling of all the above chemicals to avoid direct contactwith skin.Potentiometric Determination of Formation Constants

of the Cadmium and Mercury Chelates of DMSA Stereo-isomers. Standard 0.1 M KOH solutions were prepared from45% (w/w) KOH solution and CO2-free double-deionized H2O.The prepared base solutions were kept under a nitrogenatmosphere to prevent the absorption of CO2. The carbonatecontent, determined as described by Martell et al. (12), wasfound to be 0.5% in the KOH solution; the exact molarity of KOHsolution was determined with potassium acid phthalate withphenolphthalein as indicator. A stock 0.1 M HNO3 solution wasprepared from concentrated HNO3 and CO2-free double-deion-ized H2O and standardized by KOH titration. A standard 0.1M ethylenediaminetetraacetic acid (EDTA) solution was pre-pared from the disodium salt of EDTA and CO2-free double-deionized H2O and standardized by titration with a standardzinc solution prepared from zinc oxide (13). A stock EDTAsolution (0.01 M) was prepared from the standard 0.1 M EDTAsolution. Stock Cd2+ and Hg2+ solutions (0.01 M) were preparedfrom Cd(NO3)2 and HgCl2 and CO2-free double-deionized H2O,and their concentrations were obtained by titration with EDTA.The diastereoisomeric DMSA stock solutions (0.01 M) wereprepared by dissolving about 91.1 mg of the solid DMSA in 20mL of CO2-free double-deionized H2O containing 2 equiv ofKOH, and diluting the solution to 50 mL with CO2-free double-deionized H2O. The stock solution of DMSA was transferredto a small polystyrene bottle, deaerated with argon for 1-2 min,and stored in a freezer. The concentrations of the DMSAsolutions were calculated from the initial weights of the ligandsand then confirmed by KOH titration with methyl red asindicator.The apparatus and procedures used for potentiometric mea-

surement of hydrogen ion concentration have been describedpreviously (11). The titration solutions were prepared accordingto the following procedure for the determination of the formationconstants of the lead-DMSA complexes. Thirty milliliters ofCO2-free deionized water was placed in a titration vessel intowhich 6 mL of the stock EDTA solution was transferred,followed by addition of 6 mL of the stock lead nitrate solutionand 6 mL of the stock DMSA solution. Finally, a desiredamount of KNO3 was weighed and added into the titration vesselto maintain the ionic strength equal to 0.10 ( 0.01 throughoutthe titration. The accuracy of volume transfer in the prepara-tion of the stock solutions and the titration solutions wasensured by the exclusive use of volumetric pipets which were

previously siliconized with siliconizing liquid and calibrated withdouble-deionized water. The titration vessel was also siliconizedbefore use.

Results

Determination of Formation Constants. The po-tentiometric titration data were used to calculate theformation constants of the cadmium and mercury che-lates of EDTA and DMSA stereoisomers with the aid ofthe BEST program (12). The BEST program calculatesthe formation constants of the species which are assumedto be present in the titrated solution, minimizing thediscrepancies between the calculated values of -log[H+]and the experimental values of -log[H+] obtained froma potentiometric titration. Since the presence of thespecies which were included in the calculation by theBEST program cannot be proved by the potentiometrictitration itself, the formation constants were determinedonly on the basis of statistics. A statistically favorablemodel is a set of chemical species assumed to be presentin the titrated solution, which, with the aid of BESTprogram, yields a best fit for the experimental titrationdata.Because of the low solubilities of the mercury and

cadmium chelates of the DMSA stereoisomers, the po-tentiometric titrations of the mercury- and cadmium-DMSA solutions were performed in the presence of acompeting ligand, EDTA. Although the formation con-stants of various mercury and cadmium complexes ofEDTA are available from the database of the NationalInstitute of Standards and Technology (NIST) (14), theywere verified independently in our laboratory to ensurethe accuracy of the data as well as to validate ourpotentiometric method.(A) Formation Constants of Mercury and Cad-

mium Chelates of EDTA. The statistically favorablemodel for the titration of solutions containing equimolaramounts of Hg2+ and H2EDTA2- consisted of HgEDTA,HHgEDTA, H2HgEDTA, H3HgEDTA, and HgEDTAOH.The mean log formation constants calculated from twotitrations are listed in Table 1. The formation constantsof H3HgEDTA and HgEDTAOH were not available forcomparison. The mean log formation constant of HgED-TA and its stepwise log protonation constants along withtheir standard deviations determined from two repetitiveexperiments in our laboratory are 21.3 ( 0.3, 3.80 ( 0.01,and 2.75 ( 0.02, respectively. These values are in goodagreement with those selected from the NIST database(14), 21.5 ( 0.1, 3.2 ( 0.1, and 2.1, respectively. Ourvalue of the log formation constant of HgEDTA alsoagrees very well with that determined by Casas andJones (15), 22.14 ( 0.15. This demonstrates the reli-ability of using a statistically favorable model to calculatethe formation constants of metal chelates by using theBEST program. All five Hg-EDTA complexes were laterincluded in the models for calculations of the forma-tion constants of the mercury complexes of DMSA stereo-isomers.The statistically favorable model for the titration of

solutions containing equimolar amounts of Cd2+ andH2EDTA2- consisted of CdEDTA, HCdEDTA, H2CdEDTA,and CdEDTAOH. The average log formation constantscalculated from two titrations are listed in Table 1. Thedifference in the reported value of the log formationconstant of CdEDTAOH (14) and that determined in ourlaboratory is 1.5, which may be attributed in part to the

Mercury and Cadmium Chelates of meso- and rac-DMSA Chem. Res. Toxicol., Vol. 9, No. 1, 1996 285

+ +

ionic strength of 1.0 that was employed by the previousinvestigators (14). The log formation constant of Cd-EDTA and its stepwise log protonation constants deter-mined in our laboratory are 16.57 ( 0.01, 2.83 ( 0.01,and 1.3 ( 0.6, respectively, which are in good agreementwith those selected from the NIST database (14), 16.5 (0.1, 2.9, and 1.6 ( 0.1, respectively. This again demon-strates the reliability of calculating the formation con-stants of metal chelates on the basis of a statisticallyfavorable model. All four Cd-EDTA complexes werealways included later in the models for calculation of theformation constants of the cadmium complexes of DMSAstereoisomers.The potentiometric titration points of solutions con-

taining equimolar amounts of Hg2+ and H2EDTA2- andequimolar amounts of Cd2+ and H2EDTA2- are plottedin Figure 1, and so are their corresponding simulatedtitration curves by the BEST program (solid lines).(B) Formation Constants of Mercury Chelates of

meso- and rac-DMSA. The formation constants ofmercury chelates of DMSA stereoisomers were calculatedfrom two repetitive titrations of solutions containingequimolar amounts of Hg2+, H2EDTA2-, and H2DMSA2-

stereoisomers. The presence of three mercury complexes(HgL, HHgL, and HgL2) of meso-DMSA was deducedfrom the refinements by the BEST program, and theircorresponding log formation constants, listed in Table 2,are 27.5 ( 0.2, 32.4 ( 0.2, and 34.2 ( 0.2, respectively.The presence of five mercury complexes (HgL, HHgL, H2-HgL, HgL2, and HHgL2) of rac-DMSA was deduced fromthe refinements by the BEST program, and their corre-sponding log formation constants, listed in Table 2, are28.5 ( 0.1, 33.0 ( 0.1, 35.8 ( 0.2, 36.7 ( 0.6, and 46 (1, respectively. The values of the standard deviation forthe formation constants of HgL2 and HHgL2 of rac-DMSAare greater than those for the other chelates because themaximum distributions of the two chelates in the entirecourse of the titration are comparatively small, 15% and7%, respectively, of the total amount of mercury ionspresent in solution. The sample titration curves obtainedwith the solutions containing equimolar amounts ofmercury, EDTA, and the DMSA stereoisomer are plottedin Figure 1, and so are the corresponding simulatedtitration curves by the BEST program (solid lines).(C) Formation Constants of Cadmium Chelates

of meso- and rac-DMSA. The formation constants of

cadmium chelates of EDTA determined in our laboratorywere used in the calculations. The formation constantsof cadmium chelates of DMSA stereoisomers were cal-culated from two repetitive titrations of solutions con-taining equimolar amounts of Cd2+, H2EDTA2-, andH2DMSA2- stereoisomers. The presence of four cadmiumcomplexes (CdL, HCdL, CdL2, and HCdL2) ofmeso-DMSAwas deduced from the refinements by the BEST program,and their corresponding log formation constants, listedin Table 2, are 16.6 ( 0.4, 22.8 ( 0.1, 22.1 ( 0.1, and32.2 ( 0.1, respectively. The presence of three cadmiumcomplexes (HCdL, H2CdL, and CdL2) of rac-DMSA wasdeduced from the refinements by the BEST program, andtheir corresponding log formation constants, listed inTable 2, are 25.1 ( 0.2, 28.1 ( 0.1, and 29.3 ( 0.2,

Table 1. Formation Constants of Mercury and Cadmium EDTA Chelates at µ ) 0.10 and T ) 25.0 °C

specieslog âpqrap q r

Hg:EDTA ) 1:1 1 1 0 21.3 ( 0.3b 21.5 ( 0.1 (14) 22.14 ( 0.14 (15)1 1 1 25.1 ( 0.3b 24.7 ( 0.2c (14)1 1 2 27.8 ( 0.2b 26.8d (14)1 1 3 30.2 ( 0.3b1 1 -1 11.7 ( 0.3b

Cd:EDTA ) 1:1 1 1 0 16.57 ( 0.01b 16.5 ( 0.1 (14)1 1 1 19.40 ( 0.01b 19.4e (14)1 1 2 20.7 ( 0.4b 21.0f (14)1 1 -1 4.81 ( 0.05b 3.3g (14)

a âpqr ) [HgpLqHr(4q-2p-r)-]/[Hg2+]p[L4-]q[H+]r or [CdpLqHr

(4q-2p-r)-]/[Cd2+]p[L4-]q[H+]r; L4- represents completely deprotonated EDTAligand. b The values were determined in our laboratory, and the estimated errors were calculated from the log formation constants obtainedby the BEST program from two sets of titration data with σ(pH)fit of 0.0047 and 0.0092, respectively, for Hg-EDTA system and withσ(pH)fit of 0.0017 and 0.0025, respectively, for Cd-EDTA system (12). c The value was calculated from the reported log protonation constantof HgL of EDTA, 3.2 ( 0.1. d The value was calculated from the reported log protonation constant (2.1) of HHgL of EDTA measured atan ionic strength 1.0. e The value was calculated from the reported log protonation constant (2.9) of CdL of EDTA. f The value was calculatedfrom the reported log protonation constant (1.6) of HCdL of EDTA measured at an ionic strength 1.0. g The value was calculated from thereported log formation constant (13.2) of CdLOH of EDTA, which was defined by the following: log KCdLOH ) log([CdL2-]/[CdLOH3-][H+]),measured at an ionic strength 1.0.

Figure 1. Potentiometric titration curves of the solutionscontaining the following: (9) 1.296 mM Na2H2EDTA and 1.264mM Hg2+ (Hg2+:H2EDTA2- ) 1:1); (b) 1.255 mM Na2H2EDTA,1.247 mM Hg2+, and 1.290 mM Na2H2(rac-DMSA) (Hg2+:H2-EDTA2-:H2(rac-DMSA)2- ) 1:1:1); (2) 1.255 mM Na2H2EDTA,1.247 mM Hg2+, and 1.259 mM Na2H2(meso-DMSA) (Hg2+:H2-EDTA2-:H2(meso-DMSA)2- ) 1:1:1); (0) 1.315 mMNa2H2EDTAand 1.268 mM Cd2+ (Cd2+:H2EDTA2- ) 1:1); (O) 1.255 mMNa2H2EDTA, 1.249 mM Cd2+, and 1.290 mMNa2H2(rac-DMSA)(Cd2+:H2EDTA2-:H2(rac-DMSA)2- ) 1:1:1); (4) 1.255 mMNa2H2-EDTA, 1.249 mM Cd2+, and 1.259 mM Na2H2(meso-DMSA)(Cd2+:H2EDTA2-:H2(meso-DMSA)2- ) 1:1:1). The solid lines arecorresponding titration curves simulated by the BEST program.

286 Chem. Res. Toxicol., Vol. 9, No. 1, 1996 Fang et al.

+ +

respectively. The formation constant of CdL of rac-DMSA is not reported because the complex has a highlog protonation constant (greater than 10) and theobserved maximum distribution of the complex in theentire course of the titration is small (less than 0.2% ofthe total amount of cadmium ions present in solution).The sample titration curves obtained with the solutionscontaining equimolar amounts of cadmium, EDTA, andthe DMSA stereoisomers are plotted in Figure 1, and soare the corresponding simulated titration curves by theBEST program (solid lines).

Discussion

rac-DMSA Is More Effective Thanmeso-DMSA inMobilizing High Concentrations of Mercury atPhysiological pH. rac-DMSA was found to be superiorto meso-DMSA for the treatment of acute mercurypoisoning in laboratory animals (9). The formation ofcomplexes between mercury(II) and rac-DMSA, however,has not been studied, and their formation constants havenot been reported. meso-DMSA has been used for thetreatment of heavy metal poisoning by researchers in theformer Soviet Union (16) since 1958 and in China (17,18) since 1965. In the U.S., meso-DMSA was approvedby the Food and Drug Administration (FDA) in 1991 (19)for the treatment of children with blood lead levelsgreater than 45 µg/dL, and it is used at present for thedetoxification of children with blood lead levels greater

than 25 µg/dL. meso-DMSA is a useful antidote for acuteand chronic inorganic mercury poisoning as well as formethylmercury poisoning (1, 2, 9, 20, 21). There is,however, very little information about the reaction ofmercury compounds with this ligand. The formationconstant of the complex ofmeso-DMSA, HgL, determinedin our laboratory is 1028.5, which is substantially largerthan the value, 1017.3, reported by Okonishnikova et al.(22), but smaller than the value, 1039.4, suggested byJones (23). Mercaptan compounds are far better com-plexing agents than EDTA for the removal of mercury-(II) and organomercury compounds from the mammalianbody (23). It can be inferred, therefore, that the forma-tion constants of Hg-mercaptan compounds are probablylarger than the formation constant of HgEDTA. Theformation constant determined by Okonishnikova et al.,however, is much smaller than the formation constant,1021.5, of HgEDTA (14), and most likely is underesti-mated. The formation constant suggested by Jones (23)was not determined experimentally, but was estimatedon the basis of the formation constants of mercurycomplexes of other mercaptan compounds. Casas andJones pointed out (15) that the formation constants ofthe mercury complexes of several mercaptan compoundsdetermined in their laboratory could be overestimatedbecause of the systematic errors associated with theirexperimental procedures.The relative efficacy of rac- over meso-DMSA in

mobilizing mercury ions was evaluated, with the aid ofthe computer program SPE (12), by calculation of thedistributions of various mercury species as a function ofpH in the presence of both DMSA stereoisomers. It wasassumed in the calculations that only the mercurychelates of rac- and meso-DMSA listed in Table 2 arepresent in aqueous solution. The distributions of mer-cury species were calculated at various total mercuryconcentrations ranging from 0.2 mg/dL (10 µM) to 30 µg/dL (1.5 µM) and at various concentrations of ligand upto 3.3 mM to simulate the conditions of mercury poison-ing and the in vivo treatment with DMSA. The distribu-tion curves, calculated for solutions containing 60 µg/dLmercury, for various mercury species which contributedmore than 0.1% of total mercury in the pH region 3.0-10.0 are plotted in Figure 2. The behavior of rac- andmeso-DMSA in the complexation of mercury is similar,as shown by the titration curves obtained with the twoisomers (filled circles and filled triangles, respectively,in Figure 1). Both stereoisomers are very effective inmobilizing mercury, because mercury is completely com-plexed by the two DMSA stereoisomers in the entire pHregion when the ratio of total ligand to total mercury insolution is equal to or greater than 1, as shown in Figure2. At a molar ratio of total DMSA and mercury insolution equal to 1, meso-DMSA is as effective as rac-DMSA in binding mercury, with 50% of the mercurybeing complexed by meso-DMSA and the other 50% byits racemic isomer (Figure 2(a)). The mercury chelatesformed in this solution can be represented as HgL, whereL is one of the stereoisomers of DMSA. When theconcentration of the stereoisomers of DMSA is elevated(Figure 2(b)-(d)), HHgL2 of rac-DMSA starts to form atpH 7.4 and the distribution pattern at pH 7.4 undergoesa gradual change, with an increase and decrease in theconcentrations of HgL of rac- and meso-DMSA, respec-tively. Despite its formation, the contribution of HHgL2

of rac-DMSA is not significant unless the ligand:mercuryratio exceeds 30.

Table 2. Formation Constants of Mercury and CadmiumChelates of DMSA Stereoisomers at µ ) 0.10 and

T ) 25.0 °C

specieslog âpqrap q r

Hg:EDTA:rac-DMSA )1:1:1b

1 1 0 28.5 ( 0.1

1 1 1 33.0 ( 0.11 1 2 35.8 ( 0.21 2 0 36.7 ( 0.61 2 1 46 ( 1

Hg:EDTA:meso-DMSA )1:1:1c

1 1 0 27.5 ( 0.2 17.3 (22) 39.4 (23)

1 1 1 32.4 ( 0.21 2 0 34.2 ( 0.2

Cd:EDTA:rac-DMSA )1:1:1d

1 1 1 25.1 ( 0.2

1 1 2 28.1 ( 0.11 2 0 29.3 ( 0.2

Cd:EDTA:meso-DMSA )1:1:1e

1 1 0 16.6 ( 0.4 17.11 (6)

1 1 1 22.8 ( 0.1 23.50 (6)1 2 0 22.1 ( 0.11 2 1 32.2 ( 0.1

a âpqr ) [HgpLqHr(4q-2p-r)-]/[Hg2+]p[L4-]q[H+]r; L4- represents

completely deprotonated DMSA. b The values were determined inour laboratory, and the standard deviations were calculated fromthe log formation constants obtained by the BEST program fromtwo sets of titration data with σ(pH)fit of 0.011 and 0.016,respectively (12). c The values were determined in our laboratory,and the estimated errors were calculated from the log formationconstants obtained by the BEST program from two sets of titrationdata with σ(pH)fit of 0.059 and 0.065, respectively (12). d The valueswere determined in our laboratory, and the estimated errors werecalculated from the log formation constants obtained by theBEST program from two sets of titration data with σ(pH)fit of0.035 and 0.0095, respectively (12). e The values were determinedin our laboratory, and the estimated errors were calculated fromthe log formation constants obtained by the BEST program fromtwo sets of titration data with σ(pH)fit of 0.010 and 0.010,respectively (12).

Mercury and Cadmium Chelates of meso- and rac-DMSA Chem. Res. Toxicol., Vol. 9, No. 1, 1996 287

+ +

The concept of metal RPMI was developed by Fang etal. (24) to quantitatively assess the relative efficacy oftwo chelating agents in mobilizing a metal at physiologi-cal pH. For mercury mobilization, the RPMI of rac- andmeso-DMSA is defined as:

The RPMI was calculated for the solutions containingmercury and the stereoisomers of DMSA using theconcentrations of the mercury species obtained with theaid of the SPE program (12). The results are plotted inFigure 3 in which it is apparent that the RPMI is alwaysgreater than 1. This indicates that rac-DMSA is moreeffective than its meso isomer in mobilizing mercury atpH 7.4. The RPMI curves corresponding to solutionscontaining various amounts of mercury are almost su-perimposable at ligand:mercury ratios below 20. Atratios above 20, they deviate from each other. At aconstant ligand:mercury ratio, the higher the concentra-tion of mercury present in solution, the higher the RPMI.On the basis of our complex formation studies with

mercury, we conclude that rac-DMSA is superior tomeso-DMSA in the treatment of mercury poisoning, especiallyin the case of acute poisoning when a high mercury

concentration is found in blood and a high dose of DMSAis administered. This prediction has been verified byearly investigators (9, 10, 12) with the results obtainedfrom animal experiments. In addition, the high hydro-philicity and lipophilicity of rac-DMSA imply that, unlikemeso-DMSA, it will be readily absorbed and transportedto sites where there are high levels of mercury. Webelieve, therefore, that rac-DMSA is a better antidotethanmeso-DMSA for acute mercury poisoning. In casesof chronic mercury poisoning, the mercury concentrationin the body is comparatively low. As shown in Figure 3,the effect of ligand:Hg ratio on RPMI decreases when theconcentration of mercury decreases. Hence, the dose-related enhancement of mercury chelation by rac-DMSAmay not be significant when rac-DMSA is used for thetreatment of chronic mercury poisoning.rac-DMSA Is More Effective Thanmeso-DMSA in

Mobilizing Cadmium at Physiological pH. rac-DMSA was found to be more effective than its mesoisomer in the elimination of cadmium from rats whenthe drug was administered simultaneously with a singlesubcutaneous tracer dose of Cd115Cl2 solution (10). Thestoichiometry of some cadmium complexes of DMSAstereoisomers was studied by Egorova et al. (25); how-ever, the cadmium complexation of rac-DMSA has notbeen quantitatively studied, and the formation constantsof the cadmium complexes of rac-DMSA have not beenreported. Jones et al. (6) compared the relative effective-ness of several chelating agents in removing cadmiumfrom humans using chemical speciation models andshowed that meso-DMSA and DMPS are likely to be themost efficacious ligands. Formation constants of severalmeso-DMSA cadmium chelates have been reported byJones et al. (6) and are the only such constants that areavailable in the literature. The formation constants ofmeso-DMSA cadmium chelates determined in our labora-tory and by Jones et al., listed in Table 2, are in fairagreement in view of the experimental difficulties en-countered in the measurement of the formation constantsof the complexes of heavy metals with chelating agentscontaining mercaptan groups.

Figure 2. Distribution curves of all mercury species presentin solutions containing 3 µM mercury ion and rac- and meso-DMSA at (a) 1.5 µM each, (b) 3 µM each, (c) 30 µM each, and(d) 300 µM each.

RPMI )(total amount of mercury bound in rac-DMSA)pH)7.4

(total amount of mercury bound in meso-DMSA)pH)7.4

Figure 3. Relative plasma mobilizing index (RPMI) of rac- overmeso-DMSA for Hg2+ versus the molar ratio of the total amountof DMSA stereoisomers to the total amount of mercury at totalmercury concentrations of (9) 1.5 µM, (2) 3 µM, and (b) 10 µM;the total concentrations of rac-DMSA are kept equal to the totalconcentrations of meso-DMSA, i.e., half of the total concentra-tions of the DMSA stereoisomers, in the calculations.

288 Chem. Res. Toxicol., Vol. 9, No. 1, 1996 Fang et al.

+ +

The relative efficacy of rac- and meso-DMSA in mobi-lizing cadmium ions was evaluated in the same manneras the relative efficacy of the two DMSA isomers inmobilizing mercury ions. The distributions of cadmiumspecies were calculated at various total cadmium con-centrations ranging from 112 µg/dL (10 µM) to 17 µg/dL(1.5 µM) and at various concentrations of ligand up to3.3 mM to simulate the conditions of cadmium poisoningand the in vivo treatment with DMSA. The distributioncurves, calculated for solutions containing 34 µg/dLcadmium, for various cadmium species which contributedmore than 0.1% of the total cadmium in the pH region3.0-10.0 are plotted in Figure 4. The behavior of rac-and meso-DMSA in complexing with cadmium is drasti-cally different from each other, as can be clearly seenfrom their titration curves (empty circles and emptytriangles, respectively, in Figure 1). Neither rac- normeso-DMSA is as effective in mobilizing cadmium as itis in mobilizing mercury, because unlike in Figure 2, asignificant concentration of free cadmium ions wasobserved at pH below 7, in the solution containing 1:1ligand:cadmium, and at pH below 5, in the solutioncontaining 2:1 ligand:cadmium (Figure 4(a),(b)).In a solution containing equimolar concentrations of

ligand and cadmium, meso-DMSA is as effective as rac-DMSA in mobilizing cadmium at pH 7.4, with 50% of thecadmium being complexed bymeso-DMSA and the other50% by its racemo isomer. The major types of cadmium

chelates formed in this solution with the two DMSAstereoisomers, however, are different, since rac-DMSAforms HCdL andmeso-DMSA forms mainly CdL (Figure4(a)). Formation of the protonated cadmium chelate,HCdL, of rac-DMSA may facilitate the transport ofintracellular cadmium because of the charge reductionin the cadmium chelate. When the concentration of thestereoisomers of DMSA is increased, rac-DMSA exhibitsa higher cadmium mobilizing ability than meso-DMSA.As shown in Figure 4(b)-(d), with an increase in ligandconcentration the contribution of the cadmium complex,CdL, of meso-DMSA at pH 7.4 decreases, while thecontribution of HCdL of rac-DMSA first increases andthen decreases as a consequence of the domination of thecadmium complex, CdL2, of rac-DMSA at a high ligandconcentration.The relative cadmium mobilizing ability of rac- and

meso-DMSA was also assessed quantitatively usingRPMI, which is defined in this case as:

The values of RPMI for cadmium were calculated in thesame manner as the RPMI for mercury, and the resultsare plotted in Figure 5. Although both rac- and meso-DMSA are less effective in mobilizing cadmium than inmobilizing mercury, rac-DMSA alone is much moreeffective than meso-DMSA in its ability to mobilizecadmium. This is demonstrated in Figure 5 by anincrease of ten to several hundredfold in the RPMIobserved for solutions containing a wide range of cad-mium concentrations. Also shown in Figure 5 is thesubstantial increase in the RPMI with an increase incadmium concentration, thereby demonstrating the su-periority of rac-DMSA as an effective antidote for acutecadmium poisoning.Implicit in Figure 4(a),(b) are two important features

which have a significant bearing on the use of rac- andmeso-DMSA in chelation therapy. It is reasonable to

Figure 4. Distribution curves of all cadmium species presentin solutions containing 3 µM cadmium ion and rac- and meso-DMSA at (a) 1.5 µM each, (b) 3 µM each, (c) 30 µM each, and(d) 300 µM each.

Figure 5. Relative plasma mobilizing index (RPMI) of rac- overmeso-DMSA for Cd2+ versus the molar ratio of the total amountof DMSA stereoisomers to the total amount of cadmium at thetotal cadmium concentrations of (9) 1.5 µM, (2) 3 µM, and (b)10 µM; the total concentrations of rac-DMSA are kept equal tothe total concentrations of meso-DMSA, i.e., half of the totalconcentrations of the DMSA stereoisomers, in the calculations.

RPMI )(total amount of cadmium bound in rac-DMSA)pH)7.4

(total amount of cadmium bound in meso-DMSA)pH)7.4

Mercury and Cadmium Chelates of meso- and rac-DMSA Chem. Res. Toxicol., Vol. 9, No. 1, 1996 289

+ +

assume from the experimental evidence that cadmiumis mobilized by meso-DMSA in vivo as CdL and HCdL.In the absence of excess ligand (this may happen shortlyafter a sudden cessation of chelation therapy with highdose administrations), these chelates will reestablishequilibrium in the tissues where they are transported,while maintaining a 1:1 ligand:cadmium ratio. The freecadmium distribution in this situation was calculated andfound to be similar to that shown in Figure 4(a), withabout 13% of the cadmium being dissociated from thetransported chelates as free cadmium ion at pH around5.5, which corresponds to the pH in urine. The freecadmium ion released may accumulate in the kidneyduring urinary excretion and cause additional damageto the kidney. In contrast, cadmium is mobilized by rac-DMSA mainly in the form of CdL2 and the reestablish-ment of equilibrium by the transported chelates isgoverned by the distribution diagram corresponding toa solution in which a 2:1 ligand:cadmium ratio is main-tained, and in which the free cadmium distribution issimilar to that shown in Figure 4(b). In this case, thefree cadmium ion contribution around pH 5.5 is es-sentially zero, and this indicates that the potential kidneydamage that is likely caused by cadmium chelationtherapy can be minimized if rac- instead of meso-DMSAis administered.We conclude that rac-DMSA is superior tomeso-DMSA

for the treatment of cadmium poisoning and far superiorto meso-DMSA for the treatment of acute cadmiumpoisoning, on the assumption that thermodynamic equi-librium is attained in solution at various pH values inthe presence of the chelating agents and the variouscadmium species.

Acknowledgment. This work was supported in partby NIEHS Grant ES 03356.

References(1) Planas-Bohne, F. (1981) The influence of chelating agents on the

distribution and biotransformation of methylmercuric chloride inrats. J. Pharmacol. Exp. Ther. 217, 500-504.

(2) Kostyniak, P. J. (1982) Mobilization and removal of methylmer-cury in the dog during extracorporeal complexing hemodialysiswith 2,3-dimercaptosuccinic acid (DMSA). J. Pharmacol. Exp.Ther. 221, 63-68.

(3) Friberg, L., Elinder, C. G., Kjellstrom, T., and Nordberg, G. F.(1986) Cadmium and Health, CRC Press, Boca Raton, FL.

(4) Lucis,O. J., Lynk, M. E., and Lucis, R. (1969) Turnover of 109Cdin rats. Arch. Environ. Health 18, 307-310.

(5) Eybl, V., Sykora, J., Koutensky, J., Caisova, D., Schwartz, A., andMertl, F. (1984) Interaction of chelating agents with cadmium inmice and rats. Environ. Health. Perspect. 54, 267-73.

(6) Jones, D. C., Smith, G. L., May, P. M., and Williams, D. R. (1984)Assessment of pharmaceutical agents for removing cadmium fromhumans using chemical speciation models. Inorg. Chim. Acta 93,93-100.

(7) Fang, X., and Fernando, Q. (1995) Stereoisomeric selectivity of2,3-dimercaptosuccinic acids in chelation therapy for lead poison-ing. Chem. Res. Toxicol. 8, 525-536.

(8) Egorova, L. G., Okonishnikova, I. E., Nirenburg, V. L., andPostovskii, I. Y. (1972) Complex formation of certain metals withdimercaptosuccinic acid stereoisomer. Khim. Farmat. Zhur. 6,14-16.

(9) Okonishnikova, I. E. (1970) The comparative antidotal potencyof stereoisomeric dimercaptosuccinic acids in mercury poisoning.Gig. Tr. Prof. Zabol. 14, 18-22.

(10) Okonishnikova, I. E. (1971) The effect of steric isomers ofdimercaptosuccinic acid on the elimination of some heavy metalsfrom body. Gig. Tr. Prof. Zabol. 15, 50-52.

(11) Fang, X., and Fernando, Q. (1994) Synthesis, structure, andproperties of rac-2,3-dimercaptosuccinic acid, a potentially use-ful chelating agent for toxic metals. Chem. Res. Toxicol. 7, 148-156.

(12) Martell, A. E., and Motekaitis, R. J. (1988) Determination andUse of Stability Constants, VCH Publishers, New York, NY.

(13) Meites, L., Ed. (1963)Handbook of Analytical Chemistry, Section3, pp 3-95, McGraw-Hill Book Co., Inc., New York, NY.

(14) (14) NIST Critical Stability Constants of Metal Complexes Data-base, Version 1.0, NIST Standards Reference Database 46,Gaithersburg, MD.

(15) Casas, J. S., and Jones, M. M. (1980) Mercury(II) complexes withsulfhydryl containing chelating agents: stability constant incon-sistencies and their resolution. J. Inorg. Nucl. Chem. 42, 99-102.

(16) Kostygov, N. M. (1958) The antidotal action of mercaptosuccinicacid and Unithiol against mercury. Farmakologiia i Toksikologiia21, 64-69.

(17) Ting, K. S., Liang, Y. I., Shi, J., Chen, W., and Gu, T., et al. (1965)Chelate Stability of sodium dimercaptosuccinate on the intoxica-tion from many metals. Chinese Medical Journal 51, 304-307.

(18) Wang, S. C., Ting, K. S., and Wu, C. C. (1965) Chelating therapywith Na-DMS in occupational lead and mercury intoxications.Chinese Medical Journal 84, 437-439.

(19) FDA Medical Bulletin: Succimer Approved for Severe LeadPoisoning, Vol. 21, Part 1, p 5, March, 1991.

(20) Aaseth, J., Jacobsen, D, Andersen, O., and Wickstrom, E. (1995)Treatment of mercury and lead poisoning with dimercaptosuccinicacid and sodium dimercaptopropanesulfonate. Analyst 120, 853-854.

(21) Okonishnikova, I. E., Rosenberg, E. E., and Vorontsova, A. S.(1971) The prophylactic effect of succimer in chronic experimentalpoisoning by metallic mercury vapors. Gig. Tr. Prof. Zabol. 15,28-31.

(22) Okonishnikova, I. E., Egorova, L. G., Nirenberg, V. I., andPostovskii, I. Ya. (1970) Complexing properties of stereoisomersof dimercaptosuccinic acid. Khim.-Farm. Zh. 4, 21-24.

(23) Jones, M. M., and Basinger, M. A. (1980) Chelate therapy for typeb metal ion poisoning. Inorganic Chemistry in Biology andMedicine (Martell, A. E., Ed.) ACS Symposium Series 140, pp335-348, American Chemical Society, Washington, DC.

(24) Fang, X., and Fernando, Q. (1994) A comparative study of meso-and rac-2,3-dimercaptosuccinic acids and their zinc complexes inaqueous solution. Chem. Res. Toxicol. 7, 770-778.

(25) Egorova, L. G., and Nirenburg, V. L. (1980) The study ofcomplexing properties of dimercaptosuccinic acids. IV. Complexingof cadmium ions with dl- and meso-dimercaptosuccinic, dl-dimercaptoglutaric and meso-dimercaptoadipic acids. Zh. Obsch.Khim. 50, 426-430.

TX950096O

290 Chem. Res. Toxicol., Vol. 9, No. 1, 1996 Fang et al.

+ +