Activation and Denitrosylation of Procaspase-3 in KA-induced Excitotoxicity.

BACKGROUND: It has been reported that activation of glutamate kainate receptor subunit 2 (GluK2) subunit-containing glutamate receptors and the following Fas ligand(FasL) up-regulation, caspase-3 activation, result in delayed apoptosis-like neuronal death in hippocampus CA1 subfield after cerebral ischemia and reperfusion. Nitric oxide-mediated S-nitrosylation might inhibit the procaspase activation, whereas denitrosylation might contribute to cleavage and activation of procaspases. OBJECTIVES: The study aimed to elucidate the molecular mechanisms underlying procaspase-3 denitrosylation and activation following kainic acid (KA)-induced excitotoxicity in rat hippocampus. METHODS: S-nitrosylation of procaspase-3 was detected by biotin-switch method. Activation of procaspase-3 was shown as cleavage of procaspase-3 detected by immunoblotting. FasL expression was detected by immunoblotting. Cresyl violets and TdT-mediated dUTP Nick-End Labeling (TUNEL) staining were used to detect apoptosis-like neuronal death in rat hippocampal CA1 and CA3 subfields. RESULTS: KA led to the activation of procaspase-3 in a dose- and time-dependent manner, and the activation was inhibited by KA receptor antagonist NS102. Procaspase-3 was denitrosylated at 3 h after kainic acid administration, and the denitrosylation was reversed by SNP and GSNO. FasL ASODNs inhibited the procaspase-3 denitrosylation and activation. Moreover, thioredoxin reductase (TrxR) inhibitor auranofin prevented the denitrosylation and activation of procaspase-3 in rat hippocampal CA1 and CA3 subfields. NS102, FasL AS-ODNs, and auranofin reversed the KAinduced apoptosis and cell death in hippocampal CA1 and CA3 subfields. CONCLUSIONS: KA led to denitrosylation and activation of procaspase-3 via FasL and TrxR. Inhibition of procaspase-3 denitrosylation by auranofin, SNP, and GSNO played protective effects against KA-induced apoptosis-like neuronal death in rat hippocampal CA1 and CA3 subfields. These investigations revealed that the procaspase-3 undergoes an initial denitrosylation process before becoming activated, providing valuable insights into the underlying mechanisms and possible treatment of excitotoxicity.

GPR55 contributes to neutrophil recruitment and mechanical pain induction after spinal cord compression in mice.

Pain transmission and processing in the nervous system are modulated by various biologically active substances, including lysophospholipids, through direct and indirect actions on the somatosensory pathway. Lysophosphatidylglucoside (LysoPtdGlc) was recently identified as a structurally unique lysophospholipid that exerts biological actions via the G protein-coupled receptor GPR55. Here, we demonstrated that GPR55-knockout (KO) mice show impaired induction of mechanical pain hypersensitivity in a model of spinal cord compression (SCC) without the same change in the models of peripheral tissue inflammation and peripheral nerve injury. Among these models, only SCC recruited peripheral inflammatory cells (neutrophils, monocytes/macrophages, and CD3+ T-cells) in the spinal dorsal horn (SDH), and GPR55-KO blunted these recruitments. Neutrophils were the first cells recruited to the SDH, and their depletion suppressed the induction of SCC-induced mechanical hypersensitivity and inflammatory responses in compressed SDH. Furthermore, we found that PtdGlc was present in the SDH and that intrathecal administration of an inhibitor of secretory phospholipase A2 (an enzyme required for producing LysoPtdGlc from PtdGlc) reduced neutrophil recruitment to compressed SDH and suppressed pain induction. Finally, by screening compounds from a chemical library, we identified auranofin as a clinically used drug with an inhibitory effect on mouse and human GPR55. Systemically administered auranofin to mice with SCC effectively suppressed spinal neutrophil infiltration and pain hypersensitivity. These results suggest that GPR55 signaling contributes to the induction of inflammatory responses and chronic pain after SCC via the recruitment of neutrophils and may provide a new target for reducing pain induction after spinal cord compression, such as spinal canal stenosis.

Auranofin attenuates hepatic steatosis and fibrosis in nonalcoholic fatty liver disease via NRF2 and NF- κB signaling pathways.

BACKGROUND/AIMS: We aim to evaluate the effects of auranofin, a known antioxidant, on hepatic steatosis, inflammation, and fibrosis, contributing to non-alcoholic steatohepatitis (NASH) development in vivo and in vitro. METHODS: Transcriptome analysis of LX-2 cells was that expression patterns of genes changed by auranofin, and their related pathways were estimated. We used the gene set enrichment analysis (GSEA) program to determine the pathway involved in overall genetic change. In vitro, LX-2 and HepG2 cells were treated with transforming growth factor (TGF)-ß1 and palmitic acid (PA), respectively, and the antifibrotic and antiadipogenic effect function of auranofin was evaluated. RESULTS: Transcriptome analysis revealed that auranofin decreased the expression of 15 genes, including thrombospondin 1, endothelin 1 (ET-1), fibronectin 1, and LOX. The molecular functions of these genes are involved in collagen binding. GSEA of the overall gene expression pattern revealed that many genes increased in the reactive oxygen species pathway and decreased in the inflammatory response. Auranofin decreased nuclear factor kappa B (NF-κB) and IκBα in TGF-ß1-induced LX-2 cells, thereby reducing ET-1 and fibrosis. Furthermore, increased pNRF2 in PA-induced HepG2 cells led to increased antioxidant marker expression and decreased lipid accumulation. In the bile duct ligation model mice, auranofin reduced the fibrosis area and increased the survival rate. Auranofin reduced liver fibrosis and lipid accumulation in NASH model mice fed on a Western diet. CONCLUSION: Auranofin inhibits lipogenesis and fibrosis formation and is a proposed candidate for NASH treatment.

NMR reveals the metabolic changes induced by auranofin in A2780 cancer cells: evidence for glutathione dysregulation.

NMR metabolomics represents a powerful tool to characterize the cellular effects of drugs and gain detailed insight into their mode of action. Here, we have exploited NMR metabolomics to illustrate the changes in the metabolic profile of A2780 ovarian cancer cells elicited by auranofin (AF), a clinically approved gold drug now repurposed as an anticancer agent. An early and large increase in intracellular glutathione is highlighted as the main effect of the treatment accompanied by small but significant changes in the levels of a few additional metabolites; the general implications of these findings are discussed in the frame of the current mechanistic knowledge of AF.

Native mass spectrometry of human carbonic anhydrase I and its inhibitor complexes.

Native mass spectrometry is a potent technique to study and characterize biomacromolecules in their native state. Here, we have applied this method to explore the solution chemistry of human carbonic anhydrase I (hCA I) and its interactions with four different inhibitors, namely three sulfonamide inhibitors (AAZ, MZA, SLC-0111) and the dithiocarbamate derivative of morpholine (DTC). Through high-resolution ESI-Q-TOF measurements, the native state of hCA I and the binding of the above inhibitors were characterized in the molecular detail. Native mass spectrometry was also exploited to assess the direct competition in solution among the various inhibitors in relation to their affinity constants. Additional studies were conducted on the interaction of hCA I with the metallodrug auranofin, under various solution and instrumental conditions. Auranofin is a selective reagent for solvent-accessible free cysteine residues, and its reactivity was analyzed also in the presence of CA inhibitors. Overall, our investigation reveals that native mass spectrometry represents an excellent tool to characterize the solution behavior of carbonic anhydrase.

Selenium-independent antioxidant and anti-inflammatory effects of thioredoxin reductase inhibition in alveolar macrophages.

AIMS: Interleukin-1ß (IL-1ß) contributes to the development of bronchopulmonary dysplasia (BPD). Thioredoxin reductase-1 (Txnrd1) inhibition activates nuclear factor erythroid 2-related factor 2 (Nrf2)-dependent responses. Txnrd1 activity is selenium (Se) dependent and Se deficiency is common in prematurity. Auranofin (AFN), a Txnrd1 inhibitor, decreases IL-1ß levels and increases Nrf2 activation in lipopolysaccharide (LPS) treated alveolar macrophages. In lung epithelia, AFN-induced Nrf2 activation is Se dependent. We tested the hypothesis that the effects of Txnrd1 inhibition in alveolar macrophages are Se dependent. MAIN METHODS: To establish Se sufficient (Se+) and deficient (Se-) conditions, alveolar (MH-S) macrophages were cultured in 2.5% fetal bovine serum (FBS) ± 25 nM Na2SeO3. Se- (2.5% FBS) and Se+ (2.5% FBS + 25 nM Na2SeO3) cells were cultured in the presence or absence of 0.05 µg/mL LPS and/or 0.5 µM AFN. Nrf2 activation was determined by measuring NADPH quinone oxidoreductase-1 (Nqo1) and glutathione levels. IL-1ß mRNA (Il1b) and protein levels were measured using qRT-PCR and ELISA. Data were analyzed by ANOVA followed by Tukey's post-hoc. KEY FINDINGS: We detected an independent effect of AFN, but not LPS, on Nqo1 expression and GSH levels in Se+ and Se- cells. LPS significantly increased Il1b and IL-1ß levels in both groups. AFN-mediated attenuation of this effect was not impacted by Se status. SIGNIFICANCE: The beneficial effects of Txnrd1 inhibition in alveolar macrophages are Se-independent and therefore unlikely to be diminished by clinical Se deficiency.

Formation of oxidised (OX) proteins in Entamoeba histolytica exposed to auranofin and consequences on the parasite virulence.

Metronidazole (MNZ), the first line drug for amoebiasis and auranofin (AF), an emerging antiprotozoan drug, are both inhibiting Entamoeba histolytica thioredoxin reductase. The nature of oxidised proteins (OXs) formed in AF- or MNZ-treated E. histolytica trophozoites is unknown. In order to fill this knowledge gap, we performed a large-scale identification and quantification of the OXs formed in AF- or MNZ-treated E. histolytica trophozoites using resin-assisted capture coupled to mass spectrometry (MS). We detected 661 OXs in MNZ-treated trophozoites and 583 OXs in AF-treated trophozoites. More than 50% of these OXs were shared, and their functions include hydrolases, enzyme modulators, transferases, nucleic acid binding proteins, oxidoreductases, cytoskeletal proteins, chaperones, and ligases. Here, we report that the formation of actin filaments (F-actin) is impaired in AF-treated trophozoites. Consequently, their erythrophagocytosis, cytopathic activity, and their motility are impaired. We also observed that less than 15% of OXs present in H2 O2 -treated trophozoites are also present in AF- or MNZ-treated trophozoites. These results strongly suggest that the formation of OXs in AF- or MNZ-treated trophozoites and in H2 O2 -treated trophozoites occurred by two different mechanisms.

An investigation into the interactions of gold nanoparticles and anti-arthritic drugs with macrophages, and their reactivity towards thioredoxin reductase.

Gold(I) complexes are an important tool in the arsenal of established approaches for treating rheumatoid arthritis (RA), while some recent studies have suggested that gold nanoparticles (Au NPs) may also be therapeutically efficacious. These observations prompted the current biological studies involving gold(I) anti-RA agents and Au NPs, which are aimed towards improving our knowledge of how they work. The cytotoxicity of auranofin, aurothiomalate, aurothiosulfate and Au NPs towards RAW264.7 macrophages was evaluated using the MTT assay, with the former compound proving to be the most toxic. The extent of cellular uptake of the various gold agents was determined using graphite furnace atomic absorption spectrometry, while their distribution within macrophages was examined using microprobe synchrotron radiation X-ray fluorescence spectroscopy. The latter technique showed accumulation of gold in discrete regions of the cell, and co-localisation with sulfur in the case of cells treated with aurothiomalate or auranofin. Electrospray ionization mass spectrometry was used to characterize thioredoxin reductase (TrxR) in which the penultimate selenocysteine residue was replaced by cysteine. Mass spectra of solutions of TrxR and aurothiomalate, aurothiosulfate or auranofin showed complexes containing bare gold atoms bound to the protein, or protein adducts containing gold atoms retaining some of their initial ligands. These results support TrxR being an important target of gold(I) drugs used to treat RA, while the finding that Au NPs are incorporated into macrophages, but elicit little toxicity, indicates further exploration of their potential for treatment of RA is warranted.

The mode of action of anticancer gold-based drugs: a structural perspective.

The interactions between a few representative gold-based drugs and hen egg white lysozyme were studied by X-ray crystallography. High resolution crystal structures solved for three metallodrug-protein adducts provide valuable insight into the molecular mechanism of these promising metal compounds and the inherent protein metalation processes.

Total reflection X-ray fluorescence spectrometry as a tool for the quantification of gold and platinum metallodrugs: determination of recovery rates and precision in the ppb concentration range.

Total reflection X-ray spectrometry is an instrumental technique based on X-ray fluorescence, which offers detection limits low enough to quantify trace element concentrations with negligible interference from matrix components. The technique is well established in material sciences and now reaches out to extended applications in life sciences and pharmaceutical quality control. In our study we focused on possible applications for the quantification of gold and platinum containing active ingredients in trace concentrations (ppb range) in a matrix of biological origin (cell suspensions). General aspects of sample preparation as well as selected important method performance parameters (precision, recovery rates) were investigated. Overall, TXRF represents an useful option to quantify metals in ppb concentrations with acceptable precision and recovery.

Proteomic analysis of ovarian cancer cell responses to cytotoxic gold compounds.

Platinum-based chemotherapy is the primary treatment for human ovarian cancer. Overcoming platinum resistance has become a critical issue in the current chemotherapeutic strategies of ovarian cancer as drug resistance is the main reason for treatment failure. Cytotoxic gold compounds hold great promise to reach this goal; however, their modes of action are still largely unknown. To shed light on the underlying molecular mechanisms, we performed 2-DE and MS analysis to identify differential protein expression in a cisplatin-resistant human ovarian cancer cell line (A2780/R) following treatment with two representative gold compounds, namely Auranofin and Auoxo6. It is shown that Auranofin mainly acts by altering the expression of Proteasome proteins while Auoxo6 mostly modifies proteins related to mRNA splicing, trafficking and stability. We also found that Thioredoxin-like protein 1 expression is greatly reduced after treatment with both gold compounds. These results are highly indicative of the likely sites of action of the two tested gold drugs and of the affected cellular functions. The implications of the obtained results are thoroughly discussed in the frame of current knowledge on cytotoxic gold agents.

Mitochondrial respiratory chain involvement in peroxiredoxin 3 oxidation by phenethyl isothiocyanate and auranofin.

Mitochondrial peroxiredoxin 3 (Prx 3) is rapidly oxidized in cells exposed to phenethyl isothiocyanate (PEITC) and auranofin (AFN), but the mechanism of oxidation is unclear. Using HL-60 cells deplete of mitochondrial DNA we show that peroxiredoxin 3 oxidation and cytotoxicity requires a functional respiratory chain. Thioredoxin reductase (TrxR) could be inhibited by up to 90% by auranofin without direct oxidation of peroxiredoxin 3. However, inhibition of thioredoxin reductase promoted peroxiredoxin 3 oxidation and cytotoxicity in combination with phenethyl isothiocyanate or antimycin A. We conclude that rapid peroxiredoxin 3 oxidation occurs as a consequence of increased oxidant production from the mitochondrial respiratory chain.

Free thiol group of MD-2 as the target for inhibition of the lipopolysaccharide-induced cell activation.

MD-2 is a part of the Toll-like 4 signaling complex with an indispensable role in activation of the lipopolysaccharide (LPS) signaling pathway and thus a suitable target for the therapeutic inhibition of TLR4 signaling. Elucidation of MD-2 structure provides a foundation for rational design of inhibitors that bind to MD-2 and inhibit LPS signaling. Since the hydrophobic binding pocket of MD-2 provides little specificity for inhibitors, we have investigated targeting the solvent-accessible cysteine residue within the hydrophobic binding pocket of MD-2. Compounds with affinity for the hydrophobic pocket that contain a thiol-reactive group, which mediates covalent bond formation with the free cysteine residue of MD-2, were tested. Fluorescent compounds 2-(4'-(iodoacetamido)anilino)naphthalene-6-sulfonic acid and N-pyrene maleimide formed a covalent bond with MD-2 through Cys(133) and inhibited LPS signaling. Cell activation was also inhibited by thiol-reactive compounds JTT-705 originally targeted against cholesterol ester transfer protein and antirheumatic compound auranofin. Oral intake of JTT-705 significantly inhibited endotoxin-triggered tumor necrosis factor alpha production in mice. The thiol group of MD-2 also represents the target of environmental or endogenous thiol-reactive compounds that are produced in inflammation.

In vitro distribution of gold in serum proteins after incubation of sodium aurothiomalate and auranofin with human blood and its pharmacological significance.

This study presents a comparative drug-protein, in vitro, binding profile of sodium aurothiomalate and auranofin. It was found that about 40% of total protein-bound gold is attached to albumin after incubation of aurothiomalate with whole blood for 24 h and about 29% of it was with alpha(1)-globulin and the least amount was found with gamma-globulin (6.1%). On the other hand, approximately 84% of the protein-bound auranofin gold attached to globulins of which 51% was found with beta-globulin band. It was almost equally distributed among albumin, alpha(2)-globulin and gamma-globulin, and showed least affinity for alpha(1)-globulin. The gold analyses were performed by standardless instrumental neutron activation method duly validated by use of an established atomic absorption method. The results of this study explain to some extent the difference in, in vivo, pharmacokinetics and pharmacodynamics of the two drugs.

Auranofin disrupts selenium metabolism in Clostridium difficile by forming a stable Au-Se adduct.

Clostridium difficile is a nosocomial pathogen whose incidence and importance are on the rise. Previous work in our laboratory characterized the central role of selenoenzyme-dependent Stickland reactions in C. difficile metabolism. In this work we have identified, using mass spectrometry, a stable complex formed upon reaction of auranofin (a gold-containing drug) with selenide in vitro. X-ray absorption spectroscopy supports the structure that we proposed on the basis of mass-spectrometric data. Auranofin potently inhibits the growth of C. difficile but does not similarly affect other clostridia that do not utilize selenoproteins to obtain energy. Moreover, auranofin inhibits the incorporation of radioisotope selenium ((75)Se) in selenoproteins in both Escherichia coli, the prokaryotic model for selenoprotein synthesis, and C. difficile without impacting total protein synthesis. Auranofin blocks the uptake of selenium and results in the accumulation of the auranofin-selenide adduct in the culture medium. Addition of selenium in the form of selenite or L-selenocysteine to the growth medium significantly reduces the inhibitory action of auranofin on the growth of C. difficile. On the basis of these results, we propose that formation of this complex and the subsequent deficiency in available selenium for selenoprotein synthesis is the mechanism by which auranofin inhibits C. difficile growth. This study demonstrates that targeting selenium metabolism provides a new avenue for antimicrobial development against C. difficile and other selenium-dependent pathogens.

Identifying and characterizing the biological targets of metallotherapeutics: Two approaches using Au(I)-protein interactions as model systems.

A significant challenge in the field of medicinal inorganic chemistry is the identification of biological targets of metal-based drugs and the characterization of the metal-biomolecule adducts. A classic example is Au(I), which has long been used to treat rheumatoid arthritis despite a poor understanding of its biological targets due to the lability, reactivity, and "spectroscopic silence" that are characteristic of Au(I). Here, we report two qualitative methods for characterizing Au(I)-protein adducts: a thiol-reactive probe that facilitates the identification of biological cysteine-Au(I) adducts and a photoreactive Au(I) complex that produces a covalent bond between the Au(I) complex and the biomolecule.

A mass spectrometric investigation of the binding of gold antiarthritic agents and the metabolite [Au(CN)2]- to human serum albumin.

Electrospray ionisation (ESI) mass spectrometry was used to examine the reactions of the clinically used antiarthritic agent [Au(S2O3)2]3-, and AuPEt3Cl, a derivative of another clinically used agent auranofin, with human serum albumin (HSA) obtained from a human volunteer. Both compounds reacted readily with HSA to form complexes containing one or more covalently attached gold fragments. In the case of AuPEt3Cl, binding was accompanied by the loss of the chloride ligand, while for [Au(S2O3)2]3- the mass spectral data indicated binding of Au(S2O3) groups. Experiments performed using HSA with Cys34 blocked by reaction with iodoacetamide were consistent with reaction of both gold compounds with this amino acid. Separate blocking experiments using diethylpyrocarbonate and AuPEt3Cl also provided evidence for histidine residues acting as lower-affinity binding sites for this gold compound. ESI mass spectra of solutions containing [Au(S2O3)2]3- or [Au(CN)2]-, and HSA, provided evidence for the formation of protein complexes in which intact gold molecules were non-covalently bound. In the case of [Au(S2O3)2]3-, these non-covalent complexes proved to be transitory in nature. However, for [Au(CN)2]- a non-covalent complex containing a single gold molecule bound to HSA was found to be stable, and constituted the main adduct formed in solutions containing low-to-medium Au-to-HSA ratios. Evidence was also obtained for the formation of a covalent adduct in which a single Au(CN) moiety was bonded to Cys34 of the protein. AuPEt3Cl reacted to a much lower extent with HSA that had Cys34 modified by formation of a disulfide bond to added cysteine, than with unmodified HSA. This suggests that the extent of modification of the protein in vivo may have an important influence on the transport and bioavailability of gold antiarthritic drugs.

(13)C, (31)P and (15)N NMR studies of the ligand exchange reactions of auranofin and chloro(triethylphosphine)gold(I) with thiourea.

The interaction of thiourea (Tu) with auranofin (Et(3)PAuSATg) and its analogue, Et(3)PAuCl has been studied using (13)C, (31)P and (15)N NMR spectroscopy. It is observed that Tu is able to replace both the ligands, Et(3)P and SATg(-) simultaneously from gold(I) in auranofin, forming [Et(3)P-Au-Tu](+) and Tu-Au-SATg complexes. However, no separate resonances for these species were observed either due to their rapid exchange with auranofin and thus giving only the average resonances or because the chemical shifts of either two species are same so that they cannot be resolved. The displaced SATg(-) is oxidized to its disulfide, (SATg)(2). However, some of the displaced Et(3)P is oxidized to Et(3)PO while the remaining reacts with Tu to form Et(3)P-Tu species, characterized by delta 31P of 1.0 ppm, assigned after an independent reaction between Et(3)P and Tu. In an experiment using a 0.05 M solution of auranofin, the Et(3)PO resonance appeared in auranofin spectrum after 4 days of addition of 1.0 equivalent of Tu, showing that the reaction is slow. A resonance for free Et(3)P is also detected in 31P NMR on the addition of CN(-). It is also observed that Tu reacts with Et(3)PAuCl to form [Et(3)P-Au-Tu](+) via displacement of Cl(-), consistent with an upfield shift of 6.2 ppm in >C [double bond] S resonance of Tu in (13)C NMR. In (15)N NMR, a smaller downfield, instead of an upfield shift, in NH(2) resonance of Tu on its addition to auranofin and Et(3)PAuCl indicates that it is not binding to gold(I) through nitrogen.

Comparative (13)C and (31)P NMR studies of the ligand exchange reactions of auranofin with ergothionine, imidazolidine-2-thione and diazinane-2-thione.

The interaction of auranofin (Et(3)PAuSATg) with ergothionine (ErS), imidazolidine-2-thione (Imt) and diazinane-2-thione (Diaz) has been studied using (13)C and (31)P NMR spectroscopy. It is observed that these thiones are able to replace both Et(3)P and SATg(-) ligands simultaneously from gold(I) in auranofin forming >C [double bond] S-Au-SATg and [Et(3)P-Au-S [double bond] C<](+) type complexes. The displaced SATg(-) is oxidized to its disulfide (SATg)(2). However, some of the displaced Et(3)P is oxidized to Et(3)PO while the remaining reacts with thiones to form Et(3)P-S [double bond] C< species characterized by delta (31)P NMR of 1.0-1.5 ppm. The Et(3)PO resonance appeared in the 31P NMR spectrum, after 10 days of the addition of ErS, after 19 days of the addition of Imt and after 6 days of the addition of Diaz, to auranofin solution showing that the thiones react with auranofin very slowly.

Inhibition of erythrocyte selenium-glutathione peroxidase by auranofin analogues and metabolites.

The effect of gold ligation on the inhibition of bovine erythrocyte selenium-glutathione peroxidase (GSH-Px) was examined. The anti-arthritic drug auranofin [2,3,4,6-tetra-O-acetyl-1-thio-beta-D-glucopyranosato-S)(triethylp hosphine) gold(I)] (Et3PAuSATg) and its analogue, Et3PAuCl, exhibited experimentally equivalent Ki values (11.6+/-0.8 and 10.8+/-0.5 microM, respectively), despite the greatly disparate affinities of their ligands for gold(I): 2,3,4,6-tetra-O-acetyl-1-thiolato-beta-D-glucopyranose (ATgS-) >> Cl-. This similarity reflects ligand exchange reactions that generate the glutathione complex Et3PAuSG from the excess glutathione (GSH, 1 mM) used in the assay. The Ki values for bis(glutathionato)gold(l) (Au(SG)2-) and gold(I) thioglucose (AuSTg) were also found to be equal (2.8+/-0.4 and 2.4+/-0.5 microM, respectively). This confirms the previous postulate of Chaudiere and Tappel (J Inorg Biochem 20: 313-325, 1984) that Au(SG)2- is generated from AuSTg in the presence of excess glutathione. Since auranofin metabolites accumulate in red blood cells, the inhibition of intracellular GSH-Px was examined by using intact erythrocytes. There was greater inhibition of the reaction when the cells were resuspended in isotonic buffer than in whole blood, because serum albumin in the latter competes for the auranofin and decreases the uptake by erythrocytes. After correction for the extent of gold uptake, the Ki values were determined to be the same as those observed for Au(SG)2- in the extracellular assay, indicating loss of both the Et3P and ATgS- ligands from auranofin. Thus, the inhibition of GSH-Px by gold complexes is dependent on their ligation, and the ultimate gold(I) compound that interacts with erythrocyte GSH-Px in intact red cells, Au(SG)2-, is radically different from the original auranofin molecule.