The core control system of intracellular iron homeostasis: a mathematical model.

Iron is a metal essential for cellular metabolism. However, excess iron available for reactions contributes to the formation of dangerous reactive oxygen species, such as the hydroxyl radical, via the Fenton reaction. Therefore, intracellular iron levels are tightly constrained by a control system of proteins. This paper contains a mathematical model, in the form of a system of five ordinary differential equations, of the core of this control system, including the labile iron pool as well as proteins that regulate uptake, storage, and export and are connected through negative feedback loops. The model is validated using data from an overexpression experiment with cultured human breast epithelial cells. The parameters in the mathematical model are not known for this particular cell culture system, so the analysis of the model was done for a generic choice of parameters. Through a mixture of analytical arguments and extensive simulations it is shown that for any choice of parameters the model reaches a unique stable steady state, thereby ruling out oscillatory behavior. It is shown furthermore that the model parameters are identifiable through suitable experiments.

Diverse genotypical features and impacts on clinical course and severity of cystic fibrosis: early childhood experience.

AIM: Very little is known about the relationship between genotype and phenotype of cystic fibrosis (CF) from the Turkish children. The aim of the study was to analyze the genotype and phenotype of 24 children with CF and to investigate the correlation between type of mutation in cystic fibrosis transmembrane conductance regulator (CFTR) protein gene and clinical manifestation of the disease. METHODS: Patients were evaluated retrospectively and prospectively. History, clinical findings, sweat test and mutation analysis were used for the definitive diagnosis of CF. Phenotypical features of 24 cases were evaluated according to clinical findings. We compared the clinical phenotype and age at diagnosis, genotypic features. A total of 36 mutations were analyzed by polymerase chain reaction (PCR) and reverse hybridization methods. Statistical analysis was done by using χ2, Fisher exact and Pearson correlation tests. RESULTS: The mean age of the cases that were admitted to our out-patient clinic was 5.3±4 years. The median age of diagnosis was three months. Parents were consanguineous in 37.5% of cases and loss of a sibling at one year of age was stated in a quarter. The most frequent symptom was recurrent diarrhoea (79.2%) and there was severe growth retardation in 12 (50%) and pseudo-Bartter (PB) syndrome in 11 of the cases. The incidence of PB was higher in cases that were diagnosed at one year of age. Out of 18 cases with mutation analysis, nine (50%) were positive for DF508 mutation, and four cases were homozygous out of nine cases. Two separate mutations were determined in two cases with severe clinical picture. The incidence of respiratory tract infection during the admission was lower in DF508 positive cases (P=0.016). There was no statistically significant relation between DF508 positivity and diarrhea, severe growth retardation and PB (P>0.05). The other mutations that were determined in our patients were rarely seen mutations such as 3120+1 G-A, R347P, 1677delTA, 2789+5G-A, 2183AA-G, and R1066C. CONCLUSION: DF508 mutation rates in our cases diagnosed in early childhood were higher than the rates reported previously in Turkish children. The definition of molecular defect in CFTR gene has an impact on verifying the diagnosis and decreasing morbidity and mortality. An adequately large sample size is needed to evaluate the mutation profiles and genotype-phenotype characteristics in our country.

A new assay to quantify in vivo repair of G:T mispairs by base excision repair.

The double mismatch reversion (DMR) assay quantifies the repair of G:T mispairs exclusively by base excision repair in vivo. Synthetic oligonucleotides containing two G:T mispairs on opposite strands were placed into the suppressor tRNA gene supF in the shuttle plasmid pDMR. Placement of two mispairs on opposite strands of supF creates a one to one correspondence between the number of correct repair events prior to replication in which G:T mispairs are converted to G:C base pairs and the number of post-replication progeny plasmids with functional supF. Replication of unrepaired or incorrectly repaired mispairs cannot produce progeny plasmids containing functional supF. Indeed, direct transformation of Escherichia coli strain MBL50, which reports the functional status of supF, with pDMR constructs containing two G:T or G:G mispairs yielded <0.5% wild-type supF-containing colonies. In contrast, passage of G:T mispair-containing pDMR constructs through human 5637 bladder carcinoma cells for 48h prior to plasmid recovery and transformation of the reporter E. coli strain MBL50 produced 47% wild-type supF-containing colonies. This finding was indicative of repair prior to the onset of replication in 5637 cells. However, passage of G:G mispair-containing pDMR constructs through 5637 cells yielded <0.5% wild-type supF-containing colonies. Moreover, no difference was observed in the rate of G:T mispair repair by HCT 116 colorectal carcinoma cells deficient in long-patch mismatch repair and a long-patch mismatch repair proficient HCT 116 subline. These data demonstrate that repair measured by the DMR assay is exclusively attributable to short-patch pathways. The DMR assay proved useful in the analysis of the effect of the base 5' to a mispaired G on the rate of G:T base excision repair by 5637 cells, indicating the sequence preference CpG approximately 5mCpG>TpG>GpG approximately ApG, and in the comparison of G:T base excision repair rates between cell lines.

Comparison of tropisetron and granisetron in the control of nausea and vomiting in children receiving combined cancer chemotherapy.

Tropisetron and granisetron are selective serotonin (5-HT3) antagonists that have been proven effective in the prevention of nausea and vomiting in adults and children receiving cancer chemotherapy. This prospective, randomised study was designed to compare the efficacy of the two agents in the prevention of vomiting and nausea in children receiving highly emetogenic chemotherapy for various malignancies. A total of 51 children (mean age: 7.7 +/- 4.8 year) were studied in 133 chemotherapy cycles. In 66 chemotherapy cycles, the children received tropisetron as an antiemetic agent in a dose of 0.2 mg/kg/24 h intravenously and, in 67 cycles, they received granisetron 40 micrograms/kg/24 h intravenously before cytotoxic drug administration during the days they received chemotherapy. The response per 24 h of chemotherapy was defined as complete (no nausea and vomiting), partial (1-4 events of vomiting and/or nausea), and failure (more than 4 events of vomiting and/or nausea). Efficacy of antiemetic therapy was evaluated as acute (Day 1) and overall was based on the worst day during the chemotherapy. Complete control of acute vomiting was achieved in 74% of tropisetron and 88% of granisetron patients (P = 0.04), and complete control of acute nausea in 56% and 82% respectively (p = 0.002). Overall response by means of complete control of both vomiting and nausea during the whole therapy period was 29% of tropisetron group and 55% of granisetron group (p = 0.007). The statistical analysis (depending on the emetogenicity of the chemotherapy cycles) showed increased efficacy of granisetron in highly (grade 3) emetogenic chemotherapy cycles (p = 0.002), whereas there was no difference in the very highly emetogenic cycles (p = 0.7). Also, granisetron was found to be more effective than tropisetron, especially in patients heavier than 25 kg (p = 0.02). The adverse reactions were few and mild. There were no differences in the tolerability of the two antiemetic therapy modalities. In conclusion, granisetron was found to be more effective than tropisetron in controlling nausea and vomiting in children receiving highly emetogenic chemotherapy. This increased antiemetic efficacy of ganisetron might have been related to maximal dose differences according to body weight.

Oxidative DNA base modifications in peripheral blood mononuclear cells of patients treated with high-dose infusional doxorubicin.

In prior studies, it was demonstrated that the redox metabolism of doxorubicin leads to the formation of promutagenic oxidized DNA bases in human chromatin, suggesting a potential mechanism for doxorubicin-related second malignancies. To determine whether a similar type of DNA damage is produced in the clinic, peripheral blood mononuclear cell DNA from 15 women treated with infusional doxorubicin (165 mg/m(2)) as a single agent was examined for 14 modified bases by gas chromatography/mass spectrometry with selected ion monitoring. Prior to the 96-hour doxorubicin infusion, 13 different oxidized bases were present in all DNA samples examined. Chemotherapy, producing a steady-state level of 0.1 microM doxorubicin, increased DNA base oxidation up to 4-fold compared to baseline values for 9 of the 13 bases studied. Maximal base oxidation was observed 72 to 96 hours after doxorubicin treatment was begun; the greatest significant increases were found for Thy Gly (4.2-fold), 5-OH-Hyd (2.5-fold), FapyAde (2.4-fold), and 5-OH-MeUra (2.4-fold). The level of the promutagenic base FapyGua increased 1.6-fold (P < .02), whereas no change in 8-OH-Gua levels was observed in peripheral blood mononuclear cell DNA during the doxorubicin infusion. These results suggest that DNA base damage similar to that produced by ionizing radiation occurs under clinical conditions in hematopoietic cells after doxorubicin exposure. If doxorubicin-induced DNA base oxidation occurs in primitive hematopoietic precursors, these lesions could contribute to the mutagenic or toxic effects of the anthracyclines on the bone marrow.

Mapping oxidative DNA damage using ligation-mediated polymerase chain reaction technology.

Reactive oxygen species induce a pharmacopoeia of oxidized bases in DNA. DNA can be cleaved at most of the sites of these modified bases by digestion with a combination of two base excision repair glycosylases from Escherichia coli, Fpg glycosylase, and endonuclease III. The frequency of the resulting glycosylase-dependent 5'-phosphoryl ends can be mapped at nucleotide resolution along a sequencing gel autoradiogram by a genomic sequencing technique, ligation-mediated polymerase chain reaction (LMPCR). In cultured rat cells, the frequency of endogenous oxidized bases in mitochondrial DNA is sufficiently high, about one oxidized base per 100 kb, to be directly mapped from 0.1 microg of total cellular DNA preparations by LMPCR. Nuclear DNA has a lower frequency of endogenous oxidative base damage which cannot be mapped from 1-microg preparations of total cellular DNA. Preparative gel electrophoresis of the PGK1 and p53 genes from 300 microg of restriction endonuclease-digested genomic DNA showed a 25-fold enrichment for the genes and, after endonuclease digestion followed by LMPCR, gave sufficient signal to map the frequency of oxidized bases from human cells treated with 50 microM H2O2.

Mapping oxidative DNA damage and mechanisms of repair.

We developed a method to map oxidative-induced DNA damage at the nucleotide level using ligation-mediated polymerase chain reaction (LMPCR) technology. In vivo and in vitro DNA base modification patterns inflicted by reactive oxygen species (ROS) in the human P53 and PGK1 gene were nearly identical in vitro and in vivo. In human male fibroblasts, these patterns are independent of the transition metal used (Cu (II), Fe(II), or Cr(VI). Therefore, local probability of H2O2-mediated DNA base damage is determined primarily by DNA sequence. Moreover, in cells undergoing severe oxidative stress, extranuclear sites contribute metals that enhance nuclear DNA damage. The role of the base excision repair pathway in human cells responsible for the repair of the majority of ROS base damage is also discussed.

Mapping of peroxyl radical induced damage on genomic DNA.

We have examined the DNA damage produced by reaction of peroxyl radicals with human fibroblast DNA. DNA damage consisted of both strand breaks and base modifications. The extent of strand breaks and base modifications induced as a function of peroxyl radical concentration was determined by quantitation of fragment size distributions using denaturing glyoxal-agarose gel electrophoresis. Both strand breaks and base modifications increased in a log linear fashion with respect to peroxyl radical concentration. Oxidative base modifications were observed to occur to a greater extent than strand breaks at every concentration measured. The sequence-specific distribution of peroxyl radical induced base damage was mapped for 803 nucleotide positions using the method of ligation mediated PCR. A total of 87% of all guanine positions in the examined sequences was found to be significantly oxidized. The order of reactivity of DNA bases toward oxidation by peroxyl radicals was found to be G >> C > T. Adenine is essentially unreactive. The yield of oxidative base modifications at guanines and cytosines by peroxyl radicals depends on the exact specification of 5' and 3' flanking bases in a polarity dependent manner. Every guanine in the 5'XGC3' motif was found to be oxidized, where X is any 5' neighbor. In contrast, 5' and 3' purine flanks drastically reduced the extent of peroxyl radical G oxidation. The pattern of base modification and the influence of nearest neighbors differs substantially from that previously reported for hydrogen peroxide damage mediated by low valent transition metal ions for the identical DNA sequences.

Large scale isolation of genes as DNA fragment lengths by continuous elution electrophoresis through an agarose matrix.

A simple procedure for the isolation of genes as DNA fragment lengths is described. By using a commercial continuous elution protein electrophoresis apparatus and incorporating an agarose matrix, preparative scale amounts (300 microg) of DNA can be purified by fragment lengths from a mixture of genomic fragment lengths with high recovery yields. Fractions corresponding to unique fragment length ranges are screened for individual genes by dot-blot analysis. Using this technique, we have isolated two genes: PGK1, a single copy housekeeping gene; and p53 (exons 3-11), a tumor suppressor gene, whose DNA fragment lengths elute at 4 and 7.5 kbp, respectively, from a single preparative run. As an example of the utility of the technique, we applied it to improving the sensitivity of the ligation-mediated polymerase chain reaction (LMPCR)--a nucleic acid amplification technique used for the detection and mapping of DNA damage along genes. By eliminating excess nontargeted genomic DNA, the agarose matrix continuous elution electrophoresis (CEE) procedure provided a 24-fold increase in signal strength attributable to base damage caused by exposing DNA to reactive oxygen species. Genomic DNA fragment length purification by agarose matrix CEE should also prove useful in other research areas requiring gene isolation, such as genomics and molecular biology.

Mapping oxidative DNA damage at nucleotide level.

DNA damage induced by reactive oxygen species (ROS) is considered an important intermediate in the pathogenesis of human conditions such as cancer and aging. By developing an oxidative-induced DNA damage mapping version of the Ligation-mediated polymerase chain reaction (LMPCR) technique, we investigated the il vivo and in vitro frequencies of DNA base modifications caused by ROS in the human p53 and PGK1 gene. Intact human male fibroblasts were exposed to 50mM H2O2, or purified genomic DNA was treated with 5 mM H2O2, 100 microM Ascorbate, and 50 microM, 100 microM, or 100 microM of Cu(II), Fe(II), or Cr(VI) respectively. The damage pattern generated in vivo was nearly identical to the in vitro Cu(II) or Fe(III) damage patterns; damage was non-random with guanine bases heavily damaged. Cr(VI) generated an in vitro damage pattern similar to the other metal ions, although several unique thymine positions were damaged. Also, extra nuclear sites are a major contributor of metal ions (or metal-like ligands). These data show that the local probability of H2O2-mediated DNA damage is determined by the primary DNA sequence, with chromatin structure having a limited effect. The data suggest a model in which DNA-metal ion binding domains can accommodate different metalions. LMPCR's unique aspect is a blunt-end ligation of an asymmetric double-stranded linker, permitting exponential PCR amplification. An important factor limiting the sensitivity of LMPCR is the representation of target gene DNA relative to non-targeted genes; therefore, we recently developed a method to eliminate excess non-targeted genomic DNA. Restriction enzyme-digested genomic DNA is size fractionated by Continuous Elution Electrophoresis (CEE), capturing the target sequence of interest. The amount of target DNA in the starting material for LMPCR is enriched, resulting in a stronger amplification signal. CEE provided a 24-fold increase in the signal strength attributable to strand breaks plus modified bases created by ROS in the human p53 and PGK1 genes, detected by LMPCR. We are currently taking advantage of the enhanced sensitivity of target gene-enriched LMPCR to map DNA damage induced in human breast epithelial cells exposed to non-cytotoxic concentrations of H2O2.

The identification and characterization of a G4-DNA resolvase activity.

There is increasing evidence that four-stranded Hoogsteen-bonded DNA structures, G4-DNA, play an important role in cellular processes such as meiosis and recombination. The Hoogsteen-bonded G4-DNA is thermodynamically more stable than duplex DNA, and many guanine-rich genomic DNA sequences with the ability to form G4-DNA have been identified. A protein-dependent activity that resolves G4-DNA into single-stranded DNA has been identified in human placental tissue. The resolvase activity was purified from any apparent nuclease activity and is dependent on NTP hydrolysis and MgCl2. Resolvase activity is optimal with 5 mM MgCl2. The Vmax/Km of ATP is 0. 055%/min/microM, higher than the Vmax/Km of the other dNTPs. The products of the resolvase reaction are unmodified single-stranded DNA. The resolvase is not a duplex DNA helicase or a topoisomerase II activity and does not unwind Hoogsteen-bonded triplex DNA. Resolvase is a novel activity that unwinds stable G4-DNA structures using a dNTP-dependent mechanism producing unmodified single-stranded DNA. Potential in vivo roles for this G4-DNA resolvase activity are discussed.

Metal ion-dependent hydrogen peroxide-induced DNA damage is more sequence specific than metal specific.

The frequency of oxidative base damage along the human p53 and PGK1 genes was determined at nucleotide resolution by cleaving DNA at oxidized bases with endonuclease III and formamidopyrimidine DNA glycosylase and then using the ligation-mediated PCR technique to map induced break frequency. Damage was induced either in vivo by exposing cultured human male fibroblasts to H2O2 or in vitro by exposing purified genomic DNA to H2O2 plus ascorbate in the presence of Cu(II), Fe(III), or Cr(VI) metal ions. All four base damage patterns from either in vivo or in vitro treatments were nearly identical in both regions of the genome. The frequency of base damage varied along the DNA, with guanine being the most commonly damaged base. In the Fe(III)-mediated in vitro reactions, single-stranded breaks were almost completely suppressed by addition of sucrose, which facilitated mapping of base damage. The in vitro base damage pattern generated by Cr(VI), ascorbate, and H2O2 was similar to that of the other metal ions, with the exception of several unique positions; these were heavily damaged only in the presence of Cr(VI). Isolated nuclei suffered little oxidative base damage in the presence of ascorbate and H2O2, and we conclude that during H2O2 in vivo treatment of cells, metal ions (or metal-like ligands) are freed from the cytoplasm to migrate into the nucleus and supply the redox cycling ligands necessary for oxidative base damage. These data simplify the complexity of H2O2-induced oxidative damage and mutagenesis studies by demonstrating the commonality of damage catalyzed by different transition metal ions and by showing that the pattern of H2O2-mediated oxidative base damage is determined almost entirely by the primary DNA sequence, with chromatin structure having a limited effect. Our data suggest a model for base damage in which DNA-metal ion binding domains can equally accommodate a variety of different metal ions and thus are a key factor in determining the local probability of DNA damage.

A hot spot for hydrogen peroxide-induced damage in the human hypoxia-inducible factor 1 binding site of the PGK 1 gene.

Using ligation-mediated polymerase chain reaction to separately map the distribution of induced oxidized bases and strand breaks along the human PGK1 promoter at nucleotide resolution, we previously described the pattern of oxidative DNA damage induced in vitro by Cu(II)/ascorbate/H2O2 [J. Biol. Chem. 270, 17633-17640 (1995)]. Here we report that the pattern of in vivo base damage caused by H2O2 is almost identical to that of the previously used in vitro system with the exception of transcription factor-associated footprints. An unusually strong positive footprint for both strand breaks and oxidized bases is associated with binding of the hypoxia-inducible transcription factor-1. Base damage at this footprint was 52-91% repaired in 24 h, which was similar to the global base damage repair rate. However, strand breaks at this footprint were only 39-55% repaired in 24 h or approximately 100-fold slower than the global strand break repair rate.

Pharmacokinetics and toxicity of 120-hour continuous-infusion hydroxyurea in patients with advanced solid tumors.

A group of 18 patients with advanced cancer were entered on a phase I study of a 120-h continuous intravenous infusion of hydroxyurea. The dose of hydroxyurea was escalated in cohorts of patients from 1 to 2 to 3.2 g/ m2 per day. The primary dose-limiting toxicity was neutropenia, often accompanied by leukopenia, thrombocytopenia and generalized skin rash. Prophylactic treatment of patients with dexamethasone and diphenhydramine hydrochloride prevented the skin rash, but not the hematopoietic toxicities. The pharmacokinetics of hydroxyurea were studied in all patients. The steady-state concentrations of hydroxyurea were linearly correlated with the dose (R2 = 0.71, n = 18, P < 0.0001). The mean +/- SE concentrations were 93 +/- 16, 230 +/- 6 and 302 +/- 27 microM at 1, 2 and 3.2 g/m2 per day, respectively. The mean +/- SE renal and nonrenal clearances of hydroxyurea were 2.14 +/- 0.18 and 3.39 +/- 0.28 l/h per m2 (n = 16), neither of which correlated with the dose. The concentration of hydroxyurea in plasma decayed monoexponentially with a mean +/- SE half-life of 3.25 +/- 0.18 h (n = 17). The steady-state concentration of hydroxyurea was > 200 microM in all nine patients treated at 2 g/m2 per day, a dose which was well tolerated for 5 days. We recommend this dose for phase II trials in combination with other antineoplastic agents.

In vivo mutagenesis of the reporter plasmid pSP189 induced by exposure of host Ad293 cells to activated polymorphonuclear leukocytes.

We measured the mutation frequency and spectrum induced by exposure of the mutation reporter plasmid pSP189 in vivo to phorbol myristate acetate (PMA)-activated human polymorphonuclear leukocytes (PMNs). The mutation frequency induced in the supF tRNA gene of pSP189 transfected into human Ad293 cells by a 30 min exposure to 4 x 10(6) activated PMNs/ml was 3- to 9-fold higher than the background mutation frequency of 0.1-1.8 x 10(-5). The enhanced mutation frequency caused by activated PMNs required replication of the reporter plasmid in host Ad293 cells. Fifty five unique activated PMN-associated mutants characterized by sequencing included base substitutions (55%) and deletions (45%), however, no small (1-3 bp) deletions were observed. Ninety four percent of point mutations occurred at C:G base pairs, with C:G-->T:A transitions (47%) and C:G-->A:T transversions (37%) predominating. A prominent hot-spot was observed at d(pCAGAC) on the tRNA strand. Although H202 generation was required for mutagenesis, the mutation spectrum induced in pSP189 by in vivo exposure to activated PMNs differed from that induced by in vivo exposure to H202. It also differed from the spectrum induced in single-stranded DNA in vitro by activated PMNs, suggesting that the mutational spectrum is a complex function of the kinetics of reactive oxygen generation and factors contributed by the target cell.

Cupric ion/ascorbate/hydrogen peroxide-induced DNA damage: DNA-bound copper ion primarily induces base modifications.

The kinetics of frank DNA strand breaks and DNA base modifications produced by Cu(II)/ascorbate/H2O2 were simultaneously determined in purified human genomic DNA in vitro. Modified bases were determined by cleavage with Escherichia coli enzymes Nth protein (modified pyrimidines) and Fpg protein (modified purines). Single-stranded lesion frequency before (frank strand breaks) and after (modified bases) Nth or Fpg protein digestion was quantified by neutral glyoxal gel electrophoresis. Dialysis of EDTA-treated genomic DNA purified by standard proteinase K digestion/phenol extraction was necessary to remove low molecular weight species, probably transition metal ions and metal ion chelators, which supported frank strand breaks in the presence of ascorbate + H2O2 without supplemental copper ions. We then established a kinetic model of the DNA-damaging reactions caused by Cu(II) + ascorbate + H2O2. The principal new assumption in our model was that DNA base modifications were caused exclusively by DNA-bound Cu(I) and frank strand breaks by non-DNA-bound Cu(I). The model was simulated by computer using published rate constants. The computer simulation quantitatively predicted: (1) the rate of H2O2 degradation, which was measured using an H2O2-sensitive electrode, (2) the linearity of accumulation of DNA strand breaks and modified bases over the reaction period, (3) the rate of modified base accumulation, and (4) the dependence of modified base and frank strand production on initial Cu(II) concentration. The simulation significantly overestimated the rate of frank strand break accumulation, suggesting either that the ultimate oxidizing species that attacks the sugar-phosphate backbone is a less-reactive species than the hydroxyl radical used in the model and/or an unidentified hydroxyl radical-scavenging species was present in the reactions. Our experimental data are consistent with a model of copper ion-DNA interaction in which DNA-bound Cu(I) primarily mediates DNA base modifications and nonbound Cu(I) primarily mediates frank strand break production.

Formation of DNA-protein cross-links in cultured mammalian cells upon treatment with iron ions.

Formation of DNA-protein crosslinks (DPCs) in mammalian cells upon treatment with iron or copper ions was investigated. Cultured murine hybridoma cells were treated with Fe(II) or Cu(II) ions by addition to the culture medium at various concentrations. Subsequently, chromatin samples were isolated from treated and control cells. Analyses of chromatin samples by gas chromatography/mass spectrometry after hydrolysis and derivatization revealed a significant increase over the background amount of 3-[(1,3-dihydrio-2,4-dioxopyrimidin-5-yl)-methyl]- L-tyrosine (Thy-Tyr crosslink) in cells treated with Fe(II) ions in the concentration range of 0.01 to 1 mM. In contrast, Cu(II) ions at the same concentrations did not produce this DPC in cells. No DNA base damage was observed in cells treated with Cu(II) ions, either. Preincubation of cells with ascorbic acid or coincubation with dimethyl sulfoxide did not significantly alleviate the Fe(II) ion-mediated formation of DPCs. In addition, a modified fluorometric analysis of DNA unwinding assay was used to detect DPCs formed in cells. Fe(II) ions caused significant formation of DPCs, but Cu(II) ions did not. The nature of the Fe(II)-mediated DPCs suggests the involvement of the hydroxyl radical in their formation. The Thy-Tyr crosslink may contribute to pathological processes associated with free radical reactions.

Mapping of copper/hydrogen peroxide-induced DNA damage at nucleotide resolution in human genomic DNA by ligation-mediated polymerase chain reaction.

The ligation-mediated polymerase chain reaction was used to map the frequency of reactive oxygen species-induced DNA damage at nucleotide resolution in genomic DNA purified from cultured human male fibroblasts. Damaged pyrimidine and purine bases were recognized and cleaved by the Nth and Fpg proteins from Escherichia coli, respectively. Strand breaks and modified bases were induced in vitro by copper ion-mediated reduction of hydrogen peroxide in the presence of ascorbate; reactant concentrations were adjusted to induce lesions at a frequency of 1 per 2-3 kilobases in purified genomic DNA. Glyoxal gel analysis demonstrated that the ratio of induced strand breaks to induced base damage was 0.8/2.7 in DNA dialyzed extensively to remove adventitious transition metal ions. Ligation-mediated polymerase chain reaction analysis of the damage frequency in the promoter region of the transcriptionally active phosphoglycerate kinase (PGK 1) gene revealed that (Cu(II)/ascorbate/H2O2 caused DNA base damage by a sequence-dependent mechanism, with the 5' bases of d(pGn) and d(pCn) being damage hot spots, as were the most internal guanines of d(pGGGCCC) and d(pCCCGGG). Since base damage occurs after formation of a DNA-Cu(I)-H2O2 complex, these data suggest that the local DNA sequence affects formation of DNA-Cu(I)-H2O2 complexes and/or the efficiency of base oxidation during resolution of this complex.

Phase I study of mitomycin C and menadione in advanced solid tumors.

A phase I study of mitomycin C with menadione (2-methyl-1,4-naphthoquinone, a vitamin K analogue which lowers intracellular pools of reduced glutathione) was designed as an approach to overcoming tumor cell resistance to alkylating agent chemotherapy. Patients with refractory solid tumors (n = 51) were treated with a 48-h continuous intravenous infusion of menadione followed by a bolus intravenous dose of mitomycin C at the completion of the menadione infusion. Initial menadione doses of 8.0 and 4.0 g/m2 over 48 h were associated with hemolysis, so subsequent dose levels of menadione ranged from 1.0 to 3.0 g/m2 with mitomycin C from 5 to 20 mg/m2. All three patients treated with menadione at 8.0 g/m2 and the single patient treated at 4.0 g/m2 with mitomycin C at 5 mg/m2 developed clinically significant hemolysis despite the presence of red blood cell glucose-6-phosphate dehydrogenase. Subsequently, a revised escalation scheme for menadione was used, and all patients tolerated menadione doses of 1-2.5 g/m2 over 48 h with mitomycin C doses up to 20 mg/m2. Since the 3.0 g/m2 dose of menadione was associated with mild hemolysis in three of four patients, the maximum tolerated dose of menadione was established at 2.5 g/m2. All of the mitomycin dose levels were tolerated without unexpected toxicities attributable to the combination. Prolonged infusions of menadione at doses which have been associated with lowering of intracellular glutathione pools in short-term exposure are limited by dose-dependent hemolysis, probably due to depletion of erythrocyte glutathione by menadione-related redox cycling. There was no detectable deleterious effect of pre-exposure to menadione on mitomycin C tolerance. We recommend a combination of menadione at 2.5 g/m2 as a continuous intravenous infusion and mitomycin C at 15 mg/m2 for further study in solid tumors, for which treatment with single-agent mitomycin C is appropriate.