Pentoxifylline blocks hepatic stellate cell activation independently of phosphodiesterase inhibitory activity.

Activated, but not quiescent, hepatic stellate cells (lipocytes) have a high level of collagen type I and smooth muscle actin (SMA) gene expression. Therefore, stellate cell activation is a critical step in hepatic fibrosis. The mechanisms leading to stellate cell activation in vivo are unknown. The characteristic hepatic oxidative stress cascade induced in rats by CCl4 markedly stimulated stellate cell entry into S phase, nuclear factor (NF)-kappa B activity, and c-myb expression. These changes were prevented by pentoxifylline, which also decreased CCl4-induced hepatic injury. As expected, cAMP-mediated phosphorylation of CREB-Ser133 was induced in vivo in stellate cells by pentoxifylline but not by its metabolite 5, an N-1 carboxypropyl derivative, which lacks phosphodiesterase inhibitory activity. Stellate cell nuclear extracts from CCl4-treated, but not from control, animals formed a complex with the critical promoter E box of the alpha-SMA gene, which was disrupted by c-myb antibodies and competed with by c-myb cognate DNA. Treatment with pentoxifylline or metabolite 5 prevented the molecular abnormalities characteristic of stellate cell activation induced by CCl4. These results suggest that induction of c-myb plays an important role in the in vivo activation of stellate cells. Pentoxifylline blocks stellate cell activation in vivo independently of its inhibitory effects on phosphodiesterases by interfering with the oxidative stress cascade and the activation of NF-kappa B and c-myb.

Substituted xanthines, pteridinediones, and related compounds as potential antiinflammatory agents. Synthesis and biological evaluation of inhibitors of tumor necrosis factor alpha.

A series of analogues of pentoxifylline metabolites were prepared in the purine, pteridine, [1,2,5]-thiadiazolo[3,4-d]pyrimidine, and quinazoline ring systems and evaluated for their ability to inhibit the production of tumor necrosis factor-alpha (TNF alpha) in human peripheral blood monocytes stimulated with bacterial lipopolysaccharide (LPS). The more active compounds were also tested for inhibition of cyclic AMP phosphodiesterase type IV (PDE IV) from human neutrophils in order to help determine their mechanism of action. Selected compounds which showed good activity in the in vitro TNF alpha assay were evaluated in an in vivo LPS-induced leukopenia model in mice. The most potent compounds in the TNF alpha assay, 6, 31, and 58, inhibited TNF alpha production at an IC50 of approximately 5 microM for each. Compound 58 was a very poor inhibitor of PDE IV but was the most active at preventing the leukopenia induced by TNF alpha in mice, providing more than 60% protection at 50 mg/kg. Thus, compounds such as 58, which are good inhibitors of TNF alpha production but are devoid of PDE IV inhibitory properties, may have potential as new antiinflammatory agents.

New adenosine kinase inhibitors with oral antiinflammatory activity: synthesis and biological evaluation.

Several 5-iodotubercidin analogues in the pyrazolo[3,4-d]pyrimidine ring system were synthesized as potential inhibitors of adenosine kinase by a direct Lewis acid-catalyzed glycosylation procedure using both the preformed carbohydrate and the heterocyclic base as starting materials. The 5'-hydroxyl, -chloro, -azido, -deoxy, -amino, and -fluoro derivatives were prepared and evaluated in three systems for biological activity relative to adenosine, the true substrate, and 5-iodotubercidin, a known inhibitor. First, each compound was studied kinetically for inhibition of purified human placental adenosine kinase activity. The order of potency was: iodotubercidin > hydroxyl > amino > or = deoxy > fluoro > chloro >> azido. The Ki values for the 5'-hydroxyl and 5'-amino compounds, the two most potent inhibitors, were 80 and 150 nM, respectively. The inhibition appeared to be essentially competitive in nature, although a noncompetitive component of significance for the more potent inhibitors cannot be ruled out. Second, a bioassay was conducted in which the toxicity of 6-methylmercaptopurine riboside toward human CEM lymphoblasts was reversed by varying concentrations of the compounds. The order of effectiveness of the compounds in this system, representing a functional inhibition of adenosine kinase in cultured cells, was about the same as that with the purified enzyme, except that the 5'-chloro and 5'-fluoro compounds were ineffective. Third, the 5'-hydroxyl derivative was evaluated in vivo in a rat pleurisy inflammation model and displayed biological activity at a dose of 30 mg/kg given orally. Finally, the in vitro toxicity of each compound was assessed in CEM lymphoblasts. Results indicated that the two most potent inhibitors in the pyrazolo[3,4-d]pyrimidine ring system, the 5'-hydroxyl (7) and the 5'-amino (20), were 15-fold and 75-fold, respectively, less growth inhibitory than 5-iodotubercidin.

Oral antilymphocyte activity and induction of apoptosis by 2-chloro-2'-arabino-fluoro-2'-deoxyadenosine.

2-Chlorodeoxyadenosine (CdA) is active in chronic lymphocytic leukemia, hairy-cell leukemia, and low-grade lymphomas. In part, this spectrum of activity may be attributable to the selective toxicity of CdA to nondividing lymphocytes and monocytes. However, CdA is unstable at acidic pH and is degraded by bacterial nucleoside phosphorylases. The present experiments demonstrate that the 2'-arabino-fluoro derivative of CdA, designated CAFdA, is also directly toxic to quiescent lymphocytes and macrophages. Unlike CdA, CAFdA was stable at pH 2 and resisted degradation by Escherichia coli nucleoside phosphorylase. Cell killing was preceded by the formation of DNA strand breaks and could be prevented by supplementation of the medium with deoxycytidine. The initial DNA damage initiated the pattern of oligonucleosomal DNA fragmentation characteristic of apoptosis. Mutant lymphoblasts, deficient in deoxycytidine kinase, with elevated cytoplasmic 5'-nucleotidase, or with expanded deoxynucleotide pools secondary to increased ribonucleotide reductase activity, were cross-resistant to both CAFdA and CdA toxicity. One-week oral treatment with CAFdA (1 mg/ml in drinking water) achieved an average plasma concentration of 0.56 microM and eliminated 90% of chronic lymphocytic leukemia cells transplanted into severe combined immunodeficiency (scid) mice. Under the same conditions, CdA was much less active. Collectively, these results suggest that CAFdA could be effective as an oral agent in indolent lymphoproliferative diseases and in autoimmune diseases where lymphocyte and monocyte depletion is desirable.

Dependence of cell survival on DNA repair in human mononuclear phagocytes.

Mononuclear phagocytes play a central role in the pathogenesis of chronic inflammatory diseases. It is therefore important to define chemotherapeutically exploitable metabolic pathways that distinguish monocytes from other cell types. Blood monocytes do not synthesize deoxynucleotides de novo, and their transformation to macrophages occurs without cell division. Whether or not monocytes can repair DNA damage, and whether or not DNA repair is necessary for their survival, is unknown. The present experiments demonstrate that normal human monocytes, unlike neutrophils, rapidly repair DNA strand breaks induced by gamma-irradiation. Monocyte extracts contain functional immunoreactive DNA polymerase-alpha. DNA repair synthesis in normal monocytes is blocked by aphidicolin, an inhibitor of DNA polymerase-alpha with respect to dCTP. Aphidicolin is also directly toxic to normal monocytes, but has no effect on nondividing lymphocytes or fibroblasts. Compared to most other cell types, monocytes and macrophages have very low dCTP pools, but abundant deoxycytidine kinase activity. This suggests that dCTP derived from salvage pathways is important for DNA repair in these cells. Consistent with this notion, exogenous deoxycytidine could partially protect monocytes from aphidicolin killing. The unexpected toxicity of aphidicolin toward normal human monocytes may be attributable to their high rate of spontaneous DNA strand break formation, to the importance of DNA polymerase-alpha for DNA repair in these cells, and to their minute dCTP pools.

Deoxyadenosine-resistant human T lymphoblasts with elevated 5'-nucleotidase activity.

Although several different enzymes with 5'-nucleotidase activity have been described in mammalian cells, their functions in nucleotide metabolism have not been clearly distinguished. In the present experiments, a mutant human T lymphoblastoid cell line (CEM-dAdoR) was selected specifically for resistance to deoxyadenosine toxicity. Compared to parental CEM cells, the variant had 4-fold elevated ATP-activated cytosolic 5'-nucleotidase activity. Other enzymes of potential importance for deoxyadenosine metabolism were indistinguishable in the two cell types. In medium supplemented with the adenosine deaminase inhibitor deoxycoformycin, the T cells with increased 5'-nucleotidase accumulated less nucleotides from exogenously added deoxyadenosine, or 9-beta-D-arabinofuranosyladenine, than did parental T lymphocytes. These metabolic changes were associated with resistance to the growth inhibitory effects of these nucleosides, and also to deoxyguanosine and to 9-beta-D-arabinofuranosylguanine. The T cells with elevated 5'-nucleotidase activity formed more 2',3'-dideoxyadenosine than did parental cells, in deoxycoformycin-supplemented medium. The accumulation of 2',3'-dideoxyadenosine 5'-triphosphate from 2',3'-dideoxyinosine was similarly augmented in the mutant. These data establish the importance of the cytosolic 5'-nucleotidase for the metabolism of purine 2'-deoxyribonucleosides, arabinonucleosides and 2',3'-dideoxyribonucleosides in T lymphoblasts.

Utilization of 2'-deoxynad for ADP-ribose transfer reactions.

2'-deoxyNAD was examined as a substrate for both mono(ADP-ribosyl)ation and poly(ADP-ribosyl)ation reactions. 2'-deoxyNAD is a substrate for the diphtheria toxin-catalyzed mono(ADP-ribosyl)ation of elongation factor-2, inactivating its function to enhance protein synthesis. On the other hand, 2'-deoxyNAD is a poor substrate for poly(ADP-ribose) polymerase. 2'-deoxyNAD was not synthesized intracellularly from deoxyATP, even when deoxyATP content was markedly increased by incubation of cells with deoxyadenosine and an adenosine deaminase inhibitor. 2'-deoxyNAD, because of its specificity, could be a quite useful reagent for the investigation of cellular mono(ADP-ribosyl)ation reactions.

Synthesis of 2',3'-dideoxynucleosides by enzymatic trans-glycosylation.

Recently, several pyrimidine and purine 2',3'-dideoxynucleosides have been shown to inhibit the replication of the human immunodeficiency virus-1 (HIV), the causative agent of the acquired immune deficiency syndrome (AIDS). These compounds are usually prepared by reduction of the corresponding 2'-deoxynucleosides. The present experiments demonstrate that 2',3'-dideoxynucleosides can also be made by enzymatic trans-glycosylation, using the trans-N-deoxyribosylase from Lactobacillus helveticus. The broad specificity of this enzyme makes it possible to synthesize for metabolic studies radiochemically pure 2',3'-dideoxynucleosides, using diverse purine and pyrimidine base acceptors.

Metabolism and anti-human immunodeficiency virus-1 activity of 2-halo-2',3'-dideoxyadenosine derivatives.

Both 2',3'-dideoxyadenosine and 2',3'-dideoxyinosine have been shown (Mitsuya, H., and Broder, S. (1987) Nature 325, 773-778) to have in vitro activity against the human immunodeficiency virus-1 (HIV). However, these dideoxynucleosides may be catabolized by human T cells, even when adenosine deaminase is inhibited by deoxycoformycin. To overcome this problem, we have synthesized the 2-fluoro-, 2-chloro-, and 2-bromo-derivatives of 2',3'-dideoxyadenosine. The metabolism and anti-HIV activity of the 2-halo-2',3'-dideoxyadenosine derivatives and of 2',3'-dideoxyadenosine were compared. The 2-halo-2',3'-dideoxyadenosine derivatives were not deaminated significantly by cultured CEM T lymphoblasts. Experiments with 2-chloro-2',3'-dideoxyadenosine showed that the T cells converted the dideoxynucleoside to the 5'-monophosphate, 5'-diphosphate, and 5'-triphosphate metabolites. At concentrations lower than those producing cytotoxicity in uninfected cells (3-10 microM), the 2-halo-2',3-dideoxyadenosine derivatives inhibited the cytopathic effects of HIV toward MT-2 T lymphoblasts, and retarded viral replication in CEM T lymphoblasts. Experiments with a deoxycytidine kinase-deficient mutant CEM T cell line showed that this enzyme was necessary for the phosphorylation and anti-HIV activity of the 2-chloro-2',3'-dideoxyadenosine. In contrast, 2',3'-dideoxyadenosine was phosphorylated by the deoxycytidine kinase-deficient mutant and retained anti-HIV activity in this cell line. Thus, the 2-halo derivatives of 2',3'-dideoxyadenosine, in contrast to 2',3'-dideoxyadenosine itself, are not catabolized by T cells. Their anti-HIV and anti-proliferative activities are manifest only in cells expressing deoxycytidine kinase. The in vivo implications of these results for anti-HIV chemotherapy are discussed.

Biochemical genetic analysis of 2',3'-dideoxyadenosine metabolism in human T lymphocytes.

2',3'-dideoxyadenosine (ddAdo) has been shown to inhibit the infection of cultured human T lymphoblasts with the human immunodeficiency virus-1 (HIV-1). However, the pathways of ddAdo metabolism in T lymphocytes have not been well defined. We have studied the uptake and degradation of ddAdo in human CEM T lymphoblasts, in mutant CEM T cells deficient in adenosine kinase or deoxycytidine kinase, and in normal lymphocytes and monocytes. The results indicate that ddAdo may be phosphorylated in T cells by several different enzymes, although deoxycytidine kinase predominates. However, 99% of the ddAMP formed is deaminated by AMP deaminase and subsequently dephosphorylated. Thus, the ability of ddAdo to prevent HIV-1 infection may be limited in cells with high AMP deaminase activity.

Characterization of human poly(ADP-ribose) polymerase with autoantibodies.

The addition of poly(ADP-ribose) chains to nuclear proteins has been reported to affect DNA repair and DNA synthesis in mammalian cells. The enzyme that mediates this reaction, poly(ADP-ribose) polymerase, requires DNA for catalytic activity and is activated by DNA with strand breaks. Because the catalytic activity of poly(ADP-ribose) polymerase does not necessarily reflect enzyme quantity, little is known about the total cellular poly(ADP-ribose) polymerase content and the rate of its synthesis and degradation. In the present experiments, specific human autoantibodies to poly(ADP-ribose) polymerase and a sensitive immunoblotting technique were used to determine the cellular content of poly(ADP-ribose) polymerase in human lymphocytes. Resting peripheral blood lymphocytes contained 0.5 X 10(6) enzyme copies per cell. After stimulation of the cells by phytohemagglutinin, the poly(ADP-ribose) polymerase content increased before DNA synthesis. During balanced growth, the T lymphoblastoid cell line CEM contained approximately 2 X 10(6) poly(ADP-ribose) polymerase molecules per cell. This value did not vary by more than 2-fold during the cell growth cycle. Similarly, mRNA encoding poly(ADP-ribose) polymerase was detectable throughout S phase. Poly(ADP-ribose) polymerase turned over at a rate equivalent to the average of total cellular proteins. Neither the cellular content nor the turnover rate of poly(ADP-ribose) polymerase changed after the introduction of DNA strand breaks by gamma irradiation. These results show that in lymphoblasts poly(ADP-ribose) polymerase is an abundant nuclear protein that turns over relatively slowly and suggest that most of the enzyme may exist in a catalytically inactive state.

Programmed cell death and adenine deoxynucleotide metabolism in human lymphocytes.

Agents that cause the accumulation of DNA strand breaks are directly cytotoxic to non-dividing normal human peripheral blood lymphocytes, and to chronic lymphocytic leukemia (CLL) cells. Activation of poly(ADP-ribose) polymerase (ADPRP), and the resultant consumption of NAD, play an essential role in mediating the toxicity of these agents. Human peripheral blood lymphocytes contain a substantial number of alkali-sensitive DNA sites, reflecting ongoing DNA strand breakage and repair. However, resting lymphocytes have a limited capacity to synthesize NAD. Pulse-chase experiments indicate that approximately 75% of their NAD turnover is due to ADPRP activity. Exposure of the cells in vitro to deoxyadenosine, or to 2-chlorodeoxyadenosine (CdA, an adenosine deaminase resistant deoxyadenosine congener), caused an increase in DNA strand breaks, rapid NAD consumption, ATP depletion and cell death. Supplementation of the medium with inhibitors of poly(ADP-ribose) polymerase blocks the fall in cellular NAD and ATP, and protects the lymphocytes from the toxicity of DNA damaging agents. Slowly dividing malignant lymphocytes from patients with CLL are also susceptible to lethal NAD depletion following DNA damage. 2-chlorodeoxyadenosine (CdA) induced massive DNA strand break formation in CLL cells in vitro and a fall in NAD and ATP pools. In an initial clinical trial, several CLL patients, and two patients with hairy cell leukemia, have responded to treatment with CdA, with minimal toxicity. Thus, the suicidal activation of ADPRP in response to DNA damage has been rationally exploited in the treatment of chronic lymphoid malignancies.

Pyridine nucleotide cycling and poly(ADP-ribose) synthesis in resting human lymphocytes.

NAD is a critical cofactor for the oxidation of fuel molecules. The exposure of human PBL to agents that cause DNA strand breaks to accumulate can deplete NAD pools by increasing NAD consumption for poly(ADP-ribose) formation. However, the pathways of NAD synthesis and degradation in viable PBL have not been carefully documented. The present experiments have used radioactive labeling techniques to trace the routes of NAD metabolism in resting PBL. The cells could generate NAD from either nicotinamide or nicotinic acid. PBL incubated with [14C]nicotinic acid excreted [14C]nicotinamide into the medium. Approximately 50% of a prelabeled [14C]NAD pool was metabolized during 6 to 8 hr in tissue culture. Basal NAD turnover was prolonged threefold to fourfold by 3-aminobenzamide (3-ABA), an inhibitor of poly(ADP-ribose) synthetase. Supplementation of the medium with 3-ABA also prevented the accelerated NAD degradation that ensued after exposure of PBL to deoxyadenosine plus deoxycoformycin at concentrations previously shown to cause DNA strand break accumulation. These results demonstrate that quiescent human PBL continually produce NAD and utilize the nucleotide for poly(ADP-ribose) synthesis.

DNA strand breaks, NAD metabolism, and programmed cell death.

An intimate relationship exists between DNA single-strand breaks, NAD metabolism, and cell viability in quiescent human lymphocytes. Under steady-state conditions, resting lymphocytes continually break and rejoin DNA. The balanced DNA excision-repair process is accompanied by a proportional consumption of NAD for poly(ADP-ribose) synthesis. However, lymphocytes have a limited capacity to resynthesize NAD from nicotinamide. An increase in DNA strand break formation in lymphocytes, or a block in DNA repair, accelerates poly(ADP-ribose) formation and may induce lethal NAD and ATP depletion. In this way, the level of DNA single-strand breaks in the lymphocyte nucleus is linked to the metabolic activity of the cytoplasm. The programmed removal of lymphocytes (and perhaps of other cells) with damaged DNA, may represent a novel physiologic function for poly(ADP-ribose)-dependent NAD cycling.

Inhibition of DNA repair by deoxyadenosine in resting human lymphocytes.

Profound lymphopenia is characteristic of immunodeficient children who lack adenosine deaminase (ADA). When ADA is inactive, deoxyadenosine (dAdo) is phosphorylated by immature T lymphoblasts and inhibits cell division. However, dAdo also causes the slow accumulation of DNA strand breaks in nondividing, mature human peripheral blood lymphocytes. To explore the basis for this phenomenon, we have assessed the effects of dAdo and other deoxynucleosides on the repair of gamma-radiation induced DNA strand breaks in resting normal lymphocyte cultures. As measured by a sensitive DNA unwinding assay, most DNA strand breaks were rejoined within 2 hr after exposure of lymphocytes to 500 rad. In medium supplemented with deoxycoformycin, a tight binding ADA inhibitor, dAdo retarded DNA rejoining in a dose and time dependent manner. The inhibition required dAdo phosphorylation. Over an 8-hr period, 10 microM dAdo gradually rendered peripheral blood lymphocytes incompetent for DNA repair. Among several other compounds tested, 2-chlorodeoxyadenosine, an ADA resistant dAdo congener with anti-leukemic and immunosuppressive activity, was the most powerful inhibitor of DNA repair, exerting significant activity at concentrations as low as 100 nM. Both dAdo and 2-chlorodeoxyadenosine blocked unscheduled DNA synthesis in irradiated resting lymphocytes, as measured by [3H]thymidine uptake. On the basis of this and other data, we suggest that quiescent peripheral blood lymphocytes break and rejoin DNA at a slow and balanced rate. The accumulation of dATP progressively retards the DNA repair process and thereby fosters the time-dependent accretion of DNA strand breaks. By inhibiting DNA repair, dAdo, 2-chlorodeoxyadenosine and related compounds may substantially potentiate the toxicity of DNA damaging agents to normal and malignant lymphocytes.

Lymphocyte dysfunction after DNA damage by toxic oxygen species. A model of immunodeficiency.

The metabolic causes for immune impairment in patients with severe chronic inflammatory diseases have not been clearly defined. Recently, the overproduction of poly(ADP-ribose) in resting lymphocytes with unrepaired DNA strand breaks has been suggested to contribute to immune dysfunction in adenosine deaminase-deficient patients. Our experiments have determined to what extent DNA damage and poly(ADP-ribose) synthesis might also explain the impaired mitogen responsiveness of PBL exposed to toxic oxygen species. Treatment of normal resting human lymphocytes with xanthine oxidase and hypoxanthine dose-dependently induced DNA strand breaks and triggered the rapid synthesis of poly(ADP-ribose). Subsequently, NAD+ and ATP pools decreased precipitously. Lymphocytes exposed previously to the enzymatic oxidizing system did not synthesize DNA after stimulation with PHA. However, if the medium was supplemented with 3-aminobenzamide or nicotinamide, two compounds that inhibit poly(ADP-ribose) formation, cellular NAD+ and ATP pools were preserved, and the lymphocytes responded vigorously to a mitogenic challenge. Excessive poly(ADP-ribose) synthesis, provoked by DNA strand breakage, may represent a common pathway that connects the immunodeficiency syndromes associated with (a) exposure of lymphocytes to toxic oxygen species during chronic inflammatory states, (b) adenosine deaminase deficiency, and (c) certain DNA repair disorders.