Chem. Res. Toxicol. 1995,8, 525-536
525
Stereoisomeric Selectivity of 2,3-Dimercaptosuccinic Acids in Chelation Therapy for Lead Poisoning Xiaojun Fang and Quintus Fernando* Department of Chemistry, University of Arizona, Tucson, Arizona 85721 Received December 26, 1 9 9 P The formation constants of lead chelates of the stereoisomers of 2,3-dimercaptosuccinic acid (DMSA) were determined from potentiometric titrations in the presence of the competing ligand, EDTA. The lead chelates formed at pH 7.4 with the stereoisomers of DMSA are the monomeric complexes PbL and HPbL. Formation of PbL and HPbL at pH 7.4 is independent of total concentrations of lead and DMSA present, and so is the concentration ratio of PbL:HPbL. Lead is completely chelated at pH 7.4 when the total concentration of ligand is equal to or greater t h a n the total concentration of lead present. Lead tends to bind to a greater extent with ruct h a n with meso-DMSA, and the relative extent increases with an increase in the concentration ratio of ligand to lead and finally reaches a constant value of 45. The binding sites in the chelates, PbL, of the stereoisomers of DMSA are the two thiolate groups and one carboxylate group. ruc-DMSA also forms a dimeric complex PbzLz in which both carboxylate groups of the ligands participate in binding with lead ions. The formation constants of the lead chelates of rac-DMSA were invariably found to be larger t h a n those of the corresponding of meso-DMSA chelates, because in all the lead chelates of the stereoisomers of DMSA formed in solution, ruc-DMSA existed in staggered anti conformations, whereas meso-DMSA preferred a staggered gauche conformation with respect to carboxylate groups in the ligands. The potential of using ZnL2 of ruc-DMSA as a therapeutical lead chelator was assessed by considering its leadmobilizing ability and its ability to deplete endogenous zinc; on this basis it is predicted that ZnL2 of rac-DMSA is a better chelator t h a n meso-DMSA for the treatment of lead poisoning.
Introduction The deleterious effect of low-level lead on the growth of children was reported by Schwartz et al. in the analysis of data from the second National Health and Nutrition Examination Survey (NHNES).’ In 2695 children, 6 months through 7 years of age, an inverse correlation of blood level with height, weight, and chest circumference was unequivocally observed (11. It has been shown in most (2-51, but not all (6)cross sectional studies of the correlation of low-level blood lead and the cognitive performance of preschool and school-age children that a significant decrease in the cognitive performance and behavior scores of children is associated with an elevation of average blood lead in the range 10-30 pg/dL. Topics concerning childhood lead poisoning have been reviewed by Davis and Svendsgaard (7) and Angle (8). The Clean Air Scientific Advisory Committee to the U.S. Environmental Protection Agency considers a blood lead level of 10-15 pgldL and below as a nontoxic level (9). The current understanding (10)is that a childhood blood lead level of 10 pgldL is not only the threshold for inhibition of erythrocyte aminolevulinic acid dehydrase (ALA-D, porphobilinogen synthetase) but also appears to be the threshold for association with noticeable effects on fetal maturation, mental development, and hearing in preschool children and on behavior and cognitive performance of school-age children. * To whom correspondence should be addressed.
Abstract published in Advance ACS Abstracts, April 15, 1995. Abbreviations: NHNES, National Health and Nutrition Examination Survey; ALA-D, aminolevulinic acid dehydrase; FDA, Food and Drug Administration; meso-DMSA,meso-2,3-dimercaptosuccinic acid; rac-DMSA, rac-2,3-dimercaptosuccinic acid; RPMI, relative plasma mobilizing index; IR, infrared; (R,R)-DMSA, (R,R)-2,3-dimercaptosuccinic acid; (S,S)-DMSA, (S,S)-2,3-dimercaptosuccinicacid; NIST, National Institute of Standards and Technology. @
Chelation therapy for heavy metal intoxication has been practiced in various forms for approximately 50 years, because chelation can increase the urinary excretion of metals. Despite the controversies involved in defining the efficacy of a chelator (111, the extent of urinary excretion of ingested metals is still frequently considered to be a crucial index of the efficacy of a chelator. There has been an increasing use of oral meso2,3-dimercaptosuccinic acid (meso-DMSA) for the pharmacological treatment of children with blood Pb 225 pgl dL, since its approval by the U.S. Food and Drug Administration (FDA) in 1991 (12) for the treatment of children with blood Pb 245 pgldL. Clinical studies of meso-DMSA (13-18) show meso-DMSA to be specific for lead excretion and orally effective without clinically important increases in the depletion of calcium, zinc, and iron. Despite the increasing use of meso-DMSA in the U.S. a s a chelating agent for lead poisoning, its disadvantage in human use is that it is virtually insoluble in acidic solution and very slightly soluble in water. From a pharmacokinetic perspective, this reduces the gastrointestinal absorption of the drug (only 20.6% of the administered meso-DMSA is found in the urine as total meso-DMSA (19,20) 14 h after the drug is given orally) required to and results in an increase in the time (Lax) reach a maximum concentration (Cmax)of drug in the blood and a decrease in the value of C,,,. From a clinical perspective, prescription of a comparatively high dose of meso-DMSA becomes necessary to maintain a n effective blood meso-DMSA level for chelation therapy, and doserelated side effects (21)have been observed in a human patient. “A careful monitoring for potential side effects” was recommended by Grandjean et al. (21) for meso-
0893-228x/95/2708-0525$09.00/00 1995 American Chemical Society
526 Chem. Res. Toxicol., Vol. 8, No. 4, 1995
DMSA doses greater than or equal to 10 mgkg, which is a normal therapeutical dose administered to leadpoisoned patients. On the other hand, the rac-2,3dimercaptosuccinic acid (rac-DMSA) is very soluble in strongly acidic solution, in water, and in ethyl ether (22). Consequently, rac-DMSA is highly hydrophilic and lipophilic, and it can be readily absorbed and transported through the gastrointestinal membrane and, therefore, can be a more effective chelating agent than meso-DMSA for lead poisoning. rac-DMSA has not received as much attention as mesoDMSA because it is difficult to synthesize and is unavailable from commercial sources. rac-DMSA has never been studied in either animals or human subjects for its efficacy in the treatment of lead poisoning. There is experimental evidence, however, to demonstrate that racDMSA is superior to meso-DMSA in the treatment of mercury (23)and acute cadmium (24)poisoning. It was also reported (25, 26) that 9 9 m Tchelates ~ of rac-DMSA did not deposit 99mTcin the bone but the corresponding 99mTcchelates of meso-DMSA did. The 99mTc-rac-DMSA was also cleared from blood more rapidly than 9 9 m T ~ meso-DMSA in 1-3 h after injection (26). The different in vivo behavior of the 99mTc chelates of DMSA diastereoisomers has been attributed to the difference in the steric arrangement of carboxylate groups of the isomeric ligands in the complexes (26). The experimental results obtained with the 9 9 m Tchelates ~ of DMSA diastereoisomers demonstrate that the in vivo behavior of metal chelates of DMSA diastereoisomers is closely related to the structures of the chelates, probably by altering the solubilities of the chelates or by interacting with endogenous species. A knowledge of the structures of the lead chelates formed in aqueous solution is, therefore, crucial for understanding their transport through membranes, redistribution among organs, and excretion mechanism from the body. In addition, a knowledge of the structures of the lead chelates is also important for designing more effective chelating agents. Rivera et al. (27) studied the binding sites of lead chelates of DMSA diastereoisomers in the solid state by comparing their infrared (IR) spectra obtained in KBr pellets. The binding sites of the lead chelates of mesoand ruc-DMSA formed in aqueous solution have not been directly studied to date because of their low solubilities. Recently, Fang and Fernando (28) reported the binding sites and the stoichiometry of the zinc chelates of mesoand rac-DMSA formed in aqueous solution by studying the changes in the asymmetric and symmetric stretching bands of carboxylate groups of DMSA ligands as a function of zincligand ratio. The changes in the asymmetric and symmetric stretching bands of the carboxylate groups of DMSA ligands were mainly attributed to the formation of coordinate bonds between the zinc ion and the oxygen atoms of the carboxylate groups in DMSA ligands (28). An infrared study of the binding sites of DMSA stereoisomers in their lead chelates which have low solubilities in aqueous solution is difficult, because the local environment of DMSA ions (hydration and ion pair formation with counterions) a t low concentrations can be significantly altered by a change in the concentration of the ligand alone. In this case, the change in the asymmetric and symmetric stretching bands of the carboxylate groups in DMSA ligands with a change in the 1ead:ligand ratio may not be solely attributed to the binding of the metal ions to the carboxylate groups of DMSA ligands; a prior knowledge of the behavior of
Fang and Fernando DMSA ligands in the presence of counterions and water molecules is necessary. The formation constants of the lead chelates of mesoDMSA were determined by several investigators (2931). The formation constants of lead chelates of racDMSA, however, have not been studied to date. The efficacy of a chelating agent is primarily determined by thermodynamics; from this viewpoint, an effective lead antidote must have a high mobilizing ability for lead ions. The relative lead-mobilizing ability of rac- and mesoDMSA a t physiological pH can be assessed by using the relative plasma mobilization index (RPMI) (32) if prior knowledge of the formation of the lead chelates of the stereoisomers of DMSA a t physiological pH and their formation constants are available. The objectives of this paper are to postulate the binding sites of the lead chelates of meso- and rac-DMSA in aqueous solution, to determine the formation constants of various lead chelates of meso- and rac-DMSA, to calculate the distributions of the various lead species as a function of pH, and to assess the relative mobilizing ability of rac- and meso-DMSA for lead ions a t physiological pH.
Experimental Section Materials. Deuterium oxide (99.9% deuterium) was purchased from Cambridge Isotope Laboratories (Woburn, MA); lead perchlorate hydrate (98%) was purchased from Aldrich Chemical Co. Inc. (Milwaukee, WI); rac-DMSA was synthesized in our laboratory as previously described (22);meso-2,3-dimercaptosuccinic acid was a gift from Johnson & Johnson Baby Products Co. (Skillman, NJ); buffer solutions were purchased from Fisher Scientific (Fair Lawn, NJ); siliconizing fluid was purchased from Princeton Applied Research (Princeton, NJ); all other inorganic compounds used were purchased from Mallinckrodt, Inc. (Paris, KY), and were of analytical reagent grade. Caution: Gloves are necessary in the handling of all the above chemicals to avoid direct contact with skin. IR Spectroscopy in Aqueous Solution. IR spectra of aqueous solutions were measured with a NICOLET Fourier transform IR spectrometer Model 510P at ambient temperature. A calcium fluoride cell, which has a transmission range from 66 666 t o 1110 cm-I, equipped with a thin Teflon spacer was employed, and the thickness of the cell, which was experimentally determined from the interference fringes (33),was 31 pm. All experimental IR spectra were recorded after 64 scans against purge nitrogen background, and the resolution of the IR measurement was 2 cm-I for all spectra. The processed IR spectra, plotted only in the frequency region between 1660 and 1300 cm-I to show the asymmetric and symmetric stretching bands of the carboxylate groups in the DMSA isomers, were obtained after subtraction of the spectrum of DzO from the experimental spectra followed by a partial subtraction of the spectrum of a concentrated NaCl solution, if the IR sample solution contained additional NaC1, and a partial subtraction of the spectrum of 1%HzO in DzO. The subtraction of H2O-inD20 spectrum, which displayed a strong band at 3418 cm-1 and a weak band a t 1462 cm-’ which is located between the asymmetric and symmetric stretching bands of the carboxylate groups, was performed until the strong band a t 3418 cm-l in the resulting spectrum was no longer evident. IR Spectra of Na4L of meso- and ruc-DMSA. Stock solutions of 0.10 M tetrasodium DMSA of the two DMSA isomers were prepared by dissolving 0.40 mmol of the appropriate DMSA in 4.0 mL of DzO containing 0.40 M NaOD. A stock solution of 4.8 M NaCl was prepared by dissolving a weighed amount of NaCl crystals in DzO. Two hundred microliters of the solutions, prepared by mixing 0.1 M stock DMSA solution, 4.8 M stock NaCl solution, and D20 in a varying ratios with the aid of micropipet, was used for measuring IR spectra.
Chem. Res. Toxicol., Vol. 8, No. 4, 1995 627
Lead Chelates of meso- and rac-DMSA IR Titration of N a L of m s o - and rac-DMSA with Pb(ClO&. A stock solution of 0.1 M lead perchlorate was prepared by dissolving 0.11 mmol of lead perchlorate i n 1.1mL of D2O; its concentration was confirmed by EDTA titration to be 0.102 M. Working solutions of the DMSA stereoisomers (0.02 M) were prepared by dilution of their 0.1 M stock solutions. Working solutions (0.02 M) of the stereoisomeric DMSA ligands containing a 1:l ratio of Pb2+:L4-were prepared by mixing 240 pL of 0.1 M lead perchlorate stock solution with 240 pL of 0.1 M of the appropriate DMSA stock solution i n 720 p L of DzO. The resulting 1:l Pb:ruc-DMSA is a clear solution, but the 1:l Pb:meso-DMSA is an orange colloid. Two hundred microliters of IR solution at a desired 1ead:ligand ratio was made by (1) mixing the 0.02 M DMSA working solution with the 0.02 M lead: ligand 1:1 working solution in different ratios, or (2) adding a calculated volume of DzO in a polyethylene microcentrifuge tube, followed by addition of 40 p L of appropriate 0.1 M DMSA stock solution, a calculated volume of 4.8 M NaCl solution, if necessary, and a calculated volume of 0.1 M lead perchlorate. All of the 0.02 M lead-ruc-DMSA IR solutions were clear and stable if the Pb:ligand ratio did not exceed 1.0. If the Pb:ligand ratio exceeded 1.0, a yellowish precipitate formed. All IR “solutions” of 0.02 M meso-DMSA at varying Pb:ligand ratios were virtually colloids which were unstable and darkened within 30 min.
Potentiometric Determination of Formation Constants of the Lead Chelates of DMSA Stereoisomers. Standard KOH solutions of 0.1 M were prepared from 45% (w/w) KOH solution and COz-free double-deionized HzO. The prepared base solutions were kept under a nitrogen atmosphere to prevent the absorption of COz. The carbonate content, determined as described by Martell e t al. (341, was found to be 0.5% in the KOH solution; the exact molarity of KOH solution was determined with potassium acid phthalate with phenolphthalein a s indicator. A stock HN03 solution of 0.1 M was prepared from concentrated HN03 and CO2-free double-deionized H2O and standardized by KOH titration. The standard EDTA solution of 0.1 M was prepared from disodium EDTA and CO2-free double-deionized H2O and standardized by titration of standard zinc solution prepared from zinc oxide (35). A stock EDTA solution (0.01 M) was prepared from 0.1 M standard EDTA solution. A stock Pb2+ solution (0.01 M) was prepared from Pb(N03)2 and COz-free double-deionized HzO and standardized by EDTA titration. The diastereoisomeric DMSA stock solutions of 0.01 M were prepared by weighing about 91.1 mg of the compound i n a 50 mL volumetric flask followed by addition of 20 mL of CO2-free double-deionized HzO and 2 equiv of KOH, dilution to 50 mL with COz-free double-deionized H20, transferring to a small plastic bottle, deaerating with argon for 1-2 min, and finally keeping in a freezer until use. The concentrations of the DMSA solutions were calculated from the initial weights of the ligands and then confirmed by KOH titration with methyl red a s indicator. The apparatus and procedures used for potentiometric measurement of hydrogen ion concentration have been described previously (22). The titration solutions were prepared according to the following procedures: 30 mL of COz-free deionized water was placed i n a titration vessel into which 6 mL of the stock EDTA solution was transferred, followed by addition of 6 mL of the stock lead nitrate solution and 6 mL of the stock DMSA solution when lead-DMSA complexes were titrated. Finally, a desired amount of KNO3 was weighed and added into the titration vessel to maintain the ionic strength equal to 0.10 0.01 throughout the titration. The accuracy of volume transfer i n the preparation of the stock solutions and the titration solutions was ensured by the exclusive use of volumetric pipets which were previously siliconized with siliconizing liquid and calibrated with double-deionized water.
+
Results Dependence of IR Spectra of DMSA Stereoisomers on Their Concentrations and the Concentration of Sodium Ion. The IR spectra of rac- and meso-
DMSA tetrasodium salts in aqueous solutions a t concentrations ranging from 0.1 to 0.01 M are shown in Figure 1, and the IR spectra of 0.02 M rac- and mesoDMSA ligands, L4-, a t sodium ion concentrations ranging from 0.08 to 3.9 M are shown in Figure 2. Although the rac-DMSA is composed of equal amounts of two enantiomers ((RJ3)-2,3-dimercaptosuccinic acid ((R,R)-DMSA)and (S~)-2,3-dimercaptosuccinic acid ((S,S)DMSA)), the IR stretching frequencies of the carboxylate groups, however, should not be affected by the chiral centers in the ligand. The observed absorption bands in the spectra of tetrasodium rac-DMSA are unsymmetrical and are composed of peaks a t 1578 and 1553 cm-’ under the asymmetric stretching bands and peaks a t 1394 and 1380 cm-l shown as small shoulders under the symmetric stretching bands. The ratio of the intensity of the 1553 to the 1578 cm-l peaks increases with the concentration of the rac-DMSA, and so does the ratio of the intensity of the 1394 to 1380 cm-l peaks (Figure 1A). The intensity change of the latter, however, is much smaller, and this may be explained by interaction of overtones of low frequency vibration in the rac-DMSA with the symmetric stretching band. The exact positions of the four peaks also vary slightly with the change in the ratio of the intensities because of the significant overlapping of the peaks. A similar change was observed for 0.02 M racDMSA solutions with addition of NaCl (Figure 2A). This indicates that the rac-DMSA ligands form comparatively stable ion-paired complexes with sodium ions in the solutions. The peak at 1553 cm-’ under the asymmetric stretching band can be reasonably assigned to the sodium ion-paired rac-DMSA ligand, and the peak a t 1394 cm-’ is mainly due to the sodium ion-paired ligand a s well. As shown in Figure lB, the IR spectra of tetrasodium meso-DMSA remain almost unchanged a t concentrations below 0.02 M, with the asymmetric and symmetric stretching bands of its carboxylate groups a t 1553 and 1383 cm-l, respectively. When the concentration increases from 0.02 to 0.05 M, both asymmetric and symmetric stretching bands become narrowed and shiRed to 1545 and 1385 cm-l, respectively. The narrowing and shifting of peaks may not be attributed to the formation of sodium ion-paired complexes because addition of NaCl to 0.02 M meso-DMSA tetrasodium solutions does not induce any noticeable changes in the spectra (Figure 2B). The origin of the sudden frequency shifts as a result of dilution is still not clear and may be related to the change of solvation of the carboxylate groups. meso-DMSA ligands are mainly present in the non-ion-paired form. IR Spectra of ruc-DMSA-Pb Solutions. The correlation between the asymmetric stretching frequency (v,) and symmetric stretching frequency (vs)of a carboxyl group and their separation (Av)with the nature of the coordination of a carboxylate group was summarized by Deacon and Phillips (36). Unidentate coordination of a carboxylate group is generally associated with Av values 2200 cm-l; however, bidentate coordination of a carboxylate group does not have a significant effect on Av values compared to that of a n uncoordinated carboxylate group. The Av and intensity ratio of asymmetric and symmetric stretching bands of the carboxylate groups in the stereoisomers of DMSA were used in the interpretation of the structures of the zinc chelates of DMSA by Fang and Fernando (28). Since rac-DMSA ligands form ion-paired complexes with sodium ions and the IR spectrum of the ligand alone in a n aqueous solution changes with its concentration, a
528 Chem. Res. Toxicol., Vol. 8, No. 4, 1995
Fang and Fernando 1553
139s
1545
\
1385
A
-
l
r
l
~
1600
"
1
1
(
1
1
1
1500
1
,
,
,
,
1
1400
WAVENUMBER (em-') Figure 1. IR spectra of tetrasodium DMSA solutions at different concentrations. (A) rac-DMSA; (B) meso-DMSA.
high concentration of sodium ions (2.9 M) was maintained in the solutions containing less than 0.5 equiv of lead to minimize the effect of the concentration of the ligand, that was not complexed with Pb2-, on the observed IR spectra. The IR spectra of 0.02 M tetrasodium rac-DMSA solutions a t 1ead:ligand ratios ranging from 0 to 1.0 are plotted in Figure 3. At a ligand concentration of 0.02 M and in the presence of 2.9 M sodium ions, the separation between the asymmetric and the symmetric bands of the carboxylate groups in the rac-DMSA alone was 159 cm-l and their intensity ratio is 1.02 (Figure 31, which agrees with the values, 158 cm-l and 1.08, reported by Fang and Fernando (28)for the 0.1 M tetrasodium rac-DMSA solution. When the Pb:L ratio increased from 0 to 0.5, both bands gradually became narrowed and symmetrical, which indicated that all the ligands were consumed to form PbL2. The intensity ratio remained almost constant as the Pb:L ratio increases to 0.5, but the asymmetric and the symmetric stretching bands shifted to higher and lower frequencies, respectively, resulting in a band separation of 186 cm-'. The net increase in the band separation was 28 cm-l, which implied that one of the two carboxylate groups of the rac-DMSA in the PbLz complex was unidentately bound to a lead ion. After more than 0.5 equiv of lead was added, a new complex (PbL or Pb2L2) started to form in the solution. At a Pb:L ratio between 0.5 and 0.7, the IR spectra of the solutions remained almost unchanged, indicating that the newly formed PbL complex still made use of one of
its two carboxylate groups to coordinate with the lead ion. When the Pb:L ratio became greater than 0.7, the symmetric stretching band started to broaden and split into two bands as a result of the formation and growth of a new peak at 1367 cm-l. This new peak was attributed to the formation of the dimeric complex PbzLz of rac-DMSA. Both carboxylate groups of the ruc-DMSA ligand in Pb2Lz complex were involved in binding with lead ions, because the separation between the asymmetric stretching and the new peak was 203 cm-I and their intensity ratio was 2.09. IR Spectra of meso-DMSA-Pb "Solutions". The IR spectra of 0.02 M tetrasodium meso-DMSA solutions a t 1ead:ligand ratios ranging from 0 to 1.0 are plotted in Figure 4. No extra sodium ions were added to the IR solutions because their IR spectra were not affected by the ligand concentrations below 0.02 M. At a ligand concentration of 0.02 M, the separation between the asymmetric and symmetric bands of the carboxylate groups in meso-DMSA alone was 169 cm-l and their intensity ratio is 0.96 (Figure 4). After 0.5 equiv of lead was added, the band separation increased to 187 cm-l and the intensity ratio of the asymmetric to symmetric stretching bands increased to 1.27. This indicated that the carboxylate groups of meso-DMSA ligands were involved in the formation of the complex PbLz. It is shown in Figure 1B that dilution of tetrasodium mesoDMSA solution from 0.1 to 0.02 M increases the band separation from 159 to 169 cm-l. Therefore, formation
Chem. Res. Toxicol., Vol. 8, No. 4, 1995 529
Lead Chelates of meso- and rac-DMSA 1577
1
1
1
,
1600
1
1
1
1
)
l
l
l
1500
l
,
l
l
,
,
1400
WAVENUMBER (cui') Figure 2. IR spectra of 0.02 M N a D M S A solutions i n the presence of sodium ion a t different concentrations. (A) ruc-DMSA; (B) meso-DMSA.
of PbL2 in a hypothetical 0.1 M ligand solution will increase the band separation from 159 to 187 cm-l, assuming that the effect of dilution on a lead-bound carboxylate group is negligible in the IR spectra; hence, a net increase of 28 cm-' will result. This value matches the value observed for the formation of PbLz of rac-DMSA and denotes that, like in PbLz of rac-DMSA, one of the two carboxylate groups of meso-DMSA in the PbLz complex is unidentately bound to a lead ion. After 1 equiv of lead ion was added, the band separation further increased to 199 cm-' and the intensity ratio increased to 1.65. This may be a n indication that both carboxylate groups of meso-DMSA ligand are involved in the binding with a lead ion when PbL or PbzLz is formed. This interpretation, however, should be accepted with some caution because the samples used for the IR measurements were more colloidal than the samples at a Pb:L ratio less than 0.5, and the recorded IR spectra are severely complicated by the presence of fine particles of the lead complexes. Determination of Formation Constants. The potentiometric titration data were used to calculate the formation constants of the lead complexes of EDTA and DMSA stereoisomers with the aid of the BEST program
(34). The BEST program calculated the formation con-
stants of the species which were assumed to be present in the titration solution, minimizing the discrepancies between the calculated values of -log [H+l and the experimental values of -log [H+l obtained from a potentiometric titration. Since the presence of the species which were included in the calculation by the BEST program cannot be proved by the potentiometric titration itself, the formation constants were determined only on the basis of statistics. It was noticed that several combinations of species appeared to fit the experimental titration data without a significant difference in the minimum value of dpHkt. Selection of the statistically most favorable model, which yields a smallest value of dpH)fit(341, as the true model to calculate the formation constants of the species could lead to chemically irrational conclusions. It is necessary, therefore, to report the manner in which the speciation model is generated when the reported formation constants are calculated on the basis of statistics. Because of the low solubilities of the lead chelates of DMSA stereoisomers, the potentiometric titrations of the lead-DMSA solutions were performed in the presence of a competing ligand, EDTA. In order to determine the
530 Chem. Res. Toxicol., Vol. 8, No. 4,1995
"
'
1
Fang and Fernando
"
,
1600
1500
"
,
-1
1
1
-,-
,
7 - 1
-,
1400
WAVENUMBER (cm-') Figure 3. IR spectra of 0.02 M Nar(ruc-DMSA) solutions in the presence of Pb2+ a t 1ead:ligand ratios ranging from 0 to 1.
(
I
I
I
,
1600
I
/
J
l
(
1
0
1500
1
,
(
,
1
1
1400
WAVENUMBER ( c m 3
Figure 4. IR spectra of 0.02 M Na&neso-DMSA) "solutions" in the presence of Pb2+ at 1ead:ligand ratios ranging from 0 to 1.
formation constants of the lead chelates of DMSA stereoisomers, in this case, the formation constants of the lead complexes of EDTA have to be determined. Although the formation constants of various lead complexes of EDTA are available from the NIST database (371, they were verified independently in our laboratory to ensure the
accuracy of the data as well as to validate our potentiometric method. (A) Formation Constants of Lead Chelates of EDTA. The five lead-EDTA complexes (PbEDTA, HPbEDTA, H2PbEDTA, H3PbEDTA, and PbEDTAOH) were initially assumed to be present in the titration
Chem. Res. Toxicol., Vol. 8, No. 4, 1995 531
Lead Chelates of meso- and rac-DMSA Table 1. Formation Constants of Lead EDTA Chelates at p = 0.10 and T = 25.0 "C species
IO Pb:EDTA = 1:l
a
1 1 1
1 1 1
0 1 -1
18.18f O.Olb 18.1 =k 0.1(37) 20.83f 0.01' 20.9 f 0.2c(37) 6.63=k 0.14'
Ppqr= [PbpLqH,'4~-*p-r)-]/[Pb2+lp[L4-]q[H+]r; L4- represents
completely deprotonated EDTA ligand. The values were determined in our laboratory, and the estimated errors were calculated from the formation constants obtained by the BEST program from two individual sets of titration data with u(pH)fit of 0.0014 and 0.0036,respectively (34). The value was calculated from the reported protonation constant of the PbL complex of EDTA, 2.8 f
0.1.
?
8
&
8
?
6
4
solutions containing equimolar amounts of Pb2+ and H2EDTA2- and therefore were included in the calculations as a n initial model to refine their formation constants. The protonation constants of EDTA (log K values) used in the refinement were chosen from the NIST database (37). The lead species found to be less than 1%of the total lead concentration in the entire course of the titration were removed from the model, and the remaining complexes were used to recalculate the formation constants. The procedures were iterated until no further complexes were excluded from the model, and the final resulting model was called the statistically favorable model. The formation constants of the complexes included in this model were simultaneously calculated by the BEST program. The statistically favorable model for the titration of solutions containing equimolar amounts of Pb2+ and H2EDTA2- consisted of PbEDTA, HPbEDTA, and PbEDTAOH. The average log formation constants calculated from two titrations are listed in Table 1. The formation constant of PbEDTAOH was not available\for comparison. The reported log formation constant of QbEDTA and its log protonation constant are 18.1f 0.1 aqd 2.8 f 0.1, respectively (37),which are in very good agrdpment with those determined in our laboratory, 18.18f 0.pl and 2.65 f 0.02, respectively. This validates our pot4ntiometric method for the determination of the formation constants of metal chelates. PbEDTA, HPbEDTA, and PbEDTAOH were always included later in the models for c$culations of the formation constants of the lead complexep of DMSA stereoisomers. The potentiometric titration oints of a solution containing equimolar amounts of Pb2- and H2EDTA2- are plotted in Figure 5, and the simulated titration curve by the BEST program is also shown in Figure 5 as a solid line. (B) Formation Constants of Lead Chelates of rucDMSA. At first, six species (PbL, HPbL, HzPbL, PbLz, HPbL2, and HZPbL2) of rac-DMSA were selected in a general speciation model for a solution containing equimolar amounts of Pb2', H2EDTA2-, and H2L2- of racDMSA, assuming that there were no dimeric complexes and hydroxide complexes formed during the titration. It is reasonable to exclude hydroxide species from our models because all titrations were stopped when all the protons initially associated with EDTA and rac-DMSA were completely consumed by the strong base solution. A procedure similar to that used in obtaining the statistically favorable model for Pb-EDTA titrations was employed to generate a statistically favorable model for the competitive potentiometric titration data. The protonation constants of rac-DMSA determined previously (32) and the formation constants of lead chelates of EDTA
F
0
2
1
5
3
mmol OH7mmol H2EDT% Figure 5. Potentiometric titration curves of the solutions containing the following: (0) 1.288mM NazHzEDTA and 1.268 mM Pb2+ (Pb2+:H2EDTA2- = 1:l);(A) 1.255 mM NaZHzEDTA, 1.240 mM Pb2+, and 1.290 mM NazHz(ruc-DMSA) (Pbzf: = 1:l:lk(VI 0.6276mM NaoHoEDTA. HIEDTA~-:H~(~UC-DMSA)~0.6202 mM -Pb2+, and 1.241 mM NanHz(ruc-DMSij (Pb2+i H2EDTA2-:H2(rac-DMSA)2- = 1:1:2);( 0 ) 1.338 mM NazH2EDTA, 1.323 mM Pb2+, and 1.333 mM NazHz(meso-DMSA) (Pb2+:H2EDTA2-:H2(meso-DMSA)2= 1:1:l);(open square with a cross inside) 0.6689mM NazHzEDTA, 0.6616mM Pb2+, and 1.333 mM NazHz(meso-DMSA) (Pb2+:HzEDTA2-:Hz(mesoDMSA12- = 1:1:2).
Table 2. Formation Constants of Lead Chelates of DMSA Stereoisomers at p = 0.10 and T = 25.0 "C species
1 1 1 1 2 1 Pb:EDTAruc-DMSA = 1 1:1:2c Pb:EDTAmeso-DMSA = 1 1 1:l:ld 1 Pb:EDTAmeso-DMSA = 1 1:1:2e
Pb:EDTA:ruc-DMSA = 1:l:lb
1 1 1 2 2 1 2 1 1 1 2
0 1 2 0 0 2 0 0 1 2 1
20.3 26.2 28.8 26.2 42.5 28.8 26.0 18.2 24.7 27.5 32.3
17.46(29) 17.4(31) 24.82(30)
a PPqr = [Pb,LqH,'~-2P-r)-l/[Pbz+~[L4-]~[H+lr; L4- represents (completely deprotonated DMSA. dpH)fit = 0.008(34). dpH)fit = 0.009 (34). = 0.02 (34). a(pH)fit = 0.009 (34).e~(pH)fit
determined in our laboratory were used in the calculations. The presence of four lead complexes (PbL, HPbL, HzPbL, and PbL2) of rac-DMSA was deduced from the refinements by the BEST program, and their corresponding log formation constants, listed in Table 2, are 20.3, 26.2, 28.8, and 26.2, respectively. Since it was demonstrated from the IR titration results obtained with racDMSA a t 0.02 M that the dimeric complex Pb2Lz coexisted with monomeric PbL, it is reasonable to believe that some dimeric complexes might be present in the titration solution a t a millimolar level. Therefore, two dimeric complexes (Pb2L2 and HPbZL2) were included in the above speciation model a t this point, and the revised model was refined again by the BEST program. The resulting new model contained Pb2L2, with a log formation constant of 42.5, but not HPb2Lz. The validity of the formation constants determined from potentiometric titrations can
532 Chem. Res. Toxicol., Vol. 8, No. 4, 1995
be confirmed, but not proved, by titrating several solutions containing the metal ions and the ligands at varying ratios. If a self-consistent set of complex formation constants is obtained for the complexes that are present in common in the statistically favorable models that correspond to the different solutions, the determined formation constants are generally valid. Therefore, titration of another solution containing Pb2+and H2L2- of ruc-DMSA a t a Pb:L ratio of 1:2 was performed, and the titration data were used to refine a statistically favorable model for this solution. The starting model contained six monomeric species (PbL, HPbL, HZPbL, PbL2, HPbL2, and HzPbLZ) and two dimeric complexes (Pb2L2 and HPbzL2) of rac-DMSA, the resulting statistically favorable model consisted of only two lead chelates, HzPbL and PbL2. The log formation constants of these two chelates are 28.8 and 25.9, respectively, which are in fair agreement with the respective values, 28.8 and 26.2, determined from the titration of the solution containing equimolar amounts of Pb2+,EDTA, and rac-DMSA. The potentiometric titration points of the solutions containing lead, EDTA, and rac-DMSA a t Pb:EDTADMSA ratios 1:l:l and 1:1:2 are plotted in Figure 5, and the simulated titration curves by the BEST program are shown as solid lines. (C) Formation Constants of Lead Chelates of meso-DMSA. The formation constants of various lead chelates of meso-DMSA are determined in the same manner as those of ruc-DMSA. The eight lead species (PbL, HPbL, H2PbL, PbL2, HPbL2, H2PbL2, Pb2L2, and HPb2L2)of ruc-DMSA were included in the initial model. The protonation constants of meso-DMSA used in the calculations were determined previously (32). Three complexes, PbL, HPbL, and HZPbL, were contained in the statistically favorable model obtained from the titration data of a solution containing equimolar amounts of Pb2’, H2EDTA2-,and H2L2- of meso-DMSA of approximately 1.3 mM. Two complexes, HPbL and HPbL2, were contained in the statistically favorable model obtained from the titration data of a 1:1:2 solution of Pb:EDTA meso-DMSA of approximately 1.3 mM in meso-DMSA. The formation constants of the above complexes determined by the BEST program are listed in Table 2. The complex HPbL was found in both models. The log formation constant of HPbL determined in both models is in fair agreement, and its average value of the log of the formation constant is 24.80 f 0.14, which is in good agreement with the value, 24.82 f 0.06, reported by Willes et al. (30)despite the variation in temperature and ionic strength employed in their potentiometric competition titration experiments. The temperature and ionic strength employed in Willes’ study were 37 “C and 0.15. The potentiometric titration data points of the solutions containing lead, EDTA, and meso-DMSA a t Pb:EDTA: DMSA ratios 1 : l : l and 1:1:2 are plotted in Figure 5, and the simulated titration curves by the BEST program are shown as solid lines.
Discussion ruc-and meso-DMSALigands Behave Differently in Aqueous Solution. ruc-DMSA ligands form ionpaired complexes with sodium ions because addition of sodium chloride to fully deprotonated ruc-DMSA solutions results in a shift of 25 cm-l of the asymmetric stretching of the carboxylate groups to a lower frequency. The 25 cm-l red shift could be rationalized by the effect
Fang and Fernando
(IV)
\L‘
&;.
HS
/C -0
0 (V)
Figure 6. Proposed conformations of ruc-DMSA ligand a t various protonation stages.
of electrostatic interaction between the sodium ions and the carboxylate groups involved. The carboxylate groups of the ruc-DMSA ligands in both ion-paired and “free” states in a solution are symmetrical with the two oxygen atoms of any individual carboxylate group being equivalent with respect to their local electron environments. The electrostatic attraction between positively-charged sodium ions and “free” carboxylate groups simply induces a withdrawal of electrons from the carboxylate groups and results in a reduction of the strength of the stretching vibration of the carboxylate groups, which accounts for the red shift of the asymmetric stretching frequency in the IR spectra of the ruc-DMSA ligands in solutions. The conformations of ruc- and meso-DMSA a t different pD were studied previously (22, 321, but the formation of ion-paired complexes by ruc-DMSA ligands with sodium ions was not considered. If formation of ionpaired complexes is considered, the conformations of rucDMSA ligands a t various protonation stages in aqueous solution can be revised as discussed below. The proposed formula for the sodium ion-paired complex of ruc-DMSA ligand is N a P . As shown in Figure 6 (I), this complex can adopt a conformation in which each sodium ion is virtually linked to three oxygen atoms of the two carboxylate groups of the ruc-DMSA ligand. The proposed conformation of the ion-paired complex is similar to that of ruc-DMSA in the solid state and energetically favored over the other possible conformations. In this conformation the electrostatic attraction between sodium ions and carboxylate groups is enhanced by maximizing the coordination number of the sodium ion in the complex. Monoprotonated ruc-DMSA exists as HNaL2- and adopts a conformation (Figure 6 (IT))in which the electrostatic repulsion between the two thiolate groups is minimized because they have a staggered anti-relation. Dideuterated ruc-DMSA was found to form two intramolecular
Chem. Res. Toxicol., Vol. 8, No.4, 1995 533
Lead Chelates of meso- and rac-DMSA six-membered rings via two deuterium bonds between the deuterated thiol groups and the oxygen atoms of the two carboxylate groups (22). Therefore, diprotmated racDMSA is believed to exist a s HzL2- and assumes the same conformation (Figure 6 (111))as that proposed before on the basis of lH NMR titration (22). In contrast to rac-DMSA ligand, as shown in Figure 2B,both asymmetric and symmetric stretching bands of tetrasodium meso-DMSA in solution are not affected by the addition of sodium ions, which indicates that the ionpaired complex of meso-DMSA with sodium ions may be less stable than the complex of rac-DMSA, and therefore it is not formed in the solution under our experimental conditions. If a similar type of NazL2- complex was formed with meso-DMSA, both thiolate groups and both carboxylate groups of the ligand would be expected to located in the same hemisphere, and this reduces the stability of the complex due to the electrostatic repulsion between the charged groups in the ligand. ruc-DMSAIs More Effective Than Meso-DMSAin Mobilizing Lead at Physiological pH. Although a tremendous increase in the use of meso-DMSA has been reported in the US.for the treatment of lead poisoning since its approval by the FDA (22)in 1991,there is still a lack of information about the reaction of lead with the ligand. There are only three publications which report the formation constants of several lead-meso-DMSA complexes because the low solubility of the lead complexes has precluded their study. Willes et al. (30) determined the formation constants of HPbL and PbLOH of meso-DMSA by performing a potentiometric competition titration. The lead species generated from Willes’ titration data are not exactly the same as those found in our statistically favorable models. This difference can be attributed to the manner in which the potentiometric titrations were performed. Instead of titrating a mixture of lead, EDTA, and meso-DMSA with a strong base solution, Willes et al. titrated a mixture of lead and EDTA with a n alkaline meso-DMSA solution (30). The formation constants of the lead species HPbL determined in both studies are in good agreement. Egorova (29) calculated the formation constants of PbzL and PbL on the basis of the pH dependence of the W spectra of the ligand solutions a t M. More recently, Harris et al. (32) also determined the formation constant of PbL spectrophotometrically by using a competitive reaction between meso-DMSA and N-(2-hydroxyethyl)ethylenediamine-N,N’,”-triacetic acid (HEDTA) for lead ions to prevent the precipitation of the lead-meso-DMSA complex. The log formation constants of PbL, and 1017.46, determined by Harris et al. and Egorova et al., respectively, are smaller than the value, 1018,2,determined in our laboratory. This difference is attributed to the inclusion of PbLz of meso-DMSA in our calculations. In both of the previous spectrophotometric studies, the presence and contribution of the PbLz complexes in the mass balance equations, which were developed to calculate the formation constants of PbL, were ignored. The assumption by Harris et al. for the absence of the PbLz of meso-DMSA in the presence of excess of the ligand was based on the observation that, in a titration of lead ion with the ligand, the absorbance of the solution did not increase significantly after 1 equiv of the ligand was added. There was no evidence, however, to show that the electronic spectrum of PbLz of meso-DMSA was markedly different from that of PbL at the wavelength at which the measurements were performed. In contrast,
HPblrac-DMSAl
Pb(rac4MSAl
20 n
Pblmeso-DMSAl
Pblmeso-DMSA
3 ~~
4
6
5 ~
7
8
9
10
~
-log [H’] Figure 7. Distribution curves of the lead chelates in solutions containing 2.2 pM lead and ruc-and meso-DMSA a t (a) 1.1pM each, (b) 2.2 pM each, (c) 25 ,uM each, and (d) 250 pM each. The lead species plotted are only those which contribute to the total amount of lead by more than 0.001% in the vicinity of pH 7.4.
in our titration of a solution of 1:1:2 Pb:EDTAmesoDMSA, 14% of the lead was found to form HPbLz with meso-DMSA as the ligand. There is no information about the formation constants of the lead chelates of rac-DMSA in the literature. Therefore, no literature comparison can be made with the formation constants of the lead chelates of ruc-DMSA determined in our laboratory. The efficacy of rac- over meso-DMSA in mobilizing lead ions was evaluated by calculation of lead distribution among various chelates as a function of pH in the presence of both DMSA stereoisomers with the aid of the computer program SPE (34). It was assumed in the calculations that only the lead chelates of rac- and meso-DMSA listed in Table 2 are present in aqueous solution. The distributions of lead species were calculated a t various total lead concentrations ranging from 45.5 pgIdL (2.2pM) t o 8.3pgldL (0.4 pM)and a t various concentrations of ligand up to 1.0mM to simulate the conditions of lead poisoning and the in vivo treatment with DMSA. The distribution curves, calculated for solutions containing 45.5 pgldL lead, of various lead species which contributed more than 0.1% of total lead in the pH region 4.5-9.0 are plotted in Figure 7. It is shown in Figure 7a that lead is completely complexed by the ligands, with 50% of the lead being complexed by rac-DMSA and the other 50% by its meso isomer, a t pH 7.4when 1 equiv of DMSA stereoisomers is present. The lead chelates formed in this solution are
534 Chem. Res. Toxicol., Vol. 8, No. 4, 1995
Fang and Fernando
50 i
30 .
S
I
%
)1; 10
1
l,B
0 0
0.5
1 1.5 2 2.5 log(mmo1 DMSNmmol lead)
4 3
Figure 8. Relative plasma mobilizing index (RPMI) of ruc- over meso-DMSA for Pb2+ versus the log of molar ratio of the total amount of DMSA stereoisomers to the total amount of lead at the total lead concentrations of ( 0 )2.2 pM, (0) 1.2 pM, and (a) 0.4 ,uM; the total concentrations of ruc-DMSA are kept equal t o the total concentrations of meso-DMSA, Le., half of the total concentrations of the DMSA stereoisomers, in the calculations.
PbL and HPbL of the stereoisomers of DMSA; the contribution of HPbL complexes is relatively small. When the concentration of the stereoisomers of DMSA is elevated, no new lead chelates are formed a t pH 7.4, but the distribution pattern undergoes a gradual change, shown in Figure 7b-d, with a n increase and decrease in the concentrations of PbL of rac- and meso-DMSA, respectively. The relative lead-mobilizing ability of rue- and mesoDMSA can be quantitatively analyzed using the RPMI, which is defined as:
RPMI = the total amount of lead bound in rac-DMSA the total amount of lead bound i n meso-DMSA The values of RPMI a t pH 7.4 were calculated for the solutions containing lead and the stereoisomers of DMSA using the concentrations of the lead chelates obtained with the aid of the SPE program (34). The results are plotted in Figure 8. The curves, calculated a t the varying total lead concentrations, of RPMI vs log of the ratio of the total concentrations of the stereoisomers of DMSA to lead are superimposed on each other, indicating that the value of RPMI is determined only by the ratio of L:Pb rather than the total concentration of the ligands or lead in solution. Since the RPMI value increases rapidly a t the ratio below 10 and reaches a plateau a t about 45, it can be deduced from the RPMI curves that 40-45 times more lead will be complexed with rac- than with mesoDMSA under exactly the same conditions if the total molar concentration of the ligand surpasses the lead concentration, 2.2 pM, by a factor of 10 or more in solution. The corresponding level of DMSA, 22 pM, can be safely achieved in the blood in the treatment of lead poisoning. On the basis of complex formation, it is clear that rac-DMSA alone is superior to meso-DMSA in mobilizing lead ions if thermodynamic equilibria are attained. There is another interesting feature in Figure 7a which is worthwhile discussing. It is reasonable to assume from
the experimental evidence provided that lead is mobilized by DMSA in vivo as PbL and HPbL. When these lead chelates are transported in the absence of excess ligand (this may happen shortly after a sudden cessation of chelation therapy with high dose administrations), they will reestablish equilibria with the surroundings; this situation can be reasonably described by Figure 7a. Figure 7a shows that a t pH between 5 and 6, which corresponds to the pH range in urine, a few percent of the complexed lead will be released into solution as lead ion, which may accumulate in the kidney during its urinary excretion and cause severe damage to the kidney. This demonstrates a mechanism that can account for the adverse effects of chelation therapy caused by redistribution of toxic metals by chelating agents among various tissues and organs in which pH values are very different. This observation also suggests that the use of a n appropriate chelating agent under inappropriate solution conditions may result in adverse effects. The dissociation of the lead chelates in the kidney can be prevented when DMSA is used therapeutically, if high dose treatments are followed by small doses of DMSA for a short period to ensure that the in vivo concentration of the lead chelate has been lowered to a negligible level in blood before the termination of drug administration. The results of the analysis of lead chelate distributions are summarized as below: (1)lead is completely chelated a t pH 7.4 when the total concentration of DMSA stereoisomers is equal to or greater than the total concentration of lead present; (2) both DMSA stereoisomers form monomeric lead chelates, PbL and HPbL, a t pH 7.4, with the majority of lead being bound as PbL; (3) formation and the concentration ratio of PbL and HPbL of DMSA stereoisomers a t pH 7.4 are independent of the total concentrations of lead and the ligand present in solution; (4) lead tends to bind to a greater extent with rac- than with meso-DMSA, and the ratio of the amounts of the lead bound with rac- and meso-DMSA increases with a n increase in the ratio of the total concentrations of the ligand and lead present in the solution, and finally reaches a maximum of 45. ZnL:! of ruc-DMSA Is Potentially a Better Chelator for Therapeutical Use for Lead Detoxification. Increased urinary excretion approximately by a factor of 2, of zinc by rac-DMSA, was reported (24, 38) when identical doses of meso- and rac-DMSA were administered to rats. Analysis of the distribution model of zinc predicts that much more than a 2-fold amount of zinc will be mobilized by rac-DMSA than its meso isomer (32), and therefore, there should be a concern about a potential enhancement of endogenous zinc loss when rac-DMSA is used as a clinical chelator. To prevent the loss of endogenous zinc in chelation therapy using rac-DMSA as a chelator, ZnLz of rac-DMSA instead of rac-DMSA alone was proposed by Fang and Fernando (32) as an effective chelator for the treatment of lead poisoning. This suggestion was made on the basis of an analysis of a hypothetical solution containing equimolar (0.25 mM) amounts of rac-DMSA, meso-DMSA, zinc ion, and lead ion. The analysis was performed by assuming that only the PbL complexes of the DMSA stereoisomers were formed in the solution. There have been some concerns about the validity of extending the results of this analysis to a n in vivo system, because (1)0.25 mM (5.2 mg/dL) of lead is too high in comparison with the lead burden in human blood; (2) the species, HPbL, of the DMSA stereoisomers were not included in this analysis; (3) ZnLp
Chem. Res. Toxicol., Vol. 8, No. 4, 1995 535
Lead Chelates of meso- and rac-DMSA Table 3. Equilibrium Concentrations of Pb(I1) and Zn(1I) in Solutions Containing 2.2 pM Pb(I1) and ZnLz of rac-DMSA or meso-DMSA at pH = 7.4, Ionic Strength 01) = 0.10, and T = 25.0 "C total concn @M) ZnLz 2.2
10 100 500
meso-DMSA 4.4 20 200 1000
[Pb2+l( x 180" 12a 1.7" 0.5=
M) 3066 3S6 3.46 0.76
[Zn2+lbg/dL) 2.5 0.3 0.07 0.03
Values determined from the solutions containing ZnLz of racDMSA. bValues determined from the solutions containing mesoDMSA alone. a
of rac-DMSA is a labile chelating agent which releases zinc ion in the presence of other chelating agents, but the RPMI assessment is only suitable for nonlabile chelating agents; therefore, the results obtained from this analysis are more appropriate for a comparison between the zinc chelates of the DMSA stereoisomers rather than a comparison between meso-DMSA and the chelate ZnLz of ruc-DMSA for their lead-mobilizing efficacies. The capability of ZnLz of rac-DMSA in mobilizing lead ions, therefore, was reassessed with all the concerns raised above. In the following discussion, the capability of the chelating agent (either ZnLz of rac-DMSA or meso-DMSA alone) in mobilizing lead ion is represented by the concentration of free lead ion present in the solution after a constant number of moles of the chelating agent is added. The total concentrations of lead used in the calculations were kept constant at 2.2 pM; the total concentrations of the ligand were 2.2, 10, 100, and 500 pM, respectively, if ZnL2 of rac-DMSA was assessed, or 4.4, 20, 200, and 1000 pM, respectively, if meso-DMSA was assessed. The calculations were performed with the aid of the SPE program (341, and the results are listed in Table 3. Shown in Table 3, a small amount, pg/dL to sub-pg/dL, of free zinc ion is maintained in the solution by the chelating agent itself if ZnLz of rac-DMSA is selected. This low level of free zinc ions will depress the in vivo competition of rac-DMSA for endogenous zinc with zinc-integrated proteins and, therefore, presumably prevent the increase in the loss of endogenous zinc, if ZnL2 of rac-DMSA is administered. It is predictable that, to a lower extent, the endogenous zinc will be depleted by ZnLz of rac-DMSA than by meso-DMSA alone. ZnLz of rac-DMSA is also superior to meso-DMSA in reducing the concentration of free lead ion in solution a t all the ligand concentration levels chosen, but as shown in Table 3, the superiority of ZnLz of rac-DMSA is marginal. In addition to the concern about the depletion of endogenous zinc, the main advantage of the proposed use of ZnLz of rac-DMSA over the current use of mesoDMSA in the clinical practice does not only rely on the consideration of its lead-mobilizing capability but also on the consideration of its high solubility, which facilitates gastrointestinal absorption of the drug after oral administration. At pH 2-3, which corresponds to the pH range in the stomach, ZnLz of rac-DMSA a t a concentration of 0.1 M remains in solution, forming mainly HzZnzL22-, H&, and H4L (28). About 20-25% of the ligand exists a s neutral H4L, which is highly lipophilic, and therefore will enhance the transport of the ligand through the gastrointestinal membranes. In conclusion, we postulate that ZnL2 of rac-DMSA is a safer and more effective chelator than meso-DMSA for the clinical treatment of lead poisoning.
( 111 1 ( 11 1 Figure 9. Proposed structures of (I) PbL2 of rac-DMSA, (11) PbL of rac-DMSA, and (111) PbL of meso-DMSA.
Binding of DMSA Stereoisomers with Pb2+in Aqueous Solution. The structures of PbL of the stereoisomers of DMSA in the solid state have been studied by Rivera et al. (27). Two forms of PbL of rac-DMSA were identified by IR spectroscopy. It was proposed that the ligand was bound to a lead ion via one sulfur atom and one oxygen atom in one form and via two sulfur atoms in the other form. Only one form of PbL of meso-DMSA, in which the ligand was bound to lead via one sulfur atom and one oxygen atom, was proposed by Rivera et al. (27). There is, however, no report in the literature on the structures of lead chelates of either DMSA stereoisomer formed in aqueous solution. The structures of some of the lead chelates of the DMSA stereoisomers in aqueous solution are postulated in this paper and illustrated in Figure 9. In PbL and PbL2 of rac-DMSA (Figure 9 (11) and (I))the individual ligand uses two sulfur atoms and one oxygen atom to bind with the lead ion. The absence of free thiolate groups in both chelates is deduced from their log protonation constants, 5.9 for PbL and 2.8 for PbLz; the participation of one carboxylate group in the coordination of Pb2+ is deduced from our IR evidence obtained with solutions containing the racemo ligands and lead ions a t varying ratios. In PbL of meso-DMSA (Figure 9 (111))the ligand also uses two sulfur atoms and one oxygen atom to bind with the lead ion. Similarly, the absence of free thiolate groups in the chelate is deduced from its log protonation constant, 6.6; the participation of one carboxylate group in the coordination with Pb2' is supported by IR results. It is not apparent on the basis of our IR evidence that the other carboxylate group of the ligand is free and not bound to the lead ion; however, from a viewpoint of energetics it is unlikely that both carboxylate and both thiolate groups of meso-DMSA bind with the lead ion simultaneously in the PbL complex. Our IR evidence also indicates that a lead ion binds to one carboxylate group of an individual ligand in the PbLz of meso-DMSA and to both carboxylate groups of a n individual ligand in the PbzL2 of rac-DMSA, but the manner in which the thiolate groups in these two chelates are involved in the coordination with Pb2+is still not clear because of the lack of knowledge of their protonation constants. Therefore, the structures of these two lead chelates of the stereoimers of DMSA are not shown in Figure 9. As shown in Figure 9, the ligands in the lead chelates of rac-DMSA (I and 11) adopt staggered anti conforma-
536 Chem. Res. Toxicol., Vol. 8, No. 4, 1995
tions with respect to the carboxylate groups, whereas the ligand in the lead chelate of meso-DMSA (111) adopts a staggered gauche conformation. Because the electrostatic repulsion between the two bulky carboxylate groups is minimized when they have a staggered anti relation, the lead chelates of rue-DMSA in aqueous solution are more stable than the corresponding chelates of its meso isomer. I t may be inferred, therefore, that the formation constants of the lead chelates of ruc-DMSA are larger than those of the corresponding chelates of its meso isomer. The results obtained from our potentiometric titrations, shown in Table 2, confirm this inference.
Acknowledgment. This work was supported in part by NIEHS Grant ES 03356.
References Schwartz, J., Angel, C., and Pitcher, H. (1986) Relationship between childhood blood levels and stature. Pediatrics 77, 281288. Fulton, M., Raab, G., Thomson, G., Laxen, D., Hunter, R., and Hepburn, W. (1987)The influence of blood lead on the ability and attainment of children in Edinburgh. Lancet I , 1221-25. Hatsakis, A,, Kokkevi, A,, Katsouvanni, K., Maravelias, K., Salaminios, F., et al. (1987)Lead exposure and children’s cognitive functions and behavior. In International Conference on Heavy Metals in the Environment (Lindberg, S. E., and Hutchinson, T. C., Eds.) Vol. 1, pp 204-209, Edinburgh, U.K. Bergomi, M., Borella, P., Fantuzzi, G., Vivoli, G., Sturloni, N., et al. (1989) Relationship between lead exposure indicators and neuromvcholo&al uerformance in children. Dev. Med. Child Neural. bl, 18:-196. Landsdown. R.. Urbanowicz. M. A,. and Hunter. J . (1986) The relationship between b1ood:level concentrations, intelligence, attainment and behavior in a school population: the second study. Int. Arch. Occup. Environ. Health 57, 225-235. Winneke, G., Brockhaus, A,, Ewers, U., Krammer, U.,and Neuf, M. (1990) Results from the European multicenter study for lead neurotoxicity in children: implications for risk assessment. Neurotoxicol. Teratol. 12, 553-559. Davis, J. M., and Svendsgaard, D. J. (1987) Lead and child development. Nature 329, 297-300. Angle, C. R. (1993) Childhood lead poisoning and its treatment. Annu. Reu. Pharmacol. Toxicol. 32, 409-434. Agency for Toxic Substances and Diseases Registry, Public Health Service, US DHHS (1988)The nature and extent of lead poisoning in children in the United States: A report to congress. Atlanta, GA. Centers for Disease Control, US Department of Health Human Service 11991) Preventing lead poisoning in young children: A statement by the Centers for Disease Control, October 1991, Atlanta, GA. Kosnett, M. J. (1992) Unanswered questions in metal chelation. Clin. Toxicol. 30, 529-547. FDA Medical Bulletin: Succimer approved for severe lead poisoning. Vol. 21, Part 1, p 5, March, 1991. Friedheim, E., Graziano, J. H., Popovac, D., Dragovic, D., and Kaul, B. (1978) Treatment of lead poisoning by 2,3-dimercaptosuccinic acid (DMSA). Lancet 2 , 1234-1236. Graziano, J. H., LoIacono, N. J., and Meyer, P. (1988) Doseresponse study of oral 2,3-dimercaptosuccinic acid in children with elevated blood lead concentrations. J . Pediatr. 113, 751-757. Bentur, Y., Brook, 3. G., Behar, R., and Taitelman, U. (1987) Meso-2,3-dimercaptosuccinic acid (DMSA) in the diagnosis and treatment of lead poisoning. Clin. Toxicol. 25, 39-51. Fournier, L., Thomas, G., Gamier, R., Buisine, A,, Houze, P., Pradier, S., and Dally, S. (1988) 2,3-Dimercaptosuccinic acid
Fang and Fernando treatment of heavy metal poisoning in humans. Med. Toxicol. 3, 499-504. Kuntzelman, D. R., England, K. E. and Angle, C. R. (1990) Urine lead (UPb) in outpatient treatment of lead poisoning with dimercaptosuccinic acid (DMSA). Vet. Hum. Toxicol. 4 , 364 (Abstract). Montalvan, J., Okose, P., and Marcus, S. (1990) Outpatient chelation therapy of 24 patients with lead intoxication by dimercaptosuccinic acid. Vet. Hum. Toxicol. 4, 343 (Abstract). Maiorino, R. M., Akins, J . M., Blaha, K., Castes, D. E., and Aposhian, H. V. (1990) Determination and metabolism of dithiolchelating agents: X. In humans, meso-2,3-dimercaptosuccinicacid is bound to plasma protein via mixed disulfide formation. J . Pharmacol. Exp. Ther. 254, 570-577. Aposhian, H. V., Maiorino, R. M., Dart, R. C., and Perry, D. F. (1989) Urinary excretion of meso-2,3-dimercaptosuccinic acid in human subjects. Clin. Pharmacol. Ther. 45, 520-526. Grandjean, P., Jacobsen, I. A,, and Jorgensen, P. J. (1991)Chronic lead poisoning treated with dimercaptosuccinic acid. Pharmacol. Toxicol. 68, 266-269. Fang, X., and Fernando, Q. (1994) Synthesis, structure, and properties of ruc-2,3-dimercaptosuccinic acid, a potentially useful chelating agent for toxic metals. Chem. Res. Toxicol. 7, 148-156. Okonishnikova, I. E. (1970) The comparative antidote potency of stereoisomeric dimercaptosuccinic acids in mercury poisoning. Gig. Tr. Prof. Zabol. 14, 18-22. Okonishnikova, I. E. (1971) The effect of steric isomers of dimercaptosuccinic acid on the elimination of some heavy metals from body. Gig. Tr. Prof. Zabol. 15, 50-52. Johannsen, B., Spies, H., and Syhre, R. (1978)Meso and racemic2,3-dimercaptosuccinatotechnetate(V).Radiochem. Radioanal. Lett. 36, 111-116. Awh, 0. D., Kim, J. R., and Shim, S. C. (1989) 99Tcm-diastereoisomeric dimercaptosuccinic acid complexes: comparison of their biodistribution. Nucl. Med. Commun. 10, 449-454. Rivera, M., Aposhian, H. V., and Fernando, Q. (1989) Lead chelates of meso- and racemic-dimercaptosuccinic acid. J . Inorg. Biochem. 37, 283-293. Fang, X., and Fernando, Q. (1994) Conformations of zinc chelates of meso- and rac-2,3-dimercaptosuccinic acid in aqueous solution. Chem. Res. Toxicol. 7, 882-890. Egorova, L. G. (1972) Complex-forming properties of 2,3-dimercaptosuccinic acid. Complex formation of lead with meso-2,3dimercaptosuccinic acid. Zh. Obshch. Khim. 42, 2240-2245. Willes, M. J., and Williams, D. R. (1985) The selectivity of mesodimercaptosuccinic acid for lead(I1)in blood plasma. Inorg. Chim. Acta 106, L21-L22. Harris, W. R., Chen, Y., Stenback, J.,and Shah, B. (1991) Stability constants for dimercaptosuccinic acid with bismuth(III1, zinc(II), and lead(I1). J. Coord. Chem. 23, 173-186. Fang, X., and Fernando, Q.(1994) A comparative study of mesoand rac-2,3-dimercaptosuccinicacids and their zinc complexes in aqueous solution. Chem. Res. Toxicol. 7 , 770-778. Grim, W. M., 111,Fateley, W. G., and Grasselli, J. G. (1984)Fourier Transform Infrared Spectroscopy (Theophanides, T., Ed.) pp 2542, D. Reidel Publishing Co., Dordrecht. Martell, A. E., and Motekaitis, R. J. (1988) Determination and Use of Stability Constants, VCH Publishers, New York, NY. Meites, L., Ed. (1963)Handbook ofAnalytica1 Chemistry, Section 3, p 95, McGraw-Hill Book Co., New York, NY. Deacon, G. B., and Phillips, R. J. (1980) Relationships between the carbon-oxygen stretching frequencies of carboxylate complexes and the type of carboxylate coordination. Coord. Chem. Rev. 33, 227-250. NIST Critical Stability Constants of Metal Complexes Database, version 1.0, NIST Standards Reference Database 46, Gaithersburg, MD. Egorova, L. G., Okonishnikova, I. E., Nirenburg, V. L., and Postovskii, I. Y. (1972) Complex formation of certain metals with dimercaptosuccinic acid stereoisomer. Khim. Farmat. Zhur. 6 , 14-16.
TX940178X
