Crystal structures of reduced, oxidized, and mutated human thioredoxins: evidence for a regulatory homodimer.

BACKGROUND: Human thioredoxin reduces the disulfide bonds of numerous proteins in vitro, and can activate transcription factors such as NFkB in vivo. Thioredoxin can also act as a growth factor, and is overexpressed and secreted in certain tumor cells. RESULTS: Crystal structures were determined for reduced and oxidized wild type human thioredoxin (at 1.7 and 2.1 A nominal resolution, respectively), and for reduced mutant proteins Cys73-->Ser and Cys32-->Ser/Cys35-->Ser (at 1.65 and 1.8 A, respectively). Surprisingly, thioredoxin is dimeric in all four structures; the dimer is linked through a disulfide bond between Cys73 of each monomer, except in Cys73-->Ser where a hydrogen bond occurs. The thioredoxin active site is blocked by dimer formation. Conformational changes in the active site and dimer interface accompany oxidation of the active-site cysteines, Cys32 and Cys35. CONCLUSIONS: It has been suggested that a reduced pKa in the first cysteine (Cys32 in human thioredoxin) of the active-site sequence is important for modulation of the redox potential in thioredoxin. A hydrogen bond between the sulfhydryls of Cys32 and Cys35 may reduce the pKa of Cys32 and this pKa depression probably results in increased nucleophilicity of the Cys32 thiolate group. This nucleophilicity, in tum, is thought to be necessary for the role of thioredoxin in disulfide-bond reduction. The physiological role, if any, of thioredoxin dimer formation remains unknown. It is possible that dimerization may provide a mechanism for regulation of the protein, or a means of sensing oxidative stress.

Effect of some anticancer drugs on the surface membrane electrical properties of differentiated murine neuroblastoma cells.

The effect of some anticancer agents that produce toxic effects on electrically excitable cells in vivo was studied in vitro with the use of differentiated N1E-115 murine neuroblastoma cells and single microelectrode electrical recording. In the presence of 10(-7) g/ml tetrodotoxin, following the release of 500-millisecond conditioning hyperpolarization, the cells exhibited Ca2+-dependent action potentials. Local application to N1E-115 neuroblastoma cells of cisplatin (cis-PDD) for 30 seconds from a drug-containing effusion pipette produced a dose-dependent reversible inhibition of the Ca2+-dependent action potential, with a 61% inhibition at 1.7 microM and 67% inhibition at 17 microM cis-PDD. trans-Dichlorodiammineplatinum(II) and platinic(IV) chloride, both of which lacked the growth inhibitory properties of cis-PDD against N1E-115 neuroblastoma cells, at concentrations of 170 and 120 microM produced only an 11 and 19% inhibition of the Ca2+-dependent action potential, respectively. Vincristine at a concentration of 1 microM reversibly inhibited the Ca2+-dependent action potential by 48%. 3'-Deamino-3'-(3''-cyano-4''-morpholinyl)doxorubicin, a more potent experimental antitumor agent than doxorubicin, at 10(-8) M inhibited the Ca2+-dependent action potential by 22%, similar to the inhibition previously reported for doxorubicin. None of the agents affected the cell transmembrane potential, which suggests a lack of an effect on the mechanisms responsible for maintaining the resting cell membrane potential difference. The effects of the agents on the Ca2+-dependent action potential might reflect a direct effect on a plasma membrane Ca2+ channel or on the lipid domain around the channels, or they might be produced by changes in intracellular Ca2+ homeostasis, among other mechanisms. It is not known whether a change in the membrane Ca2+ current is related to the antitumor effects of the agents, but such a change may contribute to the neurotoxicity of cis-PDD and vincristine and the cardiac toxicity of the anthracycline.

Disposition of bisantrene in humans and rabbits: evidence for intravascular deposition of drug as a cause of phlebitis.

The investigational antitumor agent bisantrene (9,10-anthracenedicarboxaldehyde bis[(4,5-dihydro-1H-imidazol-2-yl)hydrazone] dihydrochloride) causes frequent local complications of phlebitis and thromboses in patients receiving the drug by peripheral venous infusion. Bisantrene pharmacokinetics was studied in five patients. Plasma elimination was biphasic with t1/2 alpha of 65 min and t1/2 beta of 1142 min; the mean apparent volumes of distribution of the central compartment and peripheral compartments were 185 and 1662 liters/sq m, suggesting extensive uptake, binding, or deposition of drug. Total body clearance was 735 ml/min/sq m, and 11.3% of drug was excreted in urine. One hr after bisantrene (260 mg/sq m) at 1 mg/ml in 5% dextrose was infused into the marginal ear vein of a rabbit, the vein was congested with blood and contained 2.1 mg precipitated bisantrene. After 24 hr, the vein was clotted and contained 1.18 mg precipitated drug. Precipitation of bisantrene appears to be related to the low solubility of the drug at physiological pH. Maximum solubility of bisantrene in human and rabbit serum was 12.7 micrograms/ml. Intravascular precipitation of bisantrene may be responsible for phlebitis and thromboses in humans receiving the drug by i.v. infusion.

Metabolism and disposition of 3-amino-1,5-dihydro-5-methyl-1-beta-D-ribofuranosyl-1,4,5,6, 8-pentaazaacenaphthylene in the rat.

Tricyclic nucleoside (TCN, NSC 154020) is a 7-deazapurine nucleoside possessing antitumor activity towards certain tumor cell lines in vitro. In vivo, TCN is readily interconvertible with its 5'-monophosphate ester. The present study demonstrates that TCN is metabolized by rat liver microsomes to a ring-opened bicyclic metabolite with loss of cytotoxicity toward Chinese hamster ovary cells in culture. Metabolism is mediated by hydrogen peroxide generated by the rat liver microsomes but is not dependent on cytochrome P-450. Isolated hepatocytes prepared from rat do not form detectable amounts of the ring-opened bicyclic metabolite. Unchanged TCN and the ring-opened bicyclic metabolite are excreted in the bile of rats with a cannulated bile duct, comprising 42 and 12% of an i.v. dose of TCN in 8 hr, respectively. The ring-opened bicyclic metabolite is not formed by red blood cells in vitro and could not be detected in blood in vivo. The fact that the ring-opened bicyclic metabolite appears in bile suggests that liver cells not present or not active in isolated hepatocyte preparations might produce the metabolite. Alternatively, the metabolite might be formed directly from TCN in bile, perhaps by hydrogen peroxide excreted into bile. In vivo, 56% of radiolabel was found in the upper and lower gastrointestinal tract and the feces 24 hr after an i.p. or i.v. dose of 100 mg of [5-methyl-14C]TCN/sq m. Urinary excretion of radiolabel was 21% of the dose of [14C]TCN in 24 hr. Biliary excretion of radiolabel was 65% of the dose of [14C]TCN in 8 hr. The fraction of radioactivity undergoing enterohepatic cycling with reabsorption from the gastrointestinal tract after excretion in bile is 84%.

Cigarette smoking is a determinant of DT-diaphorase gene expression in human non-small cell lung carcinoma.

The levels of NAD(P)H:(quinone-acceptor) oxidoreductase (EC.1.6.99.2) (DT-diaphorase) mRNA and enzyme activity have been studied in paired human normal lung and non-small cell lung tumor samples from patients with a history of cigarette smoking. There were significantly higher levels of DT-diaphorase mRNA (1.2 kilobases) in lung tumor compared to normal lung tissue of patients who had stopped smoking more than 6 months before surgery, with relative values (normalized to beta-actin mRNA) of 29.6 +/- 7.8 (SE) in the lung tumor compared to 11.7 +/- 2.2 in normal lung tissue (P < 0.05). There was no significant difference in DT-diaphorase mRNA between lung tumor and normal lung tissue of subjects who were smokers at the time of surgery, with values of 16.5 +/- 2.1 and 15.3 +/- 2.5 (P > 0.05), respectively. DT-diaphorase enzyme activity in normal and tumor lung tissue was positively correlated with DT-diaphorase mRNA (r = 0.908, P < 0.01). The results of the study suggest that DT-diaphorase does not function as an inducible protectant enzyme in human lung against oxidant species and carcinogens present in cigarette smoke. Metabolism of some anticancer drugs by DT-diaphorase can alter their activity. Differences in DT-diaphorase between lung tumors of smokers and past smokers might alter the response to these drugs.

An alternatively spliced form of NQO1 (DT-diaphorase) messenger RNA lacking the putative quinone substrate binding site is present in human normal and tumor tissues.

DT-diaphorase is a ubiquitously expressed flavoenzyme responsible for the two-electron reduction of a number of quinone and other anticancer drugs. The majority of DT-diaphorase enzyme activity in human tissues is the product of the NQO1 gene. We have now identified a novel alternatively spliced form of human NQO1 mRNA lacking exon 4 at levels equal to or exceeding those of wild-type NQO1 mRNA. Exon 4 codes for the putative quinone substrate binding site of DT-diaphorase derived from NQO1 and the recombinant protein from alternatively spliced NQO1 mRNA lacking exon 4 has minimal enzyme activity with quinoid and other known substrates of DT-diaphorase. The physiological substrate of DT-diaphorase is unknown, and it is possible that the protein derived from the alternatively spliced NQO1 mRNA could have enzyme activity with an appropriate substrate. We found full-length DT-diaphorase protein but could not detect expression of an appropriately smaller form of DT-diaphorase in human tissues using polyclonal antibody to DT-diaphorase, suggesting that alternatively spliced NQO1 mRNA lacking exon 4 may not be translated or that the protein product is rapidly degraded. Alternative splicing of NQO1 RNA could provide an important mechanism for regulating NQO1 gene expression.

Mechanisms of the regulation of thioredoxin reductase activity in cancer cells by the chemopreventive agent selenium.

Selenium is an essential trace element, the deficiency of which is associated with an increased incidence of some human cancers. Dietary supplementation with selenium has been reported to produce a decrease in the incidence of some cancers in humans. Thioredoxin reductase (TR) is a newly discovered homodimeric selenocysteine (SeCys)-containing protein that catalyzes the NADPH-dependent reduction of the redox protein thioredoxin (Trx). Trx is overexpressed by a number of human tumors, and experimental studies have shown that Trx contributes to the growth and to the transformed phenotype of some human cancer cells. Thus, TR, by reducing Trx, could play a role in regulating the growth of normal and cancer cells. We have investigated mechanisms by which selenium, in the form of sodium selenite, added to serum-free growth medium regulates TR activity in cancer cell lines. Selenium caused a dose-dependent increase in cellular TR activity. The increase in TR activity produced by 1 microM Se compared to medium with no added selenium was: for MCF-7 breast cancer cells, 37-fold; for HT-29 colon cancer cells, 19-fold; and for A549 lung cancer cells, 8-fold. In contrast, Jurkat and HL-60 leukemia cells showed no increase in TR activity. The half-life of the time course of induction of TR in HT-29 cells after adding selenium was 10 h. The increase in TR activity was accompanied by an increase in TR protein levels up to 3-fold and an increase in the specific activity of the enzyme of 5-32-fold, depending on the cell line. Studies using 75Se showed that the amount of selenium incorporated into TR increased with increasing selenium concentration up to a ratio of 1 selenium per TR monomer. There was an increase in TR mRNA levels of 2-5-fold at 1 microM selenium and an increase in the stability of TR mRNA with a half-life for degradation of 21 h compared to 10 h in the absence of selenium. Trx mRNA and protein levels and Trx mRNA stability were not affected by selenium. The results of the study show that the increase in TR activity caused by selenium is specific and due to several effects, including an increase in the stability of TR mRNA leading to increased TR mRNA levels, an increase in TR protein, but predominantly to an increase in the specific activity of TR associated with increased incorporation of selenium into the enzyme.

Nuclear factor kappaB transactivation is increased but is not involved in the proliferative effects of thioredoxin overexpression in MCF-7 breast cancer cells.

Thioredoxin (Trx) is a small redox-active protein that provides reducing equivalents for key cysteine residues of proteins through thiol-disulfide exchange, such as the transcription factor nuclear factor-kappaB (NF-kappaB). NF-kappaB activation has been associated previously with cell growth and the inhibition of apoptosis. We have shown in earlier studies that overexpression of Trx in MCF-7 cells increases anchorage-independent growth. In this study, the activation of NF-kappaB was examined as a mechanism through which Trx overexpression might promote anchorage-independent growth. Constitutive NF-kappaB activity is elevated 4-7-fold in Trx-overexpressing cells. NF-kappaB activity was inhibited in these cells by expressing a dominant-negative mutant of the IkappaB alpha protein (IkappaB alphaM). Expression of IkappaB alphaM in Trx-overexpressing cells dramatically reduced the Trx-associated increase in NF-kappaB activity but did not affect anchorage-dependent or -independent growth. The results suggest that increased growth in MCF-7 cells overexpressing Trx is not mediated by increased activation of the transcription factor, NF-kappaB. Additionally, activator protein-1 (AP-1), another transcription factor associated with growth, was increased up to 10-fold in Trx-overexpressing cells. Thus, AP-1 activation might contribute to the growth-promoting effect of Trx.

Metabolic conversion of indicine N-oxide to indicine in rabbits and humans.

Indicine N-oxide, a pyrrolizidine alkaloid N-oxide that exhibits antitumor activity without some of the toxic effects associated with other pyrrolizidine alkaloids, is metabolized to indicine in rabbits and humans. Indicine can be detected in the plasma and is excreted in the urine in a dose-dependent manner following the i.v. administration of indicine N-oxide. The p.o. administration of indicine N-oxide leads to an increased plasma concentration and an increased urinary excretion of indicine. The hepatic microsomal fraction and the gut flora both catalyze the anaerobic reduction of indicine N-oxide to indicine in vitro. Whole-animal studies suggest that the gut flora play a major role in the metabolic reduction of indicine N-oxide by the rabbit. Indicine N-oxide is not actively excreted in the bile, and it is probable that indicine N-oxide finds its way into the gut by passive diffusion following i.v. administration. Neomycin and erythromycin, which reduce the number of anaerobic bacteria in the gut, lead to decreased plasma levels and a decreased urinary excretion of indicine.

Toxicity of intrathecally administered cytotoxic drugs and their antitumor activity against an intrathecal Walker 256 carcinosarcoma model for meningeal carcinomatosis in the rat.

Meningeal carcinomatosis in humans is refractory to most attempts at therapy and has a grave prognosis. As part of a search for new agents to treat meningeal carcinomatosis, the toxicity of a series of antitumor agents administered through an implanted catheter directly into the cerebrospinal fluid (CSF) of the lumbar spinal cord of 200-g rats has been examined. The highest non-toxic dose (HNTD) of the agents producing no signs of histological damage, motor dysfunction, or neurobehavioral changes was bleomycin (80 micrograms), cytarabine (640 micrograms), dacarbazine (1 microgram), doxorubicin (20 micrograms), 5-fluorouracil (150 micrograms), methotrexate (1000 micrograms), mitomycin C (10 micrograms), and triethylene phosphoramide (800 micrograms). No toxicity was observed at the highest dose that could be administered because of limited solubility of 1,3-bis(2-chloroethyl)-1-nitrosourea (100 micrograms), or diaziquone (70 micrograms). The rat model gave reasonable prediction of the toxicity of antitumor agents that have been administered intra-CSF to humans but with a tendency to underpredict toxicity. The antitumor activity of the agents administered intra-CSF at the HNTD against a Walker 256 carcinosarcoma model for meningeal carcinomatosis in the rat was examined. Diaziquone, doxorubicin, methotrexate, and mitomycin C gave a 25% or greater increase in lifespan and diaziquone and mitomycin C gave some long-term survivors. BCNU, bleomycin and 5-fluorouracil gave less than a 25% increase in lifespan. When mitomycin C was administered intra-CSF at the HNTD and diaziquone at the limit of solubility daily for 4 days there was no toxicity and an increase in survival of Walker 256 carcinosarcoma tumored animals compared to animals administered a single daily dose. Intrathecal diaziquone was more active against meningeal Walker 256 carcinosarcoma than i.v. diaziquone. Intravenous methotrexate had no activity against meningeal Walker 256 carcinosarcoma. We conclude that intra-CSF administration of some anticancer agents such as diaziquone and mitomycin C at doses that do not produce toxicity can significantly increase the survival of rats with experimental meningeal carcinomatosis.

Inhibition of protein tyrosine phosphatase by the antitumor agent gallium nitrate.

Protein tyrosine phosphatases (PTPases) play an important role in regulating cell growth and transformation. We report that the antitumor agent gallium nitrate is a potent inhibitor (concentration producing 50% inhibition, 2-6 microM) of detergent-solubilized cellular membrane PTPase from Jurkat human T-cell leukemia cells and HT-29 human colon cancer cells. This is the first report of a selective, small molecule drug inhibitor of PTPase. Gallium nitrate did not inhibit CD45, a PTPase found in the membranes of hemopoietic lineage cells such as Jurkat cells. Studies with gallium nitrate and a series of gallium-containing analogues revealed no correlation between growth-inhibitory activity in Jurkat and HT-29 cells and the ability to inhibit detergent-solubilized PTPase. Gallium nitrate and most of the gallium analogues penetrate poorly into cells. In contrast, a gallium-hydrogen peroxide complex inhibits DNA synthesis in Jurkat cells and induces the accumulation of phosphotyrosines on multiple intracellular proteins in this cell line. Gallium-hydrogen peroxide complex and gallium nitrate have similar inhibitory activity toward detergent-soluble PTPase. This is a new mechanism of action for gallium nitrate but it is not known if the inhibition of PTPase is related to the antitumor activity of gallium nitrate.

Disposition and metabolism of pentamethylmelamine and hexamethylmelamine in rabbits and humans.

The disposition and metabolism of pentamethylmelamine (PMM) and hexamethylmelamine (HMM) were studied in the rabbit, and the disposition of PMM was studied in humans. Parent compound and metabolites were identified by thin-layer chromatography, gas chromatography, and gas chromatography/mass spectrometry analyses. Plasma elimination in both species following i.v. administration of each drug was best described by a two-compartment open model. Both compounds were extensively demethylated with less than 1% of the total dose administered recovered in the urine over 24 hr. The areas under the plasma time-concentration curves of PMM and HMM following p.o. administration to rabbits were 5 and 25% of the areas following i.v. administration. Gastrointestinal absorption was rapid and efficient with 75 to 89% of drug equivalents recoverable in the urine after p.o. administration of [ring-14C]PMM or [ring-14C]HMM to rabbits. Reduced bioavailability of PMM and HMM p.o. appears to be a consequence of rapid metabolism presumably in the liver.

Thioredoxin, a gene found overexpressed in human cancer, inhibits apoptosis in vitro and in vivo.

The redox protein thioredoxin plays an important role in controlling cancer cell growth through regulation of DNA synthesis and transcription factor activity. Thioredoxin is overexpressed by a number of human primary cancers and its expression is decreased during dexamethasone-induced apoptosis of mouse WEHI7.2 thymoma cells. We examined the ability of WEHI7.2 cells stably transfected with human thioredoxin cDNA showing increased levels of cytoplasmic thioredoxin to undergo apoptosis in vitro and in vivo. The cells were protected from apoptosis induced by dexamethasone, staurosporine, etoposide, and thapsigargin, but not by N-acetyl-sphingosine. When inoculated into severe combined immunodeficient mice, the trx-transfected cells formed tumors that showed increased growth compared to wild-type, as well as bcl-2-transfected, WEHI7.2 cells. The trx- and bcl-2-transfected cell tumors both showed less spontaneous apoptosis than tumors formed by the wild-type cells. Unlike tumors formed by the wild-type and bcl-2-transfected WEHI7.2 cells, trx-transfected cell tumors did not show growth inhibition upon treatment with dexamethasone. This study suggests that increased thioredoxin expression in human cancers may result in an increased tumor growth through inhibition of spontaneous apoptosis and a decrease in the sensitivity of the tumor to drug-induced apoptosis.

Protection against adriamycin-induced skin necrosis in the rat by dimethyl sulfoxide and alpha-tocopherol.

Extravasation of Adriamycin during i.v. infusion can cause serious local complications. We have used a rat skin model to study the protection afforded by dimethyl sulfoxide and alpha-tocopherol (vitamin E) against Adriamycin-induced skin necrosis. Topical daily application of 1 ml dimethyl sulfoxide for 2 days produced a small decrease in ulcer diameter of up to 11% at 2 weeks. Topical daily applications of 1 ml 10% alpha-tocopherol succinate in dimethyl sulfoxide for 2 days produced a marked decrease in ulcer diameter at 2 weeks of up to 68%. Daily topical application of 1 ml 10% alpha-tocopherol succinate in dimethyl sulfoxide for 7 days offered no greater protection than 2-day application. alpha-Tocopherol acetate appeared to have activity slightly less than that of alpha-tocopherol succinate in reducing ulcer size, and both compounds were considerably more active than was alpha-tocopherol alcohol. Administration of alpha-tocopherol succinate or alpha-tocopherol acetate i.p. had no significant effect upon ulcer diameter. Topically applied dimethyl sulfoxide and alpha-tocopherol may provide an effective way of treating accidentally extravasated Adriamycin in cancer patients.

Anthracycline-induced inhibition of a calcium action potential in differentiated murine neuroblastoma cells.

The effects of some anthracyclines on a Ca2+ -dependent action potential have been studied in differentiated murine neuroblastoma cells (N1E-115 clone). The differentiated neuroblastoma cell possesses characteristics of an electrically excitable cell and can generate propagated potential spikes in which Ca2+ is the inward charge carrier. This was shown by the fact that action potentials recorded from differentiated neuroblastoma cells in the presence of 10(-7) g of tetrodotoxin per ml, which inhibits active Na+ channels, had a spike amplitude that depended upon the extracellular Ca2+ concentration in a manner close to that predicted by the Nernst equation. The peak potential changed 28.9 mV/decade change in extracellular Ca2+. Local application to a cell of 10(-8) M doxorubicin produced inhibition of this Ca2+ -dependent action potential within 5 s of drug application and a maximum inhibition of 13% 60 s after drug application. There was almost complete recovery to the initial spike amplitude value within 10 min after removing drug. The same concentration of doxorubicin also produced complete inhibition, without recovery, of a Ca2+ -dependent after-discharge which followed the initial action potential in about half the cells studied. Increasing concentrations of doxorubicin produced dose-dependent inhibition of the initial Ca2+ -dependent action potential. Cells exposed to 10(-5) M doxorubicin showed 88% inhibition of the Ca2+ -dependent action potential with no recovery even 10 min after removing the drug. Daunomycin, 10(-6) M, produced 90% inhibition of the Ca2+ -dependent action potential. Daunomycin aglycone (10(-6) M), which lacks antitumor activity, had no significant effect on the Ca2+ -dependent action potential. The rapid onset of the drug-induced response together with the low concentrations of anthracyclines needed to inhibit voltage-dependent Ca2+ channels in the neuroblastoma cells suggest a direct effect of anthracyclines on the cell surface membrane. The findings are discussed in light of the possible role of Ca2+ in cancer cells.

Topical chemotherapy of intradermal Walker 256 carcinosarcoma with diaziquone and doxorubicin in the rat.

A model for metastatic skin cancer using intradermal injection of Walker 256 carcinosarcoma has been developed in the rat. Using this model, antitumor activity of topically applied doxorubicin and diaziquone in Vanicream, Plastibase, and dimethyl sulfoxide (DMSO) as vehicles was compared with intraperitoneal injection of the drugs at the same doses beginning 4 days after injection of tumor cells. Doxorubicin applied topically at 0.5 mg/day for 4 days in Vanicream or Plastibase exhibited no antitumor activity, while i.p. administered doxorubicin at 0.5 mg/day for 4 days inhibited tumor growth at day 20 by 66%. Diaziquone applied topically at 0.1 mg/day for 4 days in Vanicream, Plastibase, or DMSO inhibited tumor growth at day 20 by 66, 86, and 43%, respectively, and cured animals of the skin tumor at a dose of 0.5 mg/day. Diaziquone administered i.p. at 0.5 mg/day for 4 days was lethal to rats, and at 0.1 mg/day it produced 93% inhibition of tumor growth at day 20. Diaziquone applied topically at 0.1 mg/day for 4 days in Plastibase cured rats of advanced tumor when treatment was begun 12 days after injection of tumor cells. The area under the plasma radioactivity time curve over 5 h for a single 0.64-mg dose of topically applied [ring-14C]diaziquone in DMSO was 0.01% that of the same dose of [ring-14C]diaziquone administered i.p. in non-tumored rats. The decrease in WBC count following topical application of diaziquone at a dose of 0.1 mg/day for 4 days, compared to the same dose of diaziquone administered i.p., was 62% in Vanicream, 81% in Plastibase and 33% in DMSO. Topical diaziquone was non-toxic to normal skin in the rat and in the domestic pig. It is concluded that topical application of diaziquone offers a therapeutic advantage over systemic treatment for metastatic cancer of the skin.

Role of metabolism and oxidation-reduction cycling in the cytotoxicity of antitumor quinoneimines and quinonediimines.

Quinone(di)imines are nitrogen analogues of quinones in which one or both quinone oxygens are replaced by an imino group. A series of quinone(di)imines with antitumor activity has been studied for its in vitro chemical reactivity, metabolism, acute toxicity to primary cultured rat hepatocytes, and growth-inhibitory activity with Chinese hamster ovary (CHO) cells. The quinone(di)imines exhibited a wide range of activity as substrates for metabolism by hepatic microsomal flavoenzymes. The maximum rate of quinone(di)imine metabolism was more than 7.5-fold greater than reported for metabolism of quinones. Some quinone(di)imines formed free radicals that could be detected by electron spin resonance spectroscopy. 2-Amino-1,4-naphthoquinoneimine gave a short-lived electron spin resonance signal that could be detected only under aerobic conditions. 2,3',6-Trichloroindophenol gave an electron spin resonance signal in air that was stable for 24 h. Most quinone(di)imines underwent oxidation-reduction cycling to form the superoxide anion radical, but some quinone(di)imines, although rapidly metabolized, formed little or no superoxide anion radical. Quinone(di)imines were relatively toxic to hepatocytes and CHO cells, and some quinone(di)imines were more toxic to one cell type than the other. The log 1-octanol/water partition coefficient showed an optimal value of 2.61 for toxicity against both cell types. In hepatocytes the more toxic quinone(di)imines were the most rapidly metabolized. For a subgroup of quinone(di)imines toxicity to hepatocytes and CHO cells appeared to be related to the ability to form a semiquinone(di)imine free radical. Toxicity of quinone(di)imines to hepatocytes and CHO cells was not related to superoxide anion radical formation, and toxicity to CHO cells was not affected by exclusion of oxygen during exposure of the cells to the compounds. The rate of chemical addition of quinone(di)imines to reduced glutathione did not correlate with toxicity. An understanding of the mechanisms of acute toxicity and growth-inhibitory activity of quinone(di)imines could lead to the design of more selective quinonoid antitumor agents.

Inhibition of the signalling enzyme phosphatidylinositol-3-kinase by antitumor ether lipid analogues.

Phosphatidylinositol-3-kinase (PtdIns-3-kinase) is an enzyme found associated with many growth factor receptor protein tyrosine kinases and oncogene protein tyrosine kinases. PtdIns-3-kinase appears to be important for mitogenesis and the malignant transformation of cells. The antitumor ether lipid analogue, 1-O-octadecyl-2-O-methyl-rac-glycero-3-phosphocholine (ET-18-OCH3), was found to be an inhibitor of Swiss mouse 3T3 fibroblast and bovine brain PtdIns-3-kinases. The concentration of ET-18-OCH3 causing 50% inhibition (IC50) was 35 microM. The inhibition of PtdIns-3-kinase by ET-18-OCH3 was noncompetitive with ATP. Other antitumor ether lipid analogues also inhibited PtdIns-3-kinase. The cyclic ether lipid analogue (+/-)-2-(hydroxy[tetrahydro-2-(octadecyloxy)methylfuran-2- yl]methoxyl)phosphinyloxy,N,N,N-trimethyethaniminium hydroxide inhibited with an IC50 of 42 microM and hexadexylphosphocholine with an IC50 of 48 microM. 1-O-Octadecyl-2-O-methyl-rac-3-glycerophospho-myo- inositol was a weaker inhibitor of PtdIns-3-kinase, with an IC50 of 96 microM and was itself phosphorylated by the enzyme. Lipid extracted from cells grown with ET-18-OCH3 for 18 h showed inhibition of PtdIns-3-kinase with endogenous PtdIns as substrate, with an ET-18-OCH3 IC50 of 18 microM. ET-18-OCH3 inhibited platelet-derived growth factor-stimulated phosphatidylinositol-3-phosphate formation by v-sis NIH 3T3 cells with an IC50 of 12.5 microM. The results of the study suggest that inhibition of PtdIns-3-kinase might contribute to the antiproliferative activity of the antitumor ether lipid analogues.

Selective inhibition of phosphatidylinositol phospholipase C by cytotoxic ether lipid analogues.

The ether lipid analogue 1-octadecyl-2-methyl-rac-glycero-3-phosphocholine (ET-18-OCH3) has been shown to be a direct inhibitor of Swiss 3T3 fibroblast and BG1 ovarian adenocarcinoma cell cytosolic phosphoinositide selective phospholipase C (PIPLC) using [3H]-phosphatidylinositol-(4, 5)-bisphosphate ([3H]PIP2) as the substrate. The inhibition occurred when ET-18-OCH3 was incorporated into the [3H]PIP2 substrate micelles, with 50% inhibition (IC50) occurring at a ET-18-OCH3: [3H]PIP2 ratio of 0.04, or an assay concentration of 0.4 microM, and when ET-18-OCH3 was added directly to the incubation, with an IC50 of 9.6 microM. Lipid prepared from cells exposed to cytotoxic concentrations of ET-18-OCH3 for 18 h also inhibited PIPLC with an IC50 less than 1 microM. The noncytotoxic analogue 1-O-alkyl-2-hydroxy-sn-glycero-3-phosphocholine inhibited PIPLC when incorporated into the [3H]PIP2 substrate micelles, but lipid from cells grown with 5 microM 1-O-alkyl-2-hydroxy-sn-glycero-3-phosphocholine did not inhibit PIPLC. BG1 cells, which were more sensitive than Swiss 3T3 fibroblasts to growth inhibition by ET-18-OCH3, had a cytosolic PIPLC activity one-third that of Swiss 3T3 cells. NIH 3T3 cells exhibited the same sensitivity to growth inhibition by ET-18-OCH3 as Swiss 3T3 cells and had a similar level of PIPLC. v-sis NIH 3T3 cells were relatively resistant (greater than 3-fold) to growth inhibition by ET-18-OCH3 and had a cytosolic PIPLC activity more than twice that of the wild type cells. ET-18-OCH3 was a weak inhibitor, IC50 greater than 100 microM, of phospholipase D activity in NIH 3T3 cell membranes. In intact NIH 3T3 cells ET-18-OCH3 at cytotoxic concentrations did not inhibit phospholipase D or phosphatidylcholine-selective phospholipase C activity. The results show that the ether lipid analogues at cytotoxic concentrations are selective inhibitors of PIPLC and that the inhibition of PIPLC may be related to the growth inhibitory activity of the ether lipid analogues.

Activation and inactivation of cancer chemotherapeutic agents by rat hepatocytes cocultured with human tumor cell lines.

While colony formation assays provide sensitive indices of tumor cell proliferation and growth inhibition imposed by many chemotherapeutic agents, drugs which require metabolic activation lack activity in such assays. In the present study, we have utilized freshly isolated rat hepatocytes for the activation of drugs which are metabolized by hepatic microsomal as well as extra-microsomal enzymes. Hepatocytes in fluid medium are placed over soft-agarose matrix containing tumor-derived cells (e.g., A204, A549) within 35-mm culture dishes; drug and/or drug vehicle is added directly to the hepatocyte layer, and cultures are incubated for 24 hr prior to removal of the hepatocyte layer. Tumor cell colony formation is assessed following 7 to 10 days of incubation. Cyclophosphamide was used as a prototype agent to assess utility of the coculture methodology. In vivo treatment of rats with phenobarbital prior to hepatocyte isolation enhances cyclophosphamide toxicity in vitro, whereas pretreatment with carbon tetrachloride markedly reduced subsequent in vitro cyclophosphamide cytotoxicity. Hepatocyte:tumor cell cocultures provide an efficient means to detect metabolic activation and inactivation of several selected cancer chemotherapeutic agents as well. In the presence of hepatocytes, the 50% growth-inhibitory concentrations for cyclophosphamide, indicine N-oxide, and procarbazine are markedly decreased, whereas the 50% growth-inhibitory concentrations for [2,5-bis(1-aziridinyl)-3,6-diazo-1,4-cyclohexadiene-1,4-diyl]bis(c arbamic acid)diethyl ester, 1,3-bis-chloro(2-chloroethyl)-1-nitrosourea, dacarbazine, 5-fluorouracil, ftorafur, 1-(2-chloroethyl)-3-(4-methylcyclohexyl)-1-nitrosourea, and vincristine are significantly increased. By contrast, the 50% growth-inhibitory concentrations for actinomycin D, mitomycin C, 6-mercaptopurine, and other agents are unaffected by hepatocyte presence. Cryopreserved hepatocytes exhibit detectable levels of drug activation, although inadequate for routine use. Results suggest that hepatocyte:tumor cell cocultures may be well-suited for assessing the degree to which hepatic metabolism may activate or inactivate new anticancer drugs.