Inactivation of the Poly(ADP-ribose) Polymerase Gene Affects Oxygen Radical and Nitric Oxide Toxicity in Islet Cells
1995; Elsevier BV; Volume: 270; Issue: 19 Linguagem: Inglês
10.1074/jbc.270.19.11176
ISSN1083-351X
AutoresBirgit Heller, Zhao‐Qi Wang, Erwin F. Wagner, Jürgen Radons, Alexander Bürkle, Karin Fehsel, Volker Burkart, Hubert Kolb,
Tópico(s)Sirtuins and Resveratrol in Medicine
ResumoActivation of the nuclear enzyme poly(ADP-ribose) polymerase (PARP) is an early response of cells exposed to DNA-damaging compounds such as nitric oxide (NO) or reactive oxygen intermediates (ROI). Excessive poly-(ADP-ribose) formation by PARP has been assumed to deplete cellular NAD+ pools and to induce the death of several cell types, including the loss of insulin-producing islet cells in type I diabetes. In the present study we used cells from mice with a disrupted and thus inactivated PARP gene to provide direct evidence for a causal relationship between PARP activation, NAD+ depletion, and cell death. We found that mutant islet cells do not show NAD+ depletion after exposure to DNA-damaging radicals and are more resistant to the toxicity of both NO and ROI. These findings directly prove that PARP activation is responsible for most of the loss of NAD+ following such treatment. The ADP-ribosylation inhibitor 3-aminobenzamide partially protected islet cells with intact PARP gene but not mutant cells from lysis following either NO or ROI treatment. Hence the protective action of 3-aminobenzamide must be due to inhibition of PARP and does not result from its other pharmacological properties such as oxygen radical scavenging. Finally, by the use of mutant cells an alternative pathway of cell death was discovered which does not require PARP activation and NAD+ depletion. In conclusion, the data prove the causal relationship of PARP activation and subsequent islet cell death and demonstrate the existence of an alternative pathway of cell death independent of PARP activation and NAD+ depletion. Activation of the nuclear enzyme poly(ADP-ribose) polymerase (PARP) is an early response of cells exposed to DNA-damaging compounds such as nitric oxide (NO) or reactive oxygen intermediates (ROI). Excessive poly-(ADP-ribose) formation by PARP has been assumed to deplete cellular NAD+ pools and to induce the death of several cell types, including the loss of insulin-producing islet cells in type I diabetes. In the present study we used cells from mice with a disrupted and thus inactivated PARP gene to provide direct evidence for a causal relationship between PARP activation, NAD+ depletion, and cell death. We found that mutant islet cells do not show NAD+ depletion after exposure to DNA-damaging radicals and are more resistant to the toxicity of both NO and ROI. These findings directly prove that PARP activation is responsible for most of the loss of NAD+ following such treatment. The ADP-ribosylation inhibitor 3-aminobenzamide partially protected islet cells with intact PARP gene but not mutant cells from lysis following either NO or ROI treatment. Hence the protective action of 3-aminobenzamide must be due to inhibition of PARP and does not result from its other pharmacological properties such as oxygen radical scavenging. Finally, by the use of mutant cells an alternative pathway of cell death was discovered which does not require PARP activation and NAD+ depletion. In conclusion, the data prove the causal relationship of PARP activation and subsequent islet cell death and demonstrate the existence of an alternative pathway of cell death independent of PARP activation and NAD+ depletion. Pancreatic islet cells are highly susceptible to the toxic effects of reactive oxygen intermediates (ROI) 1The abbreviations used are: ROIreactive oxygen intermediatesNOnitric oxidePARPpoly(ADP-ribose) polymerase3-AB3-aminobenzamideNPnitroprusside and nitric oxide (NO) (1Grankvist K. Marklund S.L. Talledal I.B. Biochem. J. 1981; 199: 393-398Crossref PubMed Scopus (429) Google Scholar, 2Malaisse W.J. Malaisse-Lagae F. Sener A. Pipeleers D.G. Proc. Natl. Acad. Sci. U. S. A. 1982; 79: 927-930Crossref PubMed Scopus (273) Google Scholar, 3Rabinovitch A. J. Lab. Clin. Med. 1992; 119: 455-456PubMed Google Scholar). Evidence is accumulating that these reactive mediators are released during prediabetic islet inflammation and thus may contribute to beta cell loss (4Wu G. Marliss E.B. Diabetes. 1993; 42: 520-529Crossref PubMed Scopus (21) Google Scholar, 5Corbett J.A. Mikhael A. Shimizu J. Frederick K. Misko T.P. McDaniel M.L. Kanagawa O. Unanue E.R. Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 8992-8995Crossref PubMed Scopus (151) Google Scholar, 6Kleemann R. Rothe H. Kolb-Bachofen V. Xie Q.-W. Nathan C. Martin S. Kolb H. FEBS Lett. 1993; 328: 9-12Crossref PubMed Scopus (132) Google Scholar). The inactivation of mitochondrial electron transfer, of aconitase and glyceraldehyde-3-phosphate dehydrogenase have been described as major mechanisms of NO toxicity in islet cells (7Welsh N. Sandler S. Biochem. Biophys. Res. Commun. 1992; 182: 333-340Crossref PubMed Scopus (84) Google Scholar, 8Molina y Vedia L. McDonald B. Reep B. Brüne B. Di Silvio M. Billiar T.R. Lapetina E.G. J. Biol. Chem. 1992; 267: 24929-24932Abstract Full Text PDF PubMed Google Scholar). However, recent evidence for NO-induced DNA damage in islet cells has led to the proposal that excessive poly(ADP-ribose) polymerase (PARP) activation and the consequent depletion of its substrate, NAD+, might be responsible for cell death (9Radons J. Heller B. Bürkle A. Hartmann B. Rodriguez M.-L. Kröncke K.-D. Burkart V. Kolb H. Biochem. Biophys. Res. Commun. 1994; 199: 1270-1277Crossref PubMed Scopus (201) Google Scholar, 10Kallmann B. Burkart V. Kröncke K.-D. Kolb-Bachofen V. Kolb H. Life Sci. 1992; 51: 671-678Crossref PubMed Scopus (115) Google Scholar, 11Fehsel K. Jalowy A. Sun Q. Burkart V. Hartmann B. Kolb H. Diabetes. 1993; 42: 496-500Crossref PubMed Scopus (198) Google Scholar). PARP is an abundant nuclear enzyme which binds to DNA strand breaks, resulting in its catalytic activation (for reviews, see Refs. 12Althaus F.R. Richter R. Mol. Biol. Biochem. Biophys. 1987; 37: 1-126PubMed Google Scholar, 13de Murcia G. Ménissier-de Murcia J. Trends Biochem. Sci. 1994; 19: 172-176Abstract Full Text PDF PubMed Scopus (763) Google Scholar, 14Berger N.A. Radiat. Res. 1985; 101: 4-15Crossref PubMed Scopus (681) Google Scholar). This pathway has recently also been held responsible for NO-induced neurotoxicity (15Zhang J. Dawson V.L. Dawson T.M. Snyder S.H. Science. 1994; 263: 687-689Crossref PubMed Scopus (1085) Google Scholar). DNA damage, PARP activation, and NAD+ depletion are also seen in islets exposed to ROI (16Heller B. Bürkle A. Radons J. Fengler E. Jalowy A. Müller M. Burkart V. Kolb H. Biol. Chem. Hoppe-Seyler. 1994; 375: 597-602Crossref PubMed Scopus (43) Google Scholar), and the same pathway has been proposed previously to account for toxic effects of the alkylating agent streptozotocin (17Yamamoto H. Uchigata Y. Okamoto H. Nature. 1981; 294: 284-286Crossref PubMed Scopus (477) Google Scholar). By using NAD+ as substrate for poly(ADP-ribosylation) of nuclear proteins, including PARP itself, PARP is thought to play an important role in DNA repair at low levels of DNA damage. On the other hand, excessive poly(ADP-ribose) formation has been assumed to be the reason for the loss of cellular NAD+ in cells exposed to high doses of DNA damaging compounds (17Yamamoto H. Uchigata Y. Okamoto H. Nature. 1981; 294: 284-286Crossref PubMed Scopus (477) Google Scholar, 18Sims J.L. Berger S.J. Berger N.A. Biochemistry. 1983; 22: 5188-5194Crossref PubMed Scopus (169) Google Scholar). This led to the hypothesis elaborated by Berger and Okamoto that excessive PARP activation may lead to cell death because of NAD+ depletion (17Yamamoto H. Uchigata Y. Okamoto H. Nature. 1981; 294: 284-286Crossref PubMed Scopus (477) Google Scholar, 18Sims J.L. Berger S.J. Berger N.A. Biochemistry. 1983; 22: 5188-5194Crossref PubMed Scopus (169) Google Scholar). reactive oxygen intermediates nitric oxide poly(ADP-ribose) polymerase 3-aminobenzamide nitroprusside The assumption of a causal relationship between PARP activation and cell death is based on the observation that PARP inhibitors, such as nicotinamide or 3-aminobenzamide (3-AB), partially prevent ROI or NO toxicity (9Radons J. Heller B. Bürkle A. Hartmann B. Rodriguez M.-L. Kröncke K.-D. Burkart V. Kolb H. Biochem. Biophys. Res. Commun. 1994; 199: 1270-1277Crossref PubMed Scopus (201) Google Scholar, 10Kallmann B. Burkart V. Kröncke K.-D. Kolb-Bachofen V. Kolb H. Life Sci. 1992; 51: 671-678Crossref PubMed Scopus (115) Google Scholar, 11Fehsel K. Jalowy A. Sun Q. Burkart V. Hartmann B. Kolb H. Diabetes. 1993; 42: 496-500Crossref PubMed Scopus (198) Google Scholar, 12Althaus F.R. Richter R. Mol. Biol. Biochem. Biophys. 1987; 37: 1-126PubMed Google Scholar, 13de Murcia G. Ménissier-de Murcia J. Trends Biochem. Sci. 1994; 19: 172-176Abstract Full Text PDF PubMed Scopus (763) Google Scholar, 14Berger N.A. Radiat. Res. 1985; 101: 4-15Crossref PubMed Scopus (681) Google Scholar, 15Zhang J. Dawson V.L. Dawson T.M. Snyder S.H. Science. 1994; 263: 687-689Crossref PubMed Scopus (1085) Google Scholar, 16Heller B. Bürkle A. Radons J. Fengler E. Jalowy A. Müller M. Burkart V. Kolb H. Biol. Chem. Hoppe-Seyler. 1994; 375: 597-602Crossref PubMed Scopus (43) Google Scholar, 17Yamamoto H. Uchigata Y. Okamoto H. Nature. 1981; 294: 284-286Crossref PubMed Scopus (477) Google Scholar, 18Sims J.L. Berger S.J. Berger N.A. Biochemistry. 1983; 22: 5188-5194Crossref PubMed Scopus (169) Google Scholar, 19Carson D.A. Seto S. Wasson D.B. Carrera C.J. Exp. Cell Res. 1986; 164: 273-281Crossref PubMed Scopus (287) Google Scholar, 20Cleaver J.E. Morgan W.F. Mutat. Res. 1991; 257: 1-18Crossref PubMed Scopus (131) Google Scholar). Unfortunately, these compounds display several other pharmacological activities such as inhibiting mono(ADP-ribosyltransferases) (albeit to a lesser degree) or nicotinamide methyltransferase, interference with transmembrane signaling or direct interactions with oxygen radicals (20Cleaver J.E. Morgan W.F. Mutat. Res. 1991; 257: 1-18Crossref PubMed Scopus (131) Google Scholar, 21Rankin P.W. Jacobson E.L. Benjamin R.C. Moss J. Jacobson M.K. J. Biol. Chem. 1989; 264: 4312-4317Abstract Full Text PDF PubMed Google Scholar, 22Johnson G.S. Biochem. Int. 1981; 2: 611-617Google Scholar, 23Nowicki M. Landon C. Sugawara S. Dennert G. Cell. Immunol. 1991; 132: 115-126Crossref PubMed Scopus (20) Google Scholar). The present study was designed to test the hypothesis of a causal relationship between PARP activation, loss of NAD+, and cell death. To exclude possible side effects resulting from pharmacological inhibitors of PARP, the experiments were performed on cells from mice with an inactivated PARP gene. By the use of these cells we were able to demonstrate that PARP activation induced by ROI- or NO-mediated DNA damage is indeed a causative factor for NAD+ depletion and cell death to occur. In addition, an alternative pathway of cell death could be identified which does not involve PARP activation and NAD+ depletion. Animals—129/BL6 mice (129/Sv × C57BL/6) with a disrupted and thus inactive PARP gene were used to investigate the role of PARP in islet cell death. The PARP gene was inactivated by homologous recombination in embryonic stem (ES) cells (derived from 129/Sv mice), which were subsequently used to generate PARP-deficient mice (PARP–/–). The procedures performed to obtain these mutant mice as well as their phenotype will be published in detail elsewhere (24Wang Z.-Q. Auer B. Stingi L. Berghammer H. Haidacher D. Schwaiger M. Wagner E. Genes & Dev. 1995; (in press)Google Scholar). The inactivation of the PARP gene was verified by analysis of genomic structure by Southern blot, by the lack of specific mRNA and protein (24Wang Z.-Q. Auer B. Stingi L. Berghammer H. Haidacher D. Schwaiger M. Wagner E. Genes & Dev. 1995; (in press)Google Scholar). PARP+/+ mice used as controls were 129/Sv × C57BL/6, thus matching the genetic background of the PARP–/– mice which were offsprings of the brothersister matings of the 129/Sv(PARP-) × C57BL/6 mice. Preparation and Culture of Murine Cells—Pancreatic islet cells were isolated from 4–6-month-old PARP–/– and PARP+/+ mice by injection of a collagenase solution in the pancreatic duct (Serva, Heidelberg, Germany, 1.5 mg/ml, 0.37 unit/ml in Hanks' balanced salt solution (HBSS, Life Technologies, Inc. EUROPE, Heidelberg, FRG). After digestion for 34 min at 37 °C the pancreatic tissue was dispersed and the islets were enriched on a Ficoll density gradient (Ficoll 400, Pharmacia GmbH, Freiburg, FRG) with subsequent handpicking. Islet single cells were prepared by trypsin treatment and cultivated in modified RPMI 1640 supplemented with 25 mg/liter ampicillin, 120 mg/liter penicillin, 270 mg/liter streptomycin (Serva GmbH, Heidelberg, Germany), 1 mM sodium pyruvate, 2 mM L-glutamine, 24 mM NaHCO3, 1 mM HEPES, pH 7.3 (Serva GmbH), nonessential amino acids (Life Technologies, Inc. EUROPE) and 10% heat-inactivated fetal calf serum (Life Technologies, Inc. EUROPE) as described previously (25Appels B. Burkart V. Kantwerk-Funke G. Funda J. Kolb-Bachofen V. Kolb H. J. Immunol. 1989; 142: 3803-3808PubMed Google Scholar). Lymphocytes were isolated from spleen by straining the organ through a sieve followed by centrifugation of the resulting cell suspension. Cells were resuspended in modified RPMI 1640 and incubated in 10-ml Petri dishes for 1 h at 37 °C in order to remove adherent macrophages. For lysis of erythrocytes, nonadherent cells were resuspended in lysis buffer (154 mM NH4Cl, 10 mM KHCO3, and 0.1 mM EDTA, pH 7.2). After 5 min at room temperature spleen lymphocytes were washed and cultured for 3 h at 37 °C in modified RPMI 1640 prior to use. Exposure of Cells to Toxic Agents—Islet cells and lymphocytes were exposed to ROI which were generated during the oxidation of hypoxanthine (Sigma, Deisenhofen, FRG) by the enzyme xanthine oxidase (EC 1.1.3.22, Sigma) (26Burkart V. Koike T. Brenner H.-H. Kolb H. Diabetologia. 1992; 35: 1028-1034Crossref PubMed Scopus (70) Google Scholar). Islet cells were treated with the NO donor nitroprusside (NP) in modified RPMI 1640, and cyanide ions were scavenged by the addition of 8 units of rhodanese and 5 mM Na2S2O3 (10Kallmann B. Burkart V. Kröncke K.-D. Kolb-Bachofen V. Kolb H. Life Sci. 1992; 51: 671-678Crossref PubMed Scopus (115) Google Scholar). The ADP-ribosylation inhibitor 3-AB was added 30 min prior to the initiation of the NO or ROI treatment. Detection of DNA Damage—Cellular DNA was isolated by salt-chloroform extraction (11Fehsel K. Jalowy A. Sun Q. Burkart V. Hartmann B. Kolb H. Diabetes. 1993; 42: 496-500Crossref PubMed Scopus (198) Google Scholar). Samples of 2 μg were electrophoresed in TAE-buffer in a 2% agarose gel containing 0.1 μg/ml ethidium bromide. DNA was visualized by exposure to UV light. For calculation of the DNA fragment length we used the Marker X, purchased from Boehringer Mannheim (FRG). Poly(ADP-ribose) Polymerase Activity—Islet cells were cultivated (37 °C, 5% CO2) on coverslips (2.5 × 104/cm2) until adherence and then exposed to ROI. The activity of PARP was monitored by indirect immunofluorescence using the monoclonal antibody 10H (27Kawamitsu H. Hoshino H. Okada H. Miwa M. Momoi H. Sugimura T. Biochemistry. 1984; 23: 3771-3777Crossref PubMed Scopus (191) Google Scholar) that recognizes the product of the enzymatic reaction, poly(ADP-ribose), as described elsewhere (28Bürkle A. Chen G. Küpper J.H. Grube K. Zeller W.J. Carcinogenesis. 1993; 14: 559-561Crossref PubMed Scopus (98) Google Scholar). Immunofluorescence was evaluated by determination of the fraction of positive cells. NAD+ Determination—Cells were washed twice with 10 mM HEPES, 150 mM NaCl, pH 7.4, at the time of maximal NAD+ depletion after ROI treatment. For NAD+ determination the samples were disrupted by freezing and thawing followed by sonication three times for 10 s each. After centrifugation (16,000 × g, 10 min, 4 °C) the samples were stored at –70 °C until further use. Cellular NAD+ was determined by an enzymatic cycling method using alcohol dehydrogenase from Saccharomyces cerevisiae, as described previously (9Radons J. Heller B. Bürkle A. Hartmann B. Rodriguez M.-L. Kröncke K.-D. Burkart V. Kolb H. Biochem. Biophys. Res. Commun. 1994; 199: 1270-1277Crossref PubMed Scopus (201) Google Scholar, 29Nisselbaum J.S. Green S. Anal. Biochem. 1969; 27: 212-217Crossref PubMed Scopus (188) Google Scholar, 30Hinz M. Katsilambros N. Maier V. Schatz H. Pfeiffer E.F. FEBS Lett. 1973; 30: 225-228Crossref PubMed Scopus (47) Google Scholar). Determination of Islet Cell Lysis—Islet cells were cultivated in 96-well flat-bottom microtiter plates in modified RPMI 1640 and exposed to toxic agents for 18 h at 37 °C, 5% CO2. The percentage of NP- or ROI-induced lysis was calculated according to the formula: NP/ROI-inducedlysis%=100x(samplelysys-spontaneouslysis)(maximallysis-spontaneouslysis) Cell lysis was determined by trypan blue exclusion assay (10Kallmann B. Burkart V. Kröncke K.-D. Kolb-Bachofen V. Kolb H. Life Sci. 1992; 51: 671-678Crossref PubMed Scopus (115) Google Scholar). Earlier extensive control experiments had verified that the dye exclusion assay identifies the same percentage of lysed islet cells as found either by electron microscopy or by determining the release of intracellular radiolabel (31Kröncke K.-D. Kolb-Bachofen V. Berschick B. Burkart V. Kolb H. Biochem. Biophys. Res. Commun. 1991; 175: 752-758Crossref PubMed Scopus (255) Google Scholar). The spontaneous cell lysis never exceeded 157c of the total cell number. Formation of Poly(ADP-ribose) Polymer after DNA Damage—To analyze the relationship between cellular NAD+ levels and PARP activity, islet cells from PARP+/+ (wild type) mice were exposed to ROI for different time intervals. PARP activation occurred as evidenced at the single-cell level by the accumulation of poly(ADP-ribose) polymers in cell nuclei (Fig. 1A). The percentage of stained nuclei of PARP+/+ cells was 98% at 10 min, 3% at 30 min, and <1% at 60 min following treatment with ROI. The disappearance of polymer at 60 min is due to the cessation of synthesis but continued enzymatic degradation of the polymer (28Bürkle A. Chen G. Küpper J.H. Grube K. Zeller W.J. Carcinogenesis. 1993; 14: 559-561Crossref PubMed Scopus (98) Google Scholar). In the presence of 10 mm of the PARP inhibitor 3-AB no poly(ADP-ribose) polymer formation was detectable at any time point (data not shown). As expected, none of the ROI-treated PARP–/– (lacking the PARP gene) islet cells produced detectable poly(ADP-ribose) polymers at any time point (Fig. 1B). Control experiments verified that xanthine oxidase generated ROI had induced DNA strand breaks in both, PARP+/+ and PARP–/– cells as evidenced by massive DNA fragmentation, visualized on agarose gels (Fig. 2). The analysis of DNA damage was done at 4 h, i.e. prior to cell lysis, in order to avoid post mortem digestion of nuclear DNA by nucleases released after cell death.FIG. 2Analysis of DNA fragmentation in ROI-treated PARP+/+ or PARP–/– cells. Cells from PARP–/– and PARP+/+ mice were exposed to 5 milliunits/ml xanthine oxidase plus 0.1 mM hypoxanthine for 4 h. The DNA was extracted and visualized by gel-electrophoresis. Untreated PARP–/– (lane 1) or PARP+/+ cells (lane 4) show intact, high molecular weight DNA, whereas ROI-treated cells showed fragmentation of DNA in PARP–/– (lane 2) as well as in PARP+/+ cells (lane 5). Molecular size standards are shown in lane 3.View Large Image Figure ViewerDownload Hi-res image Download (PPT) NAD+ Depletion—ROI-treated PARP+/+ islet cells showed a rapid and severe depletion of intracellular NAD+ to almost 12% of the initial NAD+ content (Fig. 3A). In contrast NAD+ levels in PARP–/– cells remained normal (Fig. 3A), with a reduction to about 85% of the initial control level. A parallel analysis of splenic lymphocytes exposed to ROI also showed a reduction of NAD+ levels to 23% in PARP+/+ cells and a complete preservation of cellular NAD+ in PARP-deficient cells (Fig. 3B). These results demonstrate that most of the NAD+ depletion seen in ROI-treated cells is the direct consequence of PARP activity. Addition of 3-AB led to preservation of most (65–95%) of the cellular NAD+ content in ROI-treated islet cells and lymphocytes (Fig. 3). Therefore, the protective effect of 3-AB on NAD+ levels observed in wild type cells must be due to its inhibitory action on PARP. This is in agreement with our earlier observation that 3-AB does not preserve NAD+ levels by the scavenging of ROI, since the drug could not prevent the occurrence of ROI-induced DNA strand breaks (10Kallmann B. Burkart V. Kröncke K.-D. Kolb-Bachofen V. Kolb H. Life Sci. 1992; 51: 671-678Crossref PubMed Scopus (115) Google Scholar). However, even at high concentrations of NO the lysis of PARP–/– islet cells remained significantly lower than the lysis of the PARP+/+ cells. NO- and ROI-induced Islet Cell Lysis—We next investigated the possible causal relationship between PARP activation and islet cell death for NO as well as for ROI. As chemical donor of NO the drug sodium nitroprusside was used. Earlier studies had shown that sodium nitroprusside-generated NO mimicks macrophage-generated NO with respect to the kinetics of islet cell lysis, the occurrence of DNA damage, and the protection from NO toxicity afforded by nicotinamide (10Kallmann B. Burkart V. Kröncke K.-D. Kolb-Bachofen V. Kolb H. Life Sci. 1992; 51: 671-678Crossref PubMed Scopus (115) Google Scholar, 11Fehsel K. Jalowy A. Sun Q. Burkart V. Hartmann B. Kolb H. Diabetes. 1993; 42: 496-500Crossref PubMed Scopus (198) Google Scholar, 25Appels B. Burkart V. Kantwerk-Funke G. Funda J. Kolb-Bachofen V. Kolb H. J. Immunol. 1989; 142: 3803-3808PubMed Google Scholar). When islet cells were exposed to increasing concentrations of NP for 18 h, the cells of PARP–/– mice showed resistance to low levels of NO. Up to a concentration of 0.6 mM of NP less than 10% of the PARP–/– islet cells were lysed. In contrast, PARP+/+ islet cells responded with 40–60% lysis when exposed to concentrations of 0.4–0.6 mM NP (Fig. 4A). Control experiments verified that sodium nitroprusside induced DNA strand breaks to the same extent in both, PARP+/+ and PARP–/– cells (data not shown). Interestingly, at higher concentrations of the NO donor, lysis also occurred in PARP–/– cells (Fig. 4A). At 2 mM sodium nitroprusside more than 80% of PARP–/– cells were lysed, which demonstrates that high concentrations of NO are cytotoxic also in the absence of PARP. As described in Table I, the lytic activity of low NO levels could be almost completely inhibited by 3-AB, while no protection was provided by the PARP inhibitor when islet cells were exposed to high sodium nitroprusside concentrations.Table IInhibition of radical-induced islet cell lysisToxic insultPARP+/+PARP−/−Untreated3-ABUntreated3-ABSodium nitroprusside% 0.45 mM44.0 ± 5.57.9 ± 6.15.9 ± 3.64.1 ± 2.7 1.35 mM93.1 ± 7.989.9 ± 3.474.2 ± 4.175.2 ± 3.1Xanthine oxidase 5.0 milliunits/ml79.2 ± 8.825.4 ± 8.852.7 ± 12.958.2 ± 6.8 Open table in a new tab Parallel studies were performed in ROI-treated islet cells. In contrast to the results observed with NO, PARP–/– islet cells were not more resistant to low doses of ROI than PARP+/+ cells, but were lysed to similar extent as PARP+/+ cells (Fig. 4B). However, at levels of 5 milliunits of xanthine oxidase and above, lysis was lower in PARP–/– cells (Fig. 4B, p < 0.02). The majority of wild type islet cells could be protected from lysis by 3-AB while death of PARP deficient cells was not prevented (Table 1). NO and ROI are supposed to be important mediators of toxic damage to pancreatic islet cells during prediabetic islet inflammation. Previous studies have shown that the exposure to these toxic insults leads to DNA damage, PARP activation, NAD+ depletion and subsequent islet cell lysis (9Radons J. Heller B. Bürkle A. Hartmann B. Rodriguez M.-L. Kröncke K.-D. Burkart V. Kolb H. Biochem. Biophys. Res. Commun. 1994; 199: 1270-1277Crossref PubMed Scopus (201) Google Scholar, 16Heller B. Bürkle A. Radons J. Fengler E. Jalowy A. Müller M. Burkart V. Kolb H. Biol. Chem. Hoppe-Seyler. 1994; 375: 597-602Crossref PubMed Scopus (43) Google Scholar). In those studies and in the present experiments plasma membrane lysis was used as end point of cell destruction because DNA strand breaks or impaired ATP production may represent reversible forms of cell damage. It is known that cell death may occur along different pathways. All these pathways converge in the end point of irreversible damage of the plasma membrane. However, dependent on the definition, cell death may have occurred prior to cell lysis. The conclusions of our study relate to the causal relationship between excessive poly(ADP-ribose) formation following NO- or ROI-induced DNA damage and islet cell lysis. The previously published data are descriptive and are based on correlations between NAD+ depletion and subsequent cell lysis as well as on protective effects of pharmacological inhibition of PARP (9Radons J. Heller B. Bürkle A. Hartmann B. Rodriguez M.-L. Kröncke K.-D. Burkart V. Kolb H. Biochem. Biophys. Res. Commun. 1994; 199: 1270-1277Crossref PubMed Scopus (201) Google Scholar, 10Kallmann B. Burkart V. Kröncke K.-D. Kolb-Bachofen V. Kolb H. Life Sci. 1992; 51: 671-678Crossref PubMed Scopus (115) Google Scholar, 11Fehsel K. Jalowy A. Sun Q. Burkart V. Hartmann B. Kolb H. Diabetes. 1993; 42: 496-500Crossref PubMed Scopus (198) Google Scholar, 15Zhang J. Dawson V.L. Dawson T.M. Snyder S.H. Science. 1994; 263: 687-689Crossref PubMed Scopus (1085) Google Scholar, 16Heller B. Bürkle A. Radons J. Fengler E. Jalowy A. Müller M. Burkart V. Kolb H. Biol. Chem. Hoppe-Seyler. 1994; 375: 597-602Crossref PubMed Scopus (43) Google Scholar, 17Yamamoto H. Uchigata Y. Okamoto H. Nature. 1981; 294: 284-286Crossref PubMed Scopus (477) Google Scholar, 18Sims J.L. Berger S.J. Berger N.A. Biochemistry. 1983; 22: 5188-5194Crossref PubMed Scopus (169) Google Scholar). To approach this issue we used cells from PARP-deficient mice and compared them to wild type cells. We found that both kinds of cells showed DNA strand breaks after exposure to reactive radicals. Although the xanthine oxidase concentrations used were in a range comparable to the enzymatic activity detected in rodent pancreas (26Burkart V. Koike T. Brenner H.-H. Kolb H. Diabetologia. 1992; 35: 1028-1034Crossref PubMed Scopus (70) Google Scholar, 32Parks D.A. Granger D.N. Acta Phys. Scand. Suppl. 1986; 548: 87-99PubMed Google Scholar) it is not known, whether the doses of ROI applied correspond to ROI levels possibly released during insulitis. The NO donor nitroprusside was dosed such that kinetics of DNA damage and cell lysis resembled those of NO dependent macrophage cytotoxicity toward islet cells (11Fehsel K. Jalowy A. Sun Q. Burkart V. Hartmann B. Kolb H. Diabetes. 1993; 42: 496-500Crossref PubMed Scopus (198) Google Scholar, 25Appels B. Burkart V. Kantwerk-Funke G. Funda J. Kolb-Bachofen V. Kolb H. J. Immunol. 1989; 142: 3803-3808PubMed Google Scholar, 31Kröncke K.-D. Kolb-Bachofen V. Berschick B. Burkart V. Kolb H. Biochem. Biophys. Res. Commun. 1991; 175: 752-758Crossref PubMed Scopus (255) Google Scholar). After exposure to ROI, only PARP+/+ cells showed a severe loss of NAD+. 3-Aminobenzamide did not completely restore NAD+ levels (p < 0.05), which has been observed previously in other cell types despite a 99% inhibition of PARP activity (33Carson D.A. Seto S. Wasson D.B. J. Exp. Med. 1986; 163: 746-751Crossref PubMed Scopus (71) Google Scholar). A minor reduction of intracellular NAD+ was also seen in PARP–/– cells exposed to ROI. This reduction may be due to direct inactivation of NAD+ by oxygen radicals, to damage-induced NAD+ hydrolysis by other enzymatic activities (34Richter C. Schlegel J. Toxicol. Lett. 1993; 67: 119-127Crossref PubMed Scopus (82) Google Scholar), or by a disturbance of NAD+ biosynthetic pathways. The lack of NAD+ depletion in ROI-treated PARP–/– cells was not due to a lack of DNA strand breaks in such cells. Furthermore, previous studies showed that pharmacologically reduced PARP activity does not lead to lesser DNA damage in islet cells exposed to ROI or NO (9Radons J. Heller B. Bürkle A. Hartmann B. Rodriguez M.-L. Kröncke K.-D. Burkart V. Kolb H. Biochem. Biophys. Res. Commun. 1994; 199: 1270-1277Crossref PubMed Scopus (201) Google Scholar, 16Heller B. Bürkle A. Radons J. Fengler E. Jalowy A. Müller M. Burkart V. Kolb H. Biol. Chem. Hoppe-Seyler. 1994; 375: 597-602Crossref PubMed Scopus (43) Google Scholar). In conclusion, the comparison of PARP+/+ and PARP–/– cells offers for the first time direct proof of a cause-effect relationship between PARP activity and NAD+ depletion. A second question which could be addressed by the use of PARP–/– cells was the proposed role of PARP in toxic islet cell death (10Kallmann B. Burkart V. Kröncke K.-D. Kolb-Bachofen V. Kolb H. Life Sci. 1992; 51: 671-678Crossref PubMed Scopus (115) Google Scholar, 17Yamamoto H. Uchigata Y. Okamoto H. Nature. 1981; 294: 284-286Crossref PubMed Scopus (477) Google Scholar). Our data show that PARP–/– islet cells are highly resistant to low doses of NO compared with PARP+/+ islet cells. Lysis of PARP+/+ islet cells induced by ROI or NO takes about 4–8 h to occur (10Kallmann B. Burkart V. Kröncke K.-D. Kolb-Bachofen V. Kolb H. Life Sci. 1992; 51: 671-678Crossref PubMed Scopus (115) Google Scholar, 16Heller B. Bürkle A. Radons J. Fengler E. Jalowy A. Müller M. Burkart V. Kolb H. Biol. Chem. Hoppe-Seyler. 1994; 375: 597-602Crossref PubMed Scopus (43) Google Scholar, 35Kröncke K.-D. Brenner H.-H. Rodriguez M.-L. Etzkorn K. Noack E.A. Kolb H. Kolb-Bachofen V. Biochim. Biophys. Acta. 1993; 1182: 221-229Crossref PubMed Scopus (104) Google Scholar) while viability of PARP–/– islet cells exposed to 0.45 mM nitroprusside was still in the normal range after 18 h. The different sensitivity of PARP–/– and PARP+/+ cells cannot be explained from differences in mouse strains, since both cell types had the genetic background of 129/Sv × C57BL/6 mice. This indicates that, at the lower concentrations of the NO donor, PARP-dependent NAD+ depletion is in fact the cause of islet cell death and hence PARP–/– cells are resistant. Other known toxic actions of NO, such as the inhibition of the electron transfer chain in mitochondria, of aconitase and glyceraldehyde-3-phosphate dehydrogenase (7Welsh N. Sandler S. Biochem. Biophys. Res. Commun. 1992; 182: 333-340Crossref PubMed Scopus (84) Google Scholar, 8Molina y Vedia L. McDonald B. Reep B. Brüne B. Di Silvio M. Billiar T.R. Lapetina E.G. J. Biol. Chem. 1992; 267: 24929-24932Abstract Full Text PDF PubMed Google Scholar) or the release of intracellular iron (36Corbett J.A. Lancaster J.R. Sweetland M.A. McDaniel M.L. J. Biol. Chem. 1991; 266: 21351-21354Abstract Full Text PDF PubMed Google Scholar) do not appear to be involved. However, at higher concentrations of the NO donor, lysis also occurred in PARP–/– cells. These findings reveal the existence of an alternative pathway of cell death which does not require PARP activation and NAD+ depletion. Another issue which could be resolved by the use of PARP–/– cells was whether 3-AB exerts its protection against ROI or NO toxicity via PARP inhibition or by one of its other pharmacological properties such as radical scavenging. We found here that 3-AB was able to prevent ROI and NO cytotoxicity only in PARP+/+, but not in PARP–/– cells. This identifies PARP as a target of 3-AB's cytoprotective effect. Interestingly, 3-AB prevented NO cytotoxicity only at low nitroprusside concentrations in PARP+/+ cells. These findings clearly demonstrate the existence of an alternative pathway not only in PARP–/– but also in PARP+/+ cells. In contrast to the results obtained with NO, the exposure of PARP–/– islet cells to low doses of ROI led to a similar extent of cell lysis as in PARP+/+ cells. However, the cell lysis was significantly lower in PARP–/– cells at high concentrations of ROI. As expected, 3-AB could protect only PARP+/+ cells from lysis, whereas cell death could not be prevented in PARP-deficient cells. Taken together these data suggest that PARP/NAD+-dependent and independent cell death occur alternatively in NO- and ROI-treated islet cells. In the presence of PARP the NAD+-dependent form of cell death seems to be dominant. Such alternative activation of PARP-dependent or independent cell death suggests that early steps are common to both pathways (such as DNA damage), followed by alternative routes of cell death, one of them involving PARP activation and NAD+ depletion. At present it cannot be decided whether toxic damage of beta cells during prediabetic inflammation is PARP-dependent. In favor of such a mechanism are the findings that the NO-mediated toxicity of activated macrophages as well as that of ROI or interleukin 1 toward islet cells is suppressed in the presence of the PARP inhibitors nicotinamide or 3-AB (31Kröncke K.-D. Kolb-Bachofen V. Berschick B. Burkart V. Kolb H. Biochem. Biophys. Res. Commun. 1991; 175: 752-758Crossref PubMed Scopus (255) Google Scholar, 37Pipeleers D. Van De Winkel M. Proc, Natl. Acad. Sci. U. S. A. 1986; 83: 5267-5271Crossref PubMed Scopus (105) Google Scholar, 38Buscema M. Vinci C. Gatta C. Rabuazzo M.A. Vignen R. Purrello F. Metabolism. 1992; 41: 296-300Abstract Full Text PDF PubMed Scopus (31) Google Scholar). Furthermore, in vivo administration of nicotinamide was found to prevent diabetes development in a spontaneous autoimmune diabetes model, the NOD mouse (39Yamada K. Nonaka K. Hanafusa T. Miyazaki A. Toyoshima H. Tarui S. Diabetes. 1982; 31: 749-753Crossref PubMed Scopus (0) Google Scholar). The introduction of the inactivated PARP gene into diabetes-prone mouse strains offers a direct approach to studying the role of PARP activation in islet cell death in vivo. In conclusion, the use of islet cells from mice with a disrupted PARP gene has enabled us to provide direct evidence for a causal relationship between PARP activation, NAD+ depletion and subsequent islet cell death. In general, PARP-deficient cells were less sensitive toward the NAD+-dependent toxicity, induced by ROI and NO. In the absence of PARP a second, NAD+-independent form of cell death became evident. PARP-dependent and -independent forms of cell death were found to represent alternative responses of islet cells to the toxic immune mediators ROI and NO. We thank Drs. Miwa and T. Sigumura (Tokyo, Japan) for the kind gift of 10H hybridoma cells, M. Turken for photographic work, and M. Rinker for typing the manuscript.
Referência(s)