p53R2 as a novel prognostic biomarker in nasopharyngeal carcinoma.

BACKGROUND: p53R2 is a target of p53 gene, which is essential for DNA repair, mitochondrial DNA synthesis, protection against oxidative stress, chromosomal instability, chronic inflammation and tumorigenesis. This study is aimed to investigate the expression of ribonucleotide reductase (RR) subunit p53R2 in nasopharyngeal carcinoma and its significance in the prognosis. METHODS: The expression levels of p53R2 in 201 patients with NPC were examined by immunohistochemical assay. The correlations of p53R2 expression and clinicopathological features of nasopharyngeal carcinoma patient were analysed by chi-square test. The Kaplan-Meier survival analysis and Cox multivariate regression model were used to analyze the prognostic significance of the patients with NPC. RESULTS: Immunohistochemical results showed that p53R2 was positively expressed in 92.5% (186/201) of nasopharyngeal carcinoma and the high expression rate was 38.3% (77/201). Further analysis observed that the negative correlation between expression of p53R2 and pT status had statistical significance (P < 0.05). Kaplan-Meier survival analysis found that the mean survival time of patients with high expression of p53R2 was 143.32 months, while the patients with low expression level of p53R2 was 121.63 months (P < 0.05). Cox regression analysis suggested that p53R2 protein expression could be used as an independent prognostic factor for nasopharyngeal carcinoma (P < 0.05). CONCLUSIONS: This study drew a conclusion that p53R2 could be used as a prognostic biomarker indicative of the favorable outcome for patients with nasopharyngeal carcinoma.

Hug1 is an intrinsically disordered protein that inhibits ribonucleotide reductase activity by directly binding Rnr2 subunit.

Rad53 is a conserved protein kinase with a central role in DNA damage response and nucleotide metabolism. We observed that the expression of a dominant-lethal form of RAD53 leads to significant expression changes for at least 16 genes, including the RNR3 and the HUG1 genes, both of which are involved in the control of nucleotide metabolism. We established by multiple biophysical and biochemical approaches that Hug1 is an intrinsically disordered protein that directly binds to the small RNR subunit Rnr2. We characterized the surface of interaction involved in Hug1 binding to Rnr2, and we thus defined a new binding region to Rnr2. Moreover, we show that Hug1 is deleterious to cell growth in the context of reduced RNR activity. This inhibitory effect of Hug1 on RNR activity depends on the binding of Hug1 to Rnr2. We propose a model in which Hug1 modulates Rnr2-Rnr1 association by binding Rnr2. We show that Hug1 accumulates under various physiological conditions of high RNR induction. Hence, both the regulation and the mode of action of Hug1 are different from those of the small protein inhibitors Dif1 and Sml1, and Hug1 can be considered as a regulator for fine-tuning of RNR activity.

A phase I-II evaluation of veliparib (NSC #737664), topotecan, and filgrastim or pegfilgrastim in the treatment of persistent or recurrent carcinoma of the uterine cervix: an NRG Oncology/Gynecologic Oncology Group study.

PURPOSE: The aim of this study was to evaluate the tolerability and efficacy of poly(ADP-ribose) polymerase (PARP) inhibition by veliparib during cytotoxic topotecan administration with filgrastim or pegfilgrastim neutrophil support in women with persistent or recurrent uterine cervix cancer. EXPERIMENTAL DESIGN: This phase I-II trial examined twice-daily oral veliparib (10 mg) given during once-daily intravenous topotecan (0.6 mg/m²) on days 1 to 5 of each treatment cycle. Cycles were repeated every 21 days until disease progression or until toxicity prohibited further therapy. Toxicity and objective response rate were primary endpoints. RESULTS: Twenty-seven women were enrolled. Frequently reported grade 3 or higher treatment-related toxicities were anemia (59%), thrombocytopenia (44%), leukopenia (22%), and neutropenia (19%). There were 2 partial responses (7% [90% confidence interval, 1%-22%]). Four patients had a disease progression date more than 6 months after the start of veliparib-topotecan therapy. Patients with low immunohistochemical expression (0-1+) of PARP-1 in their primary uterine cervix cancer were more likely to have a longer progression-free interval (hazard ratio, 0.25; P = 0.02) and survival (hazard ratio, 0.12; P = 0.005) after veliparib-topotecan therapy. CONCLUSIONS: Clinical activity of a veliparib-topotecan combination was minimal in women with persistent or recurrent uterine cervix cancer. Women whose uterine cervix cancers express PARP-1 at low levels may benefit preferentially from PARP inhibitors combined with cytotoxic therapies, suggesting further study of PARP expression as an integral triage biomarker.

Biomarkers for predicting the response of esophageal squamous cell carcinoma to neoadjuvant chemoradiation therapy.

This review summarizes and evaluates the literature regarding the biomarkers for predicting the response and/or prognosis of esophageal squamous cell carcinoma (ESCC) patients treated with neoadjuvant chemoradiation therapy (CRT). There are seven categories of molecules known to correlate with the response and/or prognosis: tumor suppressors (p53, p21), cell cycle regulators (Cyclin D1, CDC25B, 14-3-3sigma), DNA repair molecules (p53R2, ERCC1), drug resistance proteins [metallothionein (MT)], angiogenic factors (VEGF), molecules involved in cell proliferation/invasion/metastasis (Ki-67, COX-2) and hedgehog signaling molecules (Gli-1). Of the above molecules, the tumor suppressor p53 is expected to be a representative biomarker for predicting the response and prognosis. The cell cycle markers CDC25B and 14-3-3sigma have potential as response biomarkers independent of the p53 status. The DNA repair markers, p53R2 or ERCC1, angiogenic molecule (VEGF), and hedgehog signaling pathway factor Gli-1 also have potential to predict the response and prognosis of ESCC. However, there are still many unanswered questions with regard to predicting the clinical effects of neoadjuvant CRT.

COP9 signalosome subunit 7 from Arabidopsis interacts with and regulates the small subunit of ribonucleotide reductase (RNR2).

The COP9 Signalosome protein complex (CSN) is a pleiotropic regulator of plant development and contains eight-subunits. Six of these subunits contain the PCI motif which mediates specific protein interactions necessary for the integrity of the complex. COP9 complex subunit 7 (CSN7) contains an N-terminal PCI motif followed by a C-terminal extension which is also necessary for CSN function. A yeast-interaction trap assay identified the small subunit of ribonucelotide reductase (RNR2) from Arabidopsis as interacting with the C-terminal section of CSN7. This interaction was confirmed in planta by both bimolecular fluorescence complementation and immuoprecipitation assays with endogenous proteins. The subcellular localization of RNR2 was primarily nuclear in meristematic regions, and cytoplasmic in adult cells. RNR2 was constitutively nuclear in csn7 mutant seedlings, and was also primarily nuclear in wild type seedlings following exposure to UV-C. These two results correlate with constitutive expression of several DNA-damage response genes in csn7 mutants, and to increased tolerance of csn7 seedlings to UV-C treatment. We propose that the CSN is a negative regulator of RNR activity in Arabidopsis.

Upregulation of the p53R2 ribonucleotide reductase subunit by nitric oxide.

The p53R2 ribonucleotide reductase subunit is a p53-inducible protein involved in DNA repair and mitochondrial DNA replication. It has been shown that p53 is activated by nitric oxide, which can damage DNA at high concentrations. This suggests that NO may regulate p53R2 expression through p53 activation. We show here that NO increases p53 protein expression in p53-wt cell lines and upregulates p53R2 at the protein and mRNA levels in a p53-dependent manner. Other p53 target genes, such as DDB2, WAF1 and PCNA, are also induced by NO. Surprisingly, p53R2 is similarly upregulated by NO in two p53-deficient cell lines, showing the existence of p53-independent regulatory mechanisms. Delta Np73, which is overexpressed in many cancers, inhibits the transcriptional activity of p53 and p53 homologs. In p53-wt cells, the Delta Np73alpha isoform inhibits basal and NO-induced p53R2 protein expression. In p53-null cells, it also strongly inhibits p53R2 expression, and represses the enhancer activity of the p53-responsive element present in the p53R2-encoding gene. These results demonstrate that p53R2 expression can be controlled by p53 homologs in the absence of p53, and is downregulated by oncogenic Delta Np73 isoforms. Knocking down p53R2 in p53-wt cells dramatically enhances NO-induced DNA damages, indicating a protective function of the p53R2 ribonucleotide reductase subunit in prevention or repair of NO-mediated genotoxic injury.

Clb6-Cdc28 Promotes Ribonucleotide Reductase Subcellular Redistribution during S Phase.

A tightly controlled cellular deoxyribonucleotide (deoxynucleoside triphosphate [dNTP]) pool is critical for maintenance of genome integrity. One mode of dNTP pool regulation is through subcellular localization of ribonucleotide reductase (RNR), the enzyme that catalyzes the rate-limiting step of dNTP biosynthesis. In Saccharomyces cerevisiae, the RNR small subunit, Rnr2-Rnr4, is localized to the nucleus, whereas the large subunit, Rnr1, is cytoplasmic. As cells enter S phase or encounter DNA damage, Rnr2-Rnr4 relocalizes to the cytoplasm to form an active holoenzyme complex with Rnr1. Although the DNA damage-induced relocalization requires the checkpoint kinases Mec1-Rad53-Dun1, the S-phase-specific redistribution does not. Here, we report that the S-phase cyclin-cyclin-dependent kinase (CDK) complex Clb6-Cdc28 controls Rnr2-Rnr4 relocalization in S phase. Rnr2 contains a consensus CDK site and exhibits Clb6-dependent phosphorylation in S phase. Deletion of CLB6 or removal of the CDK site results in an increased association of Rnr2 with its nuclear anchor Wtm1, nuclear retention of Rnr2-Rnr4, and an enhanced sensitivity to the RNR inhibitor hydroxyurea. Thus, we propose that Rnr2-Rnr4 redistribution in S phase is triggered by Clb6-Cdc28-mediated phosphorylation of Rnr2, which disrupts the Rnr2-Wtm1 interaction and promotes the release of Rnr2-Rnr4 from the nucleus.

Identification of RNR4, encoding a second essential small subunit of ribonucleotide reductase in Saccharomyces cerevisiae.

Ribonucleotide reductase (RNR), which catalyzes the rate-limiting step for deoxyribonucleotide production required for DNA synthesis, is an alpha2beta2 tetramer consisting of two large and two small subunits. RNR2 encodes a small subunit and is essential for mitotic viability in Saccharomyces cerevisiae. We have cloned a second essential gene encoding a homologous small subunit, RNR4. RNR4 and RNR2 appear to have nonoverlapping functions and cannot substitute for each other even when overproduced. The lethality of RNR4 deletion mutations can be suppressed by overexpression of RNR1 and RNR3, two genes encoding the large subunit of the RNR enzyme, indicating genetic interactions among the RNR genes. RNR2 and RNR4 may be present in the same reductase complex in vivo, since they coimmunoprecipitate from cell extracts. Like the other RNR genes, RNR4 is inducible by DNA-damaging agents through the same signal transduction pathway involving MEC1, RAD53, and DUN1 kinase genes. Analysis of DNA damage inducibility of RNR2 and RNR4 revealed partial inducibility in dun1 mutants, indicating a DUN1-independent branch of the transcriptional response to DNA damage.

UV-C response of the ribonucleotide reductase large subunit involves both E2F-mediated gene transcriptional regulation and protein subcellular relocalization in tobacco cells.

E2F factors are implicated in various cellular processes including specific gene induction at the G1/S transition of the cell cycle. We present in this study a novel regulatory aspect for the tobacco large subunit of ribonucleotide reductase (R1a) and its encoding gene (RNR1a) in the UV-C response. By structural analyses, two E2F sites were identified on the promoter of this gene. Functional analysis showed that, in addition to their role in the specific G1/S induction of the RNR1a gene, both E2F sites were important for regulating specific RNR1a gene expression in response to UV-C irradiation in non-synchronized and synchronized cells. Concomitantly, western blot and cellular analyses showed an increase of a 60 kDa E2F factor and a transient translocation of a GFP-R1a protein fusion from cytoplasm to nucleus in response to UV irradiation.

Effects of N-hydroxy-N'-aminoguanidine isoquinoline in combination with other inhibitors of ribonucleotide reductase on L1210 cells.

The 1-isoquinolylmethylene derivative of N-hydroxy-N'-aminoguanidine (HAG) is the most potent agent of the recently synthesized series of HAG-derived ribonucleotide reductase inhibitors. To potentiate the effects of the HAG-isoquinoline drug [HAG-1-isoquinolylmethylene tosylate (HAG-IQ)], we combined it with other inhibitors of ribonucleotide reductase. Using mouse leukemia L1210 cell cultures, we tested drug combinations for their cytostatic and cytotoxic properties and for their effects on intracellular ribonucleotide reductase activity and nucleic acid synthesis. Deoxyguanosine or deoxyadenosine combined with HAG-IQ inhibited cell growth in an additive manner; three-drug combinations, HAG-IQ plus either deoxyguanosine/8-aminoguanosine or deoxyadenosine/deoxycoformycin, were strongly synergistic. When Desferal, an iron chelator, was added to these combinations, the four-drug combinations increased inhibition of cell growth and increased cytotoxicity. The intracellular target of these drug combinations in L1210 cells was the ribonucleotide reductase site. The formation of deoxycytidine from [14C]cytidine and incorporation into DNA were markedly inhibited by these drug combinations, while RNA synthesis was unaffected. These data show that the antiproliferative and cytotoxic effects of HAG-IQ, a potent inhibitor by itself, can be further potentiated in combinations with other ribonucleotide reductase inhibitors.

Molecular mechanisms of drug resistance involving ribonucleotide reductase: hydroxyurea resistance in a series of clonally related mouse cell lines selected in the presence of increasing drug concentrations.

Mammalian ribonucleotide reductase is a highly regulated, rate-limiting activity responsible for converting ribonucleoside diphosphates to the deoxyribonucleotide precursors of DNA. The enzyme consists of two nonidentical proteins often called M1 and M2, both of which are required for activity. Hydroxyurea is an antitumor agent which inhibits ribonucleotide reductase by interacting with the M2 component specifically at a unique tyrosyl free radical. To obtain further information about drug resistance mechanisms, we have used M1 and M2 complementary DNAs and monoclonal antibodies to investigate the properties of a series of clonally related drug-resistant mouse cell lines, selected by a step-wise procedure for increasing levels of resistance to the cytotoxic effects of hydroxyurea. Several interesting mechanisms have been identified. Each successive drug selection step leading to the isolation of highly resistant cells was accompanied by stable elevations in cellular resistance and ribonucleotide reductase activities. The changes that occurred at each step involved the M2 component. A very early event, occurring at the first step in the selection process, was the amplification of the M2 gene accompanied by an increase in M2 messenger RNA. Although cellular resistance and M2 protein levels increased significantly during drug selection, only a modest change in M2 gene copy number was observed after the initial selection step. Analysis of wild type, moderately resistant, and highly resistant cells indicated that, in addition to M2 gene amplification, posttranscriptional modification also occurred during drug selection. This second mechanism was not due to alterations in protein M2 half-life, but involved an increase in translational efficiency. By increasing the rate of M2 synthesis, without altering degradation rates, resistant cells were able to accumulate high levels of this key regulatory protein. Cells selected for the ability to proliferate in concentrations of drug as high as 4 mM exhibited changes that involved M2, without detectable changes to M1. These results provide further evidence that M1 and M2 levels are controlled by different mechanisms in mammalian cells. Eventually, however, cells required an elevation in the M1 protein, as well as the M2 protein, to survive in a hydroxyurea concentration of 5 mM. These results illustrate the complexity of the drug-resistant phenotype and provide further information about the molecular processes that lead to the development of cells resistant to low, intermediate, and high concentrations of hydroxyurea.

Combination chemotherapy directed at the components of nucleoside diphosphate reductase.

It would be expected that drugs directed at the rate-limiting step in a key metabolic pathway in tumor cell proliferation would provide a useful basis for therapy of neoplasms. Ribonucleotide reductase catalyzes the rate-limiting step in the de novo synthesis of dNTP's for DNA synthesis. Further, ribonucleotide reductase is composed of two non-identical protein subunits (non-heme iron and effector-binding subunits) which can be specifically and independently inhibited. As a result, combinations of drugs specifically directed at each of the subunits of ribonucleotide reductase have been shown to cause synergistic inhibition of L1210 cell growth in culture and synergistic cell kill. This approach offers a novel basis for the design of combination chemotherapy.

The ribonucleotide reductase subunit M2B subcellular localization and functional importance for DNA replication in physiological growth of KB cells.

Ribonucleoside diphosphate reductase (EC 1.17.4.1) (RR) is a potential target for antineoplastic agents due to its crucial role in DNA replication and repair. The expression and activity of RR subunits are highly regulated to maintain an optimal dNTP pool, which is required to maintain genetic fidelity. The human RR small subunit M2B (p53R2) is thought to contribute to DNA repair in response to DNA damage. However, it is not clear whether M2B is involved in providing dNTPs for DNA replication under physiological growth conditions. Serum starvation synchronized studies showed that a rapid increase of M2B was associated with cyclin E, which is responsible for regulation of G(1)/S-phase transition. A living cell sorting study that used KB cells in normal growth, further confirmed that M2B increased to maximum levels at the G(1)/S-phase transition, and decreased with DNA synthesis. Confocal studies revealed that M2B redistributed from the cytoplasm to the nucleus earlier than hRRM2 in response to DNA replication. Nuclear accumulation of M2B is associated with dynamic changes in dNTP at early periods of serum addition. By using M2B-shRNA expression vectors, inhibition of M2B may result in growth retardation in KB cells. We conclude that M2B may translocate from the cytoplasm into the nucleus and allow dNTPs to initiate DNA synthesis in KB cells under physiological conditions. Thus, our findings suggested that M2B might play an important role for initiating DNA replication of KB cells in normal growth.

Involvement of ribonucleotide reductase activity in the senescence of normal human diploid fibroblasts.

The levels of intracellular ribonucleotide reductase activity, a highly regulated rate-limiting step in DNA synthesis, were investigated during serial subculture of normal human diploid fibroblasts in vitro. This key enzyme activity was found to decline significantly during cellular senescence. This observation along with previous findings of a mutator gene associated with mammalian ribonucleotide reductase suggests a possible mutation mechanism for aging which involves changes in reductase activity during cellular senescence. Furthermore, in keeping with the decrease in enzyme activity, we show that cell resistance to the antitumor agent hydroxyurea, whose site of action is ribonucleotide reductase, decreases progressively with increasing passage numbers. This indicates that an important factor to be considered in drug therapy aimed at the reductase is the increased sensitivity of normal cells to drug with cell age, due to a decline in enzyme activity. Much remains to be determined about age-dependent factors involved in drug therapy; cultured normal human diploid fibroblasts provide a useful system in which to investigate these important parameters.

Effector studies of 3'-azidothymidine nucleotides with human ribonucleotide reductase.

The 5'-mono-, di- and triphosphate derivatives (N3dTMP, N3dTDP and N3dTTP respectively) of 3'-azidothymidine (N3dThd), a new drug for the treatment of the acquired immune deficiency syndrome (AIDS), were synthesized. The abilities of these analog nucleotides to mimic the effector properties of the corresponding thymidine nucleotides with human ribonucleotide reductase were studied. Surprisingly, the mode of inhibition of CDP reduction by dTTP and dTDP was found to be competitive versus CDP. The Ki values were 22 and 78 microM respectively. Inhibition by N3dTTP and N3dTDP was considerably weaker, with Ki values of 1200 and 550 microM. Neither dTMP nor N3dTMP produced significant inhibition at concentrations up to 500 microM. dTTP was an essential activator for GDP reduction. In the presence of the accessory activator, ATP, the activation constant for dTTP was 7.8 microM. N3dTTP was neither an activator of GDP reduction nor an inhibitor of the activation by dTTP. In view of the intracellular concentrations of these analog nucleotides reached after incubations with N3dThd [Furman et al., Proc. natn. Acad. Sci. U.S.A. 83, 8333 (1986)] and the weakness of their interactions with ribonucleotide reductase, it is unlikely that the antiviral or toxic effects of N3dThd can be attributed to direct effects on this enzyme. The possible indirect effects caused by alterations in the pools of the natural effectors are discussed.

Molecular markers and targeted therapy with novel agents: prospects in the treatment of non-small cell lung cancer.

Detection of genomic differences predictive of drug response or resistance in individual patients may allow therapy to be customized to the characteristics of particular tumors. Preliminary findings are that non-small cell lung cancer patients overexpressing ERCC1 mRNA have lower response to cisplatin chemotherapy, while those overexpressing ribonucleotide reductase mRNA have limited benefit from gemcitabine. In addition, overexpression of beta-tubulin III and stathmin can influence the sensitivity to microtubule interacting drugs, like vinorelbine and paclitaxel. The introduction of biological agents which target highly specific intracellular pathways offers the promise of enhancing the efficacy of cytotoxic chemotherapy. Among many promising biological agents is the monoclonal antibody C225, which blocks the EGFR receptor. The addition of C225 appears to induce responses in a proportion of colon cancer patients refractory to 5-FU or irinotecan, supporting pre-clinical evidence of synergistic activity. It also appears from xenograft data that C225 enhances the sensitivity of tumors to radiation and docetaxel or the combination.

Improved method for the measurement of ribonucleotide reductase activity.

An improved method for the measurement of herpes simplex virus type 1 encoded ribonucleotide reductase has been developed. The enzyme which catalyses the conversion of ribonucleoside diphosphates to deoxyribonucleoside diphosphates was determined by first converting the ribonucleotide substrate and deoxyribonucleotide product to the corresponding nucleosides by treatment with snake venom phosphodiesterase. Then nucleosides were separated by HPLC and measured by flow through scintillation counting and by monitoring their absorbance at 254 nm. Under the conditions used in the experiment cytidine and deoxcytidine, the derivitised substrate and product respectively, eluted from the column at approximately 4 min 33 s and 6 min 24 s. Peak heights and areas were automatically calculated by computer to ascertain the amount of product formed and thus quantitate the assay. Automation of the assay from sample injection to analysis provides a significant saving in time and an improvement in the efficiency of measurement of ribonucleotide reductase activity over other published methods.

Rapid sampling of multiple enzyme reactions.

A simple method of initiating and sampling six simultaneous reactions was devised. A commercially available vial rack was fitted with a Plexiglas overlaying sheet to stabilize the vials for the addition and sampling procedures. Glass vials were routinely used because of their thermal conductivity advantages. Samples were added and removed, and the reactions were mixed with a multichannel pipet using every other channel. The data showing six simultaneous progress curves for the rapid inactivation of herpes simplex virus ribonucleotide reductase were presented and analyzed. In addition, the time course of 12 reactions catalyzed by varicella zoster virus thymidine kinase were assayed at one min intervals generating 96 data points within 8.5 min. A second experiment generated data points every 30 s for six simultaneous replicate thymidine kinase reactions. The ease of use and high reproducibility of the method are demonstrated by these data.

Biochemical and antitumor activity of trimidox, a new inhibitor of ribonucleotide reductase.

Trimidox (3,4,5-trihydroxybenzamidoxime), a newly synthesized analog of didox (N,3,4-trihydroxybenzamide) reduced the activity of ribonucleotide reductase (EC 1.17.4.1) in extracts of L1210 cells by 50% (50% growth-inhibitory concentration, IC50) at 5 microM, whereas hydroxyurea, the only ribonucleotide reductase inhibitor in clinical use, exhibited an IC50 of 500 microM. Ribonucleotide reductase activity was also measured in situ by incubating L1210 cells for 24 h with trimidox at 7.5 microM, a concentration that inhibits cell proliferation by 50% (IC50) or at 100 microM for 2 h; these concentrations resulted in a decrease in enzyme activity to 22% and 50% of the control value, respectively. Trimidox and hydroxyurea were cytotoxic to L1210 cells with IC50 values of 7.5 and 50 microM, respectively. Versus ribonucleotide reductase, trimidox and hydroxyurea yielded IC50 values of 12 and 87 microM, respectively. A dose-dependent increase in life span was observed in mice bearing intraperitoneally transplanted L1210 tumors. Trimidox treatment (200 mg/kg; q1dx9) significantly increased the life span of mice bearing L1210 leukemia (by 82% in male mice and 112% in female mice). The anti-tumor activity appeared more pronounced in female mice than in male mice. Viewed in concert, these findings suggest that trimidox is a new and potent inhibitor of ribonucleotide reductase and that it is a promising candidate for the chemotherapy of cancer in humans.

Effects of iron, copper and zinc on the activity of ribonucleotide reductase in normal and leukemic human lymphocytes.

The activity of ribonucleotide reductase (RR) extracted from normal and leukemic human lymphocytes was assessed. The activity of the enzyme was five times higher in leukemic cells than in normal lymphocytes. Fe and Cu stimulated and Zn inhibited RR activity in both types of cells. Zn also partially reduced the stimulatory effects of Fe and Cu. Deferoxamine (DFX) alone inhibited the activity of the enzyme and enhanced a similar inhibitory action of hydroxyurea (HU). In both cases the inhibitory effects were reversed by Fe and Cu, but not Zn. These findings suggest that: 1) Fe, Cu and Zn may modulate RR activity in vitro and their action is related and interdependent. 2) DFX alone may inhibit RR activity. 3) DFX enhances HU effect on leukemic cells. 4) The effects of both inhibitors may be modified by trace metals.