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

Comparison of Cytotoxicity of Mononucleoside Phosphotriester Derivatives Bearing Biolabile Phosphate Protecting Groups in Normal Human Bone Marrow Progenitor Cells

The effects of three mononucleoside phosphotriester derivatives of 3′-azido-2′,3′-dideoxythymidine (AZT) which incorporate biolabile phosphate protecting groups, namely S-acetyl-2-thioethyl (MeSATE), S-(2-hydroxyethylsulfidyl)-2-thioethyl (DTE), and pivaloyloxymethyl (POM) were studied and compared to their nucleoside parent in human myeloid colony-forming cells. Moreover, the relative antiviral potency of these three pronucleotides were determined in primary human peripheral blood mononuclear cells infected with human immunodeficiency virus type 1. The results indicate that the SATE and DTE pro-moieties, as well as their degradation products, do not induce additional toxicity. The bis(MeSATE) phosphotriester derivative of AZT emerged as the most selective inhibitor with an in-vitro therapeutic index of the same order of magnitude as observed for AZT. This study has been extended to the corresponding bis(MeSATE) and bis(DTE) phosphotriester derivatives of 2′,3′-dideoxyuridine (ddU).

Comparison of the Mechanism of Toxicity of 3′-Azido-3′-Deoxythymidine and 3′-Fluoro-3′-Deoxythymidine in Human Lymphocytes

The thymidine analogue 3′-azido-3′-deoxythymidine (AZT), whilst a useful drug for the treatment of acquired immunodeficiency syndrome, produces toxic side-effects which can be severe and can interfere with therapy. The toxic mechanism of AZT is unknown. We have investigated the relationship between the phosphorylation and effect on natural dNTP pools of 3′-azido-3′-deoxythymidine and the closely related 3′-fluoro-3′-deoxythymidine and their toxicity in human lymphocytes. We attempted to reduce the toxicity by co-administration of natural nucleosides.

The toxicity of 3′-fluoro-3′-deoxythymidine could be reduced with thymidine or deoxyuridine five- and 10-fold, respectively. The toxicity of 3′-azido-3′-deoxythymidine could be reduced twofold with cytidine or uridine but was increased by all other nucleosides, including thymidine. Neither analogue caused significant changes in the dNTP pools at cytotoxic concentrations; the effect of the nucleosides in reducing toxicity was not owing to replacement of a depleted dNTP. Thymidine reduced the phosphorylation of 3′-azido-3′-deoxythymidine and 3′-fluoro-3′-deoxythymidine 6 and 17 times, respectively.

3′-azido-3′-deoxythymidine and 3′-fluoro-3′-deoxythymidine appear to have different mechanisms of toxicity. The toxic mechanism of 3′-fluoro-3′-deoxythymidine is probably inhibition of cellular DNA synthesis by the triphosphate. The toxicity of 3′-azido-3′-deoxythymidine in lymphocytes does not appear to be directly related to the amounts of the phosphorylated forms. The mechanism may be interference with RNA metabolism or precursors, perhaps by the nucleoside.

Effect of G-CSF and M-CSF on the in vitro Toxicity Associated with Zidovudine in Normal Human Bone Marrow Haematopoietic Progenitor Stem Cells

Zidovudine, the antiviral drug used in the treatment of acquired immunodeficiency syndrome (AIDS), causes toxicity to the haematopoietic system. Although use of the haematopoietic growth factors, GM-CSF and erythropoietin have been investigated in clinical trials to modulate antiviral toxicity, there is scant data which supports their ability to ameliorate zidovudine induced toxicity on haematopoietic progenitor cells when combined in vitro. We describe here the results of studies designed to evaluate the capacity of additional haematopoietic factors such as granulocyte-colony stimulating factor (G-CSF) and macrophage-colony stimulating factor (M-CSF) to modulate zidovudine-induced toxicity on G-CSF and M-CSF dependent-colony formation in the presence or absence of zidovudine in vitro. These factors were also studied combined with erythropoietin in culture for the early erythroid progenitor BFU-E using adherent, T-cell, depleted normal human bone marrow cells in the presence or absence of zidovudine. In the presence of zidovudine at the concentration producing 50% inhibition of G- and M-CSF dependent colony formation, (5 × 10−5M), dose-escalation of either G-CSF or M-CSF failed to ameliorate zidovudine toxicity. However, in the presence of zidovudine at the concentration that produces 50% inhibition of BFU-E (5 × 10−9M), and optimal erythropoietin (1 unit ml−1), G-CSF ameliorated zidovudine inhibition of BFU-E, which was not observed with M-CSF. In the presence of erythropoietin, G-CSF increased significantly normal BFU-E. These studies indicate that G-CSF may be useful in ameliorating zidovudine-induced anaemia and suggest G-CSF may act as a synergistic factor to enhance erythropoietin to support the growth of erythroid progenitors in conditions where erythropoitin is ineffective.

Studies on the Reversal of Azidothymidine Toxicity in Human Lymphocytes by Cytidine and Uridine

The toxicity of 3′-azidothymidine (AZT) in human lymphocytes has been shown previously to be reversed by co-incubation with the ribonucleosides cytidine or uridine. In the present paper, the effects of 3′-azidothymidine and cytidine/uridine, both alone and in combination, were studied upon key processes in lymphocytes in order to discover more about the mechanism of toxicity reversal.

In these experiments 3′-azidothymidine had only minor effects on the ribonucleoside triphosphate pools. Cytidine increased the CTP pool, and uridine the UTP pool. Co-incubation with AZT caused similar changes to incubation with cytidine or uridine alone. Toxicity reversal was not linked to replacement of deficient ribonucleoside triphosphate pools.

3′-Azidothymidine caused the excretion of thymidine from lymphocytes. Incubation with cytidine and uridine increased the intracellular cytidine and uridine pools, respectively. Co-incubation with 3′-azidothymidine increased still further the intracellular cytidine and uridine pools. Cytidine and uridine did not affect the intracellular 3′-azidothymidine pool.

The toxicity of 3′-azidothymidine was increased by co-incubation with the bases adenine, guanine, hypoxanthine, and uracil, but not by dihydrouracil, thymine, or xanthine.

Possible New Reaction Mechanisms of Dideoxynucleosides as Anti-Aids Drugs

Evidence is presented that a major class of drugs, the dideoxynucleosides (ddNs) and nucleoside/nucleotide analogues, may inhibit the symptoms of acquired immunodeficiency syndrome (AIDS) by initiation of inactivation at the ribonucleotide reductase (RNR) enzyme stage and/or inactivation of reverse transcriptase enzyme or at a stage more initial than that of the currently accepted DNA chain termination hypothesis. For example, it has been previously shown that ribonucleotide diphosphate reductase (RDPR) and ribonucleotide triphosphate reductase (RTPR) are inactivated with 2′-chloro-2 ‘-deoxyuridine 5′-diphosphate-([3′-3H]ClUDP) and triphosphate ([3′-3H]ClUTP) by reaction with an intermediate furanone, Scheme 2. RDPR has also been inactivated by 2‘-azido-2‘-deoxyuridine 5‘-diphosphate (N3UDP). Furthermore, addition of hydroxyurea to RNR can inhibit DNA synthesis which results in a rapid depletion of limiting deoxynucleotide triphosphate (dNTP) pools. There are similar perturbations of dNTP pools upon interaction of human RNR with 3‘-azido-2‘,3 ‘-dideoxythymidine (AZT), in human cell studies involving AZT/HIV and in adenosine/coformycin experiments in relation to inherited immunodeficiency, Table 1. Also, the herein proposed reduction mechanisms of nucleotides by RNR ( e.g., a single electron transfer from the nucleotide base to the phenol moiety of the tyrosyl radical of RNR via a pathway involving the thiyl radical of a cysteine residue) can also account for the chemistry of some antiretroviral drugs, the ddNs. Analyses are presented that the RNR reductions of regular unsubstituted nucleotides may occur predominantly via initial 2’ C-H abstraction instead of the originally proposed 3’ C-H abstraction mechanism. Also, it is noted that the fate of the phenol moiety of the tyrosyl unit in some RNR reactions with 2‘-halo-2‘-deoxynucleotides is not clear. The proposed reaction mechanisms may provide guidance for the development of potentially effective anti-AIDS drugs.

Effect of combination interleukin-3 (IL-3) and granulocyte-macrophage colony stimulating factor (GM-CSF) on hematopoiesis administered to retrovirus-infected immunodeficient mice receiving dose-escalation zidovudine (AZT)

We have previously demonstrated that continuous administration of dose-escalation zidovudine (AZT) in either normal or LP-BM5 MuLV immunodeficient virus-infected mice (MAIDS) was associated with the development of anemia, neutropenia, and thrombocytopenia. Hematopoietic growth factors/cytokines are being evaluated to determine their efficacy in ameliorating the hematopoietic toxicity associated with AZT. In normal mice receiving AZT, an increase in only plasma erythropoietin and not GM-CSF, Meg-CSF or TNF-alpha has been reported. This article describes studies that investigated the effect of combination interleukin-3 (IL-3) and granulocyte-macrophage colony stimulating factor (GM-CSF) administered in normal non-viral, viral-infected, and viral-infected C57BL6 mice receiving dose-escalation AZT, i.e. 0.1 mg/ml, 1.0 mg/ml, and 2.5 mg/ml placed in the drinking water. Non-viral control mice responded to IL-3/GM-CSF by increasing erythropoiesis, myelopoiesis and platelet production measured by increased bone marrow and spleen derived erythroid, myeloid and platelet precursor stem cells cultured in semi-solid media. Virus-infected control mice not receiving IL-3/GM-CSF developed pancytopenia. Administration of IL-3/GM-CSF to virus-infected mice receiving dose-escalation AZT did not ameliorate the peripheral pancytopenia associated with immunodeficiency disease and AZT treatment, even though erythroid, myeloid and platelet precursor progenitor cells were increased at certain times when compared to either normal or viral-infected mice receiving IL-3/GM-CSF. These results indicate that the combination use of IL-3 and GM-CSF in vivo is only a partially effective growth factor/cytokine treatment to ameliorate the hematopoietic toxicity associated with the use of the anti-viral drug zidovudine.

Disparate actions of hydroxyurea in potentiation of purine and pyrimidine 2',3'-dideoxynucleoside activities against replication of human immunodeficiency virus

We and other groups have recently reported the potentiation by ribonucleotide reductase inhibitors such as hydroxyurea of the anti-human immunodeficiency virus type 1 (HIV-1) activity of purine and pyrimidine 2',3'-dideoxynucleosides in both resting and phytohemagglutinin-stimulated peripheral blood mononuclear cells. Little agreement prevails, however, as to the mechanism of the synergistic effects described. We report here that in phytohemagglutinin-stimulated peripheral blood mononuclear cells, two mechanisms exist for the potentiation of the anti-HIV-1 activity by low-dose hydroxyurea of the purine-based dideoxynucleoside 2',3'-dideoxyinosine and the pyrimidine-based dideoxynucleosides 3'-azido-3'-deoxythymidine and 2',3'-dideoxycytidine. For 2',3'-dideoxyinosine, the enhancement arises from a specific depletion of dATP by hydroxyurea, resulting in a favorable shift of the 2',3'-dideoxyadenosine 5'-triphosphate/dATP ratio. For the pyrimidine dideoxynucleosides 3'-azido-3'-deoxythymidine and 2',3'-dideoxycytidine, the more modest anti-HIV enhancement results from hydroxyurea-induced increases of pyrimidine kinase activities in the salvage pathway and, hence, increased 5'-phosphorylation of these drugs, while depletion of the corresponding deoxynucleoside 5'-triphosphates (dTTP and dCTP) plays no significant role.

Effects of 2',3'-dideoxynucleosides on proliferation and differentiation of human pluripotent progenitors in liquid culture and their effects on mitochondrial DNA synthesis

2',3'-Dideoxynucleosides (ddNs) including 3'-azido-3'-deoxythymidine (AZT), 3'-fluoro-3'-deoxythymidine (FLT), 3'-amino-3'-deoxythymidine (AMT), 2',3'-dideoxycytidine (ddC), and 2',3'-didehydro-3'-deoxythymidine (D4T) were tested for their effects on proliferation and differentiation of pluripotent progenitor cells (CD34+) purified from human bone marrow cells grown in liquid cultures. These highly purified progenitor cells undergo extensive proliferation during 14 days, with a marked differentiation during the last 7 days. These differentiated cells exhibit normal morphological features in response to specific hematopoietic growth factors of both erythroid and granulocyte-macrophage lineages, as demonstrated by flow cytometry cell phenotyping. The potencies of these ddNs in inhibiting proliferation of granulocyte-macrophage lineage cells were in the order FLT > AMT = ddC > AZT >> D4T, and the potencies in inhibiting proliferation of erythroid lineage cultures were in the order FLT > AMT > AZT > ddC >> D4T. The toxic effects of ddNs assessed in these liquid cultures were in agreement with data obtained by using semisolid cultures, demonstrating the consistency of these two in vitro hematopoietic systems toward ddN toxicity. ddC was toxic to CD34+ progenitor cells and/or cells in the early stages of differentiation, whereas the inhibitory effect of AZT on the erythroid lineage was predominantly observed on a more mature population of erythroid progenitors during the differentiation process. Slot blot analysis of granulocyte-macrophage cultures demonstrated that exposure to ddC and FLT was associated with a decrease in total mitochondrial DNA (mtDNA) content, suggesting that these two ddNs inhibit mtDNA synthesis. In contrast, no difference in the ratio of nuclear DNA to mtDNA was observed in cells exposed to toxic concentrations of AZT and AMT is not associated with an inhibition of mtDNA synthesis. This human pluripotent progenitor liquid culture system should permit detailed investigations of the cellular and molecular events involved in ddN-induced hematological toxicity.

Influence of interleukin-3 on zidovudine (AZT)-induced in vitro toxicity to human hematopoietic progenitors

Zidovudine (AZT), the antiviral drug used in the treatment of acquired immunodeficiency syndrome (AIDS), produces some toxicity to the hematopoietic system. Although several hematopoietic growth factors are currently undergoing clinical trials to evaluate their ability to modulate antiviral toxicity, there are scant data which support their ability to ameliorate AZT toxicity on hematopoietic progenitor cells when combined in vitro. We describe in this report the results of studies designed to evaluate in vitro the capacity of the cytokine interleukin-3 (IL-3), in dose-escalation fashion, to modulate AZT toxicity on normal human marrow derived granulocyte/erythroid/macrophage/megakaryocyte colony-forming units (CFU-GEMM), CFU-granulocyte/macrophage (CFU-GM) and erythroid burst-forming units (BFU-E). Colony formation for each progenitor was increased in the presence of IL-3 compared to cultures plated in its absence. In the presence of AZT (ID50 dose, used for each progenitor), IL-3 reduced AZT toxicity, with the most significant response observed for CFU-GEMM, indicating IL-3 may exert an effect on early, less differentiated hematopoietic progenitors. These studies indicate IL-3 may be an effective agent in reversing the hematopoietic toxicity associated with AZT; however, further in vivo studies are required before clinical use of IL-3 is advocated.

Human ribonucleotide reductase. Activation and inhibition by analogs of ATP

Sixteen ATP analogs were studied as activators of CDP reduction catalyzed by human ribonucleotide reductase. Activation constants were determined. Three analogs, 3-deazaATP, 5'-adenylimidodiphosphate, and 3'-dATP, activated approximately as efficiently as ATP. Four analogs were partial activators. These seven activators were also accessory activators of GDP reduction. Furthermore, two other analogs, adenosine-5'-O-(1-thiotriphosphate) and 8-bromoATP, selectively stimulated GDP reduction. Ten analogs, at equal molar concentrations with ATP, inhibited ATP-dependent activation of CDP reduction and/or accessory activation of GDP reduction by greater than 45%. No analog inhibited as potently as 2'-dATP, which had an IC50 of 30-50 microM versus the stimulation of CDP and GDP reduction by 2.0 mM ATP.

Herpes and human ribonucleotide reductases. Inhibition by 2-acetylpyridine 5-[(2-chloroanilino)-thiocarbonyl]-thiocarbonohydrazone (348U87)

The mode of inactivation of herpes simplex virus type 1 and human ribonucleotide reductases by 2-acetylpyridine 5-[(2-chloroanilino)-thiocarbonyl]-thiocarbonohydrazone++ + (348U87) was determined and compared to that described previously [Porter et al. Biochem Pharmacol 39: 639-646, 1990] for 2-acetylpyridine 5-[(dimethylamino)thiocarbonyl]-thiocarbonohydrazone (A1110U). 348U87 inactivated herpes ribonucleotide reductase faster than did A1110U. Moreover, iron-complexed 348U87 was a considerably more potent inactivator than iron-complexed A1110U. It appeared to efficiently form an initial complex with the viral enzyme prior to rapid enzyme inactivation. The combination of 348U87 and iron-complexed 348U87 inactivated with a rate constant that was slightly greater than the sum of their individual rate constants of inactivation. The corresponding combination of A1110U species inactivated with a rate constant that was much greater than the sum of the individual rate constants of inactivation. Herpes ribonucleotide reductase that had been inactivated by either species of 348U87 was reactivated by diluting the enzyme and inactivators into assay medium containing excess iron. 348U87 was also an effective inactivator of herpes simplex virus type 2 and varicella zoster virus ribonucleotide reductases. The iron-complexed forms of 348U87 and A1110U exhibited very different modes of inactivation of human ribonucleotide reductase. Iron-complexed 348U87 was a tight-binding inactivator, whereas iron-complexed A1110U was only a weak, non-inactivating, inhibitor. Furthermore, the inactivation by iron-complexed 348U87 was not stimulated by either 348U87 or A1110U, whereas the weak inhibition by iron-complexed A1110U was converted to rapid inactivation by A1110U. Excess iron prevented the inactivation by iron-complexed 348U87. Uncomplexed 348U87 was similar to uncomplexed A1110U in that it was not an inhibitor of the human enzyme.

Herpes simplex virus type 1 ribonucleotide reductase: selective and synergistic inactivation by A1110U and its iron complex

2-Acetylpyridine-5-[(dimethylamino)thiocarbonyl]thiocarbonohydr azone (A1110U) inactivated herpes simplex virus Type 1 ribonucleotide reductase (EC 1.17.4.1) by a first-order process (kinact) which had a maximum value (Mkinact) of 8 hr-1 and a Kd that was less than 1 microM. The stable complex between iron and A1110U, (A1110U)2Fe+i, inactivated this enzyme with a Mkinact of 7 hr-1 and a Kd of 7 microM. Free A1110U and its iron-complex synergized as inactivators of the enzyme. For example, the kinact for the combination of 2 microM A1110U and 1 microM (A1110U)2Fe+i as independent inactivators was calculated to be about 9 hr-1, while the observed value was 32 hr-1. The bimolecular rate constant for inactivation of the viral enzyme by (A1110U)2Fe+i in the presence of a saturating concentration of A1110U was 2.5 10(7) M-1 hr-1 at 30 degrees. Human ribonucleotide reductase was less sensitive to the inhibitory effects of A1110U and its iron-complex. This enzyme was neither inhibited nor inactivated by A1110U and was weakly inhibited by (A1110U)2Fe+i. Furthermore, inactivation required the combination of A1110U and (A1110U)2Fe+i. The bimolecular rate constant for inactivation of human ribonucleotide reductase by (A1110U)2Fe+i in the presence of a saturating concentration of A1110U was considerably smaller (3.8 10(6) M-1 hr-1 at 37 degrees) than the analogous constant for the viral enzyme. Several iron-chelating reagents with unrelated structures substituted for free A1110U in its various roles with both enzymes. However, the iron complexes of these alternative chelators did not substitute for (A1110U)2Fe+i. The rates of inactivation of both enzymes were independent of the oxidation state of iron in (A1110U)2Fe+i and of CDP concentration. The inactivated enzymes were reactivated rapidly by FeSO4, but were not reactivated by CoCl2, CuSO4, or NiCl2. MnCl2 inhibited reactivation of the viral enzyme by FeSO4.

Hydroxyurea increases the phosphorylation of 3'-fluorothymidine and 3'-azidothymidine in CEM cells

The triphosphates of the nucleoside analogues 3'-azidothymidine and 3'-fluorothymidine inhibit reverse transcriptase and are of therapeutic interest for the treatment of retrovirus infections. At equimolar concentrations 3'-fluorothymidine was more effectively transformed to the triphosphate by human CEM cells than azidothymidine which mainly accumulates as the monophosphate. Hydroxyurea, a drug that inhibits de novo synthesis of deoxyribonucleotides, considerably increased the ability of cells to phosphorylate both analogues. Addition of as little as 50 microM hydroxyurea decreased the amount of dideoxynucleoside required to attain a given intracellular concentration of its triphosphate by an order of magnitude. Hydroxyurea is known to shift the balance of substrate cycles between natural deoxynucleosides and their 5'-phosphates in the direction of synthesis and thereby to increase the import and intracellular phosphorylation of the nucleoside. The present results demonstrate a similar effect for the two analogues and raise the possibility of using this effect in therapy.

Interaction of the 5'-phosphates of the anti-HIV agents, 3'-azido-3'-deoxythymidine and 3'-azido-2',3'-dideoxyuridine, with thymidylate synthase

A study has been made of the interaction of 3'-azido-3'-deoxythymidine 5'-phosphate (AZTMP) and 3'-azido-2',3'-dideoxy-uridine 5'-phosphate (AZdUMP) with thymidylate synthase. With the enzyme from L1210 cells and the tapeworm Hymenolepis diminuta, AZTMP was a weak inhibitor competitive with respect to dUMP (Ki = 6.3 mM and 0.5 mM); hence cytotoxicity of AZT, in cells in which accumulation of AZTMP is not high, is not due to inhibition of cellular thymidylate synthase. AZdUMP, with the L1210 enzyme, was a weak substrate (competition with dUMP described by apparent Ki = 4.7 mM), excluding conversion of AZdUMP to AZTMP as a source of toxicity of 3'-azido-2',3'-dideoxyuridine. An efficient procedure is described for enzymatic phosphorylation on a preparative scale of dideoxynucleosides.

Effects of 3'-azido-3'-deoxythymidine on the deoxynucleotide triphosphate pools of cultured human cells

The effects of 3'-azido-3'-deoxythymidine (AZT) on the deoxynucleotide pools of three human cell lines, HL-60, H-9, and K-562, were determined. The corresponding ED50s for inhibition of cell growth were 670, 100, and 100 microM AZT. In all three lines, exposure to 200 microM AZT caused dTTP and dGTP initially to fall and then to return towards control levels. In contrast to a previous report [Furman et al., (1986) Proc. Nat. Acad. Sci. USA 83, 8333-8337], dCTP levels increased. Pools of dATP were relatively unchanged. Qualitatively similar changes occurred in 10 microM AZT, but recovery was faster than at 200 microM AZT. After 24 hrs incubation with 200 microM AZT, AZT-5'-MP reached 2.8, 4.7, and 15.7 mM in the HL-60, H-9, and K-562 cells, respectively. When HL-60 and K-562 cells incubated in AZT were resuspended in fresh medium, AZT-5'-MP pools declined with respective t1/2 values equal to 34 and 68 min. The concentration of thymidine, and to a lesser extent deoxyuridine, increased in the media of treated cells. AZT-5'-MP was found in the media of cells treated with AZT.

Phosphorylation, anti-HIV activity and cytotoxicity of 3'-fluorothymidine

3'-fluorothymidine (FddThd) is well phosphorylated to the 5'-triphosphate in various relevant cell-lines. This results in fairly stable levels of this compound without accumulation of the 5'-monophosphate to the extent described for 3'-azidothymidine (AzT). The di- and triphosphate of FddThd seems unable to influence the ribonucleotide reductase in permeable cells. FddThd protects 50% of the MT-4 cells against the cytopathic effect of HIV at 3 nM, but reduces the number of uninfected viable cells to 50% at 1.1 microM. All other tested human cell-lines displayed a far less antiproliferative sensitivity to FddThd, comparable with that of AzT.