Despite more than 20 years of extensive research, sepsis and systemic inflammatory response syndrome remain the main causes of death in intensive care units, with mortality rates between 30 and 70% (Angus et al., 2001). The pathogenesis of endotoxin-induced organ dysfunction is multifactorial and incompletely understood. The lack of efficacy of anti-inflammatory drugs shifted the interest toward developing alternative treatments. Because PARP-1-deficient mice were resistant to lethality induced by LPS, inhibitors of this nuclear enzyme were suggested to be a promising tool for therapeutic intervention (Oliver et al., 1999; Soriano et al., 2002).

LPS plays a major role in inducing sepsis during infection caused by Gram-negative bacteria. Endothelial and epithelial cells, as well as neutrophils, macrophages, and lymphocytes produce powerful proinflammatory mediators, especially tumor necrosis factor-α (TNF-α), interleukin (IL)-6, IL-1β, and IL-8, and play important roles in the induction of sepsis (Stone, 1994). Binding of LPS to the CD14 and TLR4/MD2 complex induces the production of proinflammatory mediators by activating mitogen-activated protein (MAP) kinase pathways, including extracellular signal-regulated kinase (p42/44 MAPK or ERK1/2), p38 kinase, and c-Jun NH₂-terminal kinase (JNK) (Dumitru et al., 2000). MAPKs are capable of modulating functional responses through phosphorylation of transcription factors and activation of other kinases. The activated p42/44 MAPK will, in turn, phosphorylate an array of cellular substrates, including downstream Ser/Thr effector kinases such as p90RSK (protein of 90 kDa from the ribosomal subunit S6 kinase) (Sturgill et al., 1988).

Induction of cytokine production and the expression of proinflammatory genes are achieved by the activation of nuclear transcription factors, such as activator protein 1 (AP-1), or by activation and nuclear translocation of cytoplasmic transcription factors such as NF-κB (Bone et al., 1997; Bozinovski et al., 2002). MAPKs can mediate phosphorylation and activation of both (Dumitru et al., 2000). AP-1, an obligatory transcription factor in inflammation and innate immunity, exists either as a homodimeric c-Jun complex or as a c-Jun/c-Fos heterodimer that is regulated by transcription and direct phosphorylation. Phosphorylation of c-Fos stabilizes the transcription factor and enhances the trans-activation and DNA binding of AP-1 (Shaywitz and Greenberg, 1999). Previous reports have showed that PARP-1 plays a pivotal role in activation of transcription factors NF-κB and AP-1 (Andreone et al., 2003), and inhibition of PARP-1 can attenuate this activation (Veres et al., 2003).

Studies with the traditional PARP inhibitor 3-aminobenzamide suggested that PARP activity either does not have (Baechtold et al., 2001) or has only a partial (Albertini et al., 2000) role in the mechanisms of septic shock. However, 3-aminobenzamide is considered to be a poor inhibitor of PARP and has pronounced toxicity in vivo. On the other hand, novel, potent PARP-1 inhibitors were found to protect against LPS-induced tissue damage (Liaudet et al., 2002; Ivanyi et al., 2003; Veres et al., 2003). Disruption of the PARP gene suppresses LPS-induced cytokine expression and NF-κB activation (Oliver et al., 1999). However, there are some data suggesting that the presence of PARP protein and not its enzymatic activity is necessary for the NF-κB activation (Hassa et al., 2001). Several reports have demonstrated the importance of the PI3-kinase/Akt pathway in LPS-induced inflammatory mechanisms (Ozes et al., 1999; Bozinovski et al., 2002; Guha and Mackman, 2002), and recent evidence suggests that activation of PI3-kinase, a ubiquitous lipid-modifying enzyme, may modulate positively acting signaling pathways. It therefore seemed plausible to investigate whether the aforementioned pathways are involved in the effect of PARP-1 inhibitors on LPS-induced septic shock.

To this end, we have used an MRI technique to visualize the LPS-induced inflammatory response in vivo and its attenuation by a PARP-1 inhibitor in a rodent septic shock model. In an attempt to uncover the mechanism of this protective effect, we have tested the effect of the drug administered 1 or 6 h after the LPS treatment on the survival of mice. Additionally, we have determined its effect on TNF-α expression as well as on the activation of the transcription factors NF-κB and AP-1, and ERK1/2, p90RSK, p38, and PI3-kinase/Akt signaling pathways. We have used 4-HQN, an effective PARP-1 inhibitor (Banasik et al., 1992) in our experiments, because it was structurally unrelated to the inhibitors with previously characterized anti-inflammatory properties in addition to having protective properties regarding ischemia-reperfusion-induced cardiac damage (Halmosi et al., 2001).

Materials and Methods

Animals. BALB/c mice were purchased from Charles River Hungary Breeding Ltd. (Budapest, Hungary). The animals were kept under standardized conditions; tap water and mouse chow were provided ad libitum during the whole experimental procedure. Animals were treated in compliance with approved institutional animal care guidelines.

Materials. LPS from Escherichia coli 0127:B8 and 4-hydroxyquinazoline were purchased from Sigma-Aldrich (St. Louis, MO); primary antibodies, anti-phospho-p44/42 MAP kinase (Thr202/Tyr204), anti-phospho-Akt (Ser473, Thr308), anti-phospho-GSK-3β (Ser9), anti-phospho-stress-activated protein kinase/JNK (Thr183/Tyr185), and anti-phospho-p90RSK (Thr359/Ser363) were from Cell Signaling Technology Inc. (Waltham, MA); anti-phospho-p38 (Thr180/Tyr182) was from Sigma-Aldrich; anti-cyclooxygenase-2 was from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA) and anti-high mobility group-1 was from BD Biosciences (San Diego, CA).

Sepsis Model. To induce murine endotoxic shock, BALB/c mice were injected i.p. with LPS at a dose of 20 mg/kg body weight in a volume of 250 μl. 4-HQN (100 mg/kg) was administered i.p. in a volume of 250 μl three times on the day preceding the LPS challenge (pretreatment), or as a single dose 1 or 6 h after the LPS injection. Control mice received the same volume of sterile saline solution instead of the PARP-1 inhibitor. The mice were monitored for clinical signs of endotoxemia and lethality every hour for 48 h, after which time they were monitored three times a day for 1 week. No late deaths were observed in any of the experimental groups. Each experimental group consisted of 10 mice. Data represents means of three independent experiments.

4-HQN was reported as a potential PARP-1 inhibitor (Banasik et al., 1992) with an IC₅₀ of 9.5 μM. We have measured the inhibitory effect of 4-HQN on PARP-1, and our results were in agreement with the above-mentioned data (IC₅₀ of 8–12 μM). In accordance with the notion that an animal model requires higher dose than an in vitro system, we selected 100-mg/kg dose, which had a significant protective effect in the survival studies. In a previous study (Halmosi et al., 2001), we have established that this dose of 4-HQN effectively inhibited PARP-1 activity in Langerdorff-perfused hearts, because it prevented the ischemia-induced ADP-ribosylation of proteins. Because this substance has a high potency on PARP-1 and no effects of it on enzymes other than PARP have been documented, it seemed likely that our observations could be assigned to its PARP inhibitory effect.

Western Blot Analysis. For Western blot analysis, groups of four BALB/c mice were pretreated or not with 100 mg/kg 4-HQN three times a day preceding the day of LPS challenge (20 mg/kg). Liver, lung, and spleen were removed from the animals 6 and 16 h after the LPS treatment, frozen in liquid N₂, and processed exactly as described previously (Veres et al., 2003). Protein load was 35 μg/lane. We applied the primary antibodies at 4°C overnight at a dilution of 1:1000. The secondary antibodies were horseradish peroxidase-conjugated rabbit IgG. Peroxidase labeling was visualized with the SuperSignal West Pico chemiluminescent substrate (Pierce Chemical, Rockford, IL).

Quantification of band intensities (E₅₄₀) of the blots was performed by a DU-62 spectrophotometer equipped with a densitometry attachment (Beckman Coulter Inc., Fullerton, CA) and ImageJ software. Data representing three independent experiments are expressed as percentage of the untreated control (mean ± S.E.M.).

TNF-α Determination. Mice were treated exactly as for Western blot analysis. Serum TNF-α concentrations were determined with the Quantikine M TNF-α immunoassay kit (R&D Systems, Abingdon, UK). Blood samples were taken 1.5 h after LPS administration and were allowed to clot for 0.5 h at room temperature before centrifuging for 20 min at 2000g. The serum was removed and assayed immediately. To calculate the results, we created a standard curve by reducing the data using computer software to generate a four-parameter logistic curve fit. The mouse TNF-α standard dilution series were prepared in accordance with the protocol of the manufacturer. Data represent mean ± S.E.M. of 12 independent values (three independent experiments with four mice in each group.).

Determination of NF-κB and AP-1/c-Fos Activation. Mice were treated exactly as for Western blot analysis. For nucleus isolation, liver, spleen, and lung were removed 1.5 h after the LPS-treatment and were homogenized immediately according to the procedure described previously (Veres et al., 2003). Protein concentrations in nuclear extracts were determined using a bicinchoninic acid assay with bovine serum albumin as standard (Sigma-Aldrich). To monitor NF-κB and AP-1/c-Fos activation in tissues, we used Trans-AM transcription factor assay kits (Active Motif, Rixensart, Belgium). The kit consists of 96-well plates into which oligonucleotides containing the NF-κB and AP-1 c-Fos consensus sites (5′-GGGACTTTCC-3′ and 5′-TGAGTCA-3′, respectively) are bound. The active forms of the above-mentioned transcription factors in the nuclear extract specifically bind to these consensus sites and are recognized by primary antibodies. A horseradish peroxidase-conjugated secondary antibody provides the basis for the colorimetric quantification. Results were expressed as percentages of the positive controls (TNF-α-stimulated HeLa whole cell extract and WI-38 nuclear cell extract, respectively) provided by the manufacturer and represent mean ± S.E.M. of 12 independent values (three independent experiments with four mice in each group.).

MRI Analysis. Mice were treated exactly as for Western blot analysis. Six hours after LPS treatment, the animals were anesthetized with urethane (1.7 g/kg administered i.p.) and were placed into an epoxy resin animal holder tube.

MRI measurements were performed on a Varian ^UNITYINOVA 400 spectrometer (Varian, Palo Alto, CA) with a 89-mm vertical bore magnet of 9.4 T (Oxford Instruments Ltd., Abingdon, UK) using a 35-mm inner diameter hollow micro-imaging probe with a built-in self-shielded gradient system up to 400 mT/m (Doty Scientific, Inc., Columbia, SC). After tuning, shimming (¹H line width ≈150 Hz) and radio frequency calibration, the slice of interest was selected using a T₁-weighted multi-slice spin-echo sequence (4.0-ms sinc pulses; TR, 1000 ms; TE, 12 ms; slice thickness, 1 mm; FOV, 30 × 30 mm; acquisition matrix, 128 × 128). T2-weighted images were recorded using a multi-slice spin-echo sequence (parameters were like at T₁-weighting, except TR, 3000 ms and TE, 50 ms). One average was taken and images were reconstructed as 256 × 256 matrices. The intensities of the images were standardized to the signal of a 1-mm inner diameter tube filled with water/glycerol (9:1), which was placed near the animal during the measurements. Mean signal intensities were measured and expressed as a percentage of the signal intensity of the internal standard in characteristic regions delineated as freehand areas by experts who were blinded to the experiment. Due to individual differences among the animals, with special respect to uncertainty of their posture and position (i.e., orientation of their internal organs) within the magnet of the NMR instrument, this scoring should be considered as semiquantitative. Experiments were repeated three times.

Inflammation-affected regions show up in T2-weighted MRI with increased signal intensity, due to cellular invasion and edema, as reported recently for paraspinal or epidural inflammation and soft tissue inflammation (Ledermann et al., 2002). Agreement between MRI data and histopathology was found to be satisfactory for acute local infection in mouse muscle (Ruiz-Cabello et al., 2002).

Statistical Analysis. When pertinent, data were presented as means ± S.E.M. For multiple comparisons of groups, analysis of variance was used. Statistical difference between groups was established by paired or unpaired Student's t test, with Bonferroni's correction.

Results

Effect of 4-HQN on Survival of LPS-Treated Mice. Mice treated with a single dose of LPS (20 mg/kg) died within 30 h. When the mice were pretreated three times on the day preceding the LPS challenge by 4-HQN (100 mg/kg), 80% of the animals in the group survived. Even when groups of mice received only a single injection of 4-HQN (100 mg/kg) 1 h after the LPS challenge, the PARP-1 inhibitor significantly protected the animals against LPS-induced death with 30% surviving mice (Fig. 1). When 4-HQN was administered 6 h after LPS challenge in a single treatment (100 mg/kg), it did not have protective effect (Fig. 1). 4-HQN treatment itself did not induce death or any obvious damage (data not shown).

MRI Analysis. Untreated, 4-HQN-treated, LPS-treated, and LPS + 4-HQN-treated mice underwent MRI analysis. T2-weighted transversal spin-echo images were taken from the thoracic and lower abdominal regions (Fig. 2). Signal intensities of T2 images were proportional to the inflammatory response. Representative images selected from 12 images for each group were presented in Fig. 2, and normalized mean pixel intensities for all the images were summarized in Table 1.

T2-weighted images of the thoracic regions showed considerably increased intensities in the dorsal subcutaneous region, moderately increased intensities in the intramuscular regions, and no observable difference in and inside the pleura of the LPS-treated mice. In the abdominal regions, characteristic increases were observed around the kidneys and in the interintestinal cavities (Fig. 2). Among all the observed LPS-induced inflammatory responses, we found the most characteristic and most pronounced increases in the gastrointestinal tract (Fig. 2). On the other hand, no increase in the signal could be observed inside the kidneys and in skeletal muscle, neither in the paravertebral nor in the femoral muscles (data not shown). All increases in signal intensities were significantly attenuated in mice treated with 4-HQN (Table 1), indicating that the PARP-1 inhibitor reversed the LPS-induced morphological changes (Fig. 2). Mice treated with 4-HQN only were identical to the untreated control (data not shown).

Effect of 4-HQN on LPS-Induced TNF-α Production. LPS-treatment resulted in a rapid increase in serum TNF-α concentration that reached 2960 ± 112 pg/ml after 90 min (Fig. 3). In the 4-HQN-treated mice, LPS challenge resulted in significantly lower TNF-α concentrations (1500 ± 135 pg/ml). 4-HQN alone did not exert any significant effect on serum TNF-α levels (Fig. 3).

Effect of 4-HQN on LPS-Induced Phosphorylation of Various Kinases in Liver, Lung, and Spleen. Phospho-ERK1/2, phospho-p38, phospho-Akt, phospho-GSK-3β, phospho-p90RSK, and phospho-JNK expression were determined by Western blotting from lung, liver, and spleen of untreated mice, mice treated with LPS or 4-HQN, and from 4-HQN + LPS-treated mice. Using a phosphorylation-specific antibody against phospho-Akt (Ser⁴⁷³), we were able to demonstrate activation of Akt under our experimental conditions. We did not find Akt activation in tissues of untreated animals or in animals treated with LPS alone. However, there was a marked increase in the phosphorylation and thereby the activation of Akt in every studied tissue of 4-HQN- and 4-HQN + LPS-treated mice (Fig. 4). This activation reached 2880 ± 144% in spleen, 835 ± 54% in liver, and 957 ± 103% in lung in animals treated with 4-HQN + LPS and 416 ± 57, 554 ± 19, and 1158 ± 97% in animals treated with 4-HQN alone, respectively. Using an antibody against phospho-Akt (Thr³⁰⁸), we did not find Akt activation either with LPS or with 4-HQN treatment (data not shown). Under our experimental conditions, we could not detect phospho-GSK-3β (Ser⁹) synthesis, even though our assay system was able to detect GSK-3β phosphorylation in other systems (data not shown).

Extracellular signal-regulated kinase phosphorylation and activation were determined by Western blotting using an anti-phospho-p44/42 MAP kinase (Thr²⁰²/Tyr²⁰⁴) antibody. LPS-treatment resulted in a marked increase in activation of ERK1/2 in the spleen (4360 ± 270%), liver (5025 ± 345%), and lung (4575 ± 298%). Pretreatment with 4-HQN significantly attenuated this activation in all three tissues (2022 ± 125, 1358 ± 97, and 2174 ± 203%, respectively). 4-HQN treatment itself did not have any effect on the activation of phospho-ERK1/2 in the tissues we studied (Fig. 5).

p90RSK is a downstream target of p44/42 MAP kinase in the ERK pathway. Phosphorylation of p90RSK was determined using an anti-phospho-p90RSK antibody. In the lungs of LPS-treated mice, we found a marked activation of p90RSK (5964 ± 420%), which was attenuated by 4-HQN pretreatment (4550 ± 121%) (Fig. 6). However, in spleen and in liver, LPS did not induce the activation of p90RSK, and 4-HQN had no additional effect (data not shown).

Phosphorylation of MAPK p38 was determined using an anti-phospho-p38 antibody. As shown in Fig. 7, LPS-stimulated p38 activation in the lung (368 ± 42%) was abolished by 4-HQN pretreatment. However, in spleen and in liver, LPS did not induce the activation of p38, and 4-HQN had no additional effect (data not shown).

Under our experimental conditions, we could not detect phospho-JNK (Thr¹⁸³/Tyr¹⁸⁵) synthesis, even though our assay system was able to detect JNK phosphorylation in other systems (data not shown).

Effect of 4-HQN on LPS-Induced NF-κB Activation. Ninety minutes after LPS treatment, NF-κB activation was assessed in spleen, lung, and liver. LPS-treatment caused a significant increase in NF-κB activation in all three tissues. That was slightly but not statistically significantly attenuated by 4-HQN pretreatment in spleen (Fig. 8A). However, and in contrast to the spleen, NF-κB activation in the lung and in the liver was prevented by 4-HQN pretreatment (Fig. 8, B and C). 4-HQN treatment itself did not have any effect on NF-κB activation in the tissues studied.

Effect of 4-HQN on LPS-Induced AP-1 (c-Fos) Activation. Ninety minutes after LPS treatment, phosphorylation of AP-1 family member c-Fos and thereby the activation of this transcription factor was assessed in spleen, liver, and lung. LPS treatment caused a significant increase in c-Fos activation in all three tissues. 4-HQN pretreatment significantly inhibited this activation in the tissues we studied (Fig. 9). 4-HQN treatment itself did not have any effect on c-Fos activation in the studied tissues.

Discussion

Despite the importance of LPS as a trigger of innate host defenses and inflammation, very little is known of the pathomechanism and actual transduction pathways activated by this endotoxin. The in vivo response to endotoxin-induced septic shock and its prevention by the PARP-1 inhibitor 4-HQN was detected by MRI techniques. As we showed in our previous work (Veres et al., 2003), this method provides a real-time insight into inflammatory processes in living animals, even in cases where no histological alteration can be detected. We have observed the most characteristic and most pronounced increases in the gastrointestinal and in the thoracic tract, which were similar to those found in a porcine endotoxic shock model (Oldner et al., 1999). The importance of these tracts in the mediation of sepsis and multiple-system organ failure is well documented (Standiford et al., 1995; Boulares et al., 2003). Hyperintensity in MRI signals due to inflammation and edema is originated from the intercellular space and not the organs themselves. Simultaneously, this hyperintensity decreases or sometimes erases contrast, making it difficult or sometimes impossible to outline certain anatomical formulae within these areas.

To determine whether PARP-1 inhibition interferes with early or late mediators of LPS-induced septic shock, we performed survival studies in which we administered 4-HQN 1 or 6 h after the LPS challenge. We have found that 4-HQN had a protective effect (30% survival rate) when it was added 1 h but not when it was added 6 h after LPS injection, suggesting that 4-HQN may interfere preferentially with the early mediators of the effect of LPS. To support this observation, we studied some important late mediators of septic shock by Western blotting and found that LPS treatment did not affect the high mobility group-1 and cyclooxygenase-2 expression 16 h after the LPS challenge in lung, liver, or spleen.

TNF-α is a substantial early mediator of endotoxemia because the production of this cytokine returned to a normal level 4 h after LPS treatment (Zanetti et al., 1992; James et al., 2002). When administered before LPS treatment, 4-HQN attenuated the LPS-induced elevation in the serum TNF-α concentrations by approximately 50%, consistent with the notion that this PARP-1 inhibitor, similarly to PJ34, partially inhibits the expression of TNF-α gene (Veres et al., 2003).

Due to the importance of the functional state of different organs during septic shock, we investigated in lung, liver, and spleen various protein kinases that lead to transcription factor activation. Previous studies have shown that LPS induces activation of MAPK pathways in different cell lines. Activated MAPKs play a key role in the transduction of the LPS signal between the extracellular receptor and the intracellular (cytoplasmic and nuclear) response, resulting in activation of gene expression (Dumitru et al., 2000; Bozinovski et al., 2002). Furthermore, several articles showed that there is cross talk between ERK and other MAP kinases in different cell lines (Dumitru et al., 2000; Guha and Mackman, 2002; Xiao et al., 2002). In agreement with these results, we found that ERK and p38 MAPK did indeed show a similar LPS-induced activation in lung but not in spleen and liver. In monocytes, macrophages, and Tpl2 knockout mice, a functional association between ERK and JNK was reported that was not detectable in our BALB/c mice sepsis model. We found, however, that p90RSK, a downstream target of ERK, showed a similar activation pattern in lung as ERK and p38 in response to the LPS treatment. The significance of this finding can be understood by considering that phosphorylated (i.e., activated) p90RSK can activate nuclear transcription factors such as c-Fos and NF-κB.

Because ERK is a key element of the signal transduction pathway that regulates LPS-mediated transcription factor activity, inhibition of this MAP kinase can contribute to reducing the endotoxin-induced inflammatory responses. Under our experimental conditions, ERK was activated by LPS in all of the tissues studied (liver, lung, and spleen) but p38 MAP kinase had detectable activation only in lung, indicating that signaling from the cell surface CD14-TLR4/MD2 LPS receptor to the activation of the different branches of MAP kinases, has tissue-specific components. Furthermore, we did not find detectable activation of the JNK pathway, suggesting that there is no functional association between either the JNK pathway and the cell surface LPS receptor or ERK and JNK pathways in these tissues in our experimental system. Interestingly, the phosphorylation of the ERK substrate p90RSK also showed tissue specificity, indicating that ERK activation does not necessarily trigger the activation of p90RSK and the activation of p90RSK-dependent transcription factors. LPS induced the activation of ERK, p90RSK, and p38 MAP kinase but not JNK in lung. In addition, there was a significant increase in the activation of NF-κB and c-Fos transcription factors showing that gene expression in lung is regulated by ERK or p38 MAP kinase but not by the JNK pathway. In liver and spleen, only the ERK pathway was activated by LPS, and it is likely that this pathway activated the c-Fos and NF-κB transcription factors. In the absence of PARP-1 inhibitor, LPS did not affect the PI-3-kinase/Akt pathway; therefore, MAP kinase pathways had to play the major role in activation of AP-1 and NF-κB.

In monocytes, it was found that inhibition of the PI-3-kinase/Akt pathway can activate MAP kinases (Guha and Mackman, 2002). Therefore, it is likely that activated Akt in the presence of the PARP-1 inhibitor mediates the inhibition of MAP kinases and so prevents the activation of NF-κB and AP-1 transcription factors, resulting in inhibition of the consequent inflammatory gene expression. In conclusion, the most important protective effect of PARP-1 inhibitors in different organs could be the activation of Akt, inhibition of MAP kinases, and the inhibition of related transcription factors (Fig. 10).

Because Akt activation can be induced by structurally different PARP-1 inhibitors, such as quinazoline derivates, phenantridine derivates (Veres et al., 2003), and carboxamino benzimidazol derivates (unpublished data), it is likely that this effect is related to the PARP-inhibitory property of these molecules, although it is not clear how the inhibition of the nuclear PARP can activate cytoplasmic Akt. Furthermore, Luo et al. (2003) have recently showed in another experimental model that PARP inhibitors prevent deactivation of Akt in N-methyl-d-aspartate-induced neurotoxicity. In spite of the lack of the precise molecular mechanism, the finding that PARP-1 inhibitors reverse the inflammatory processes and organ damage (Fig. 2; Veres et al., 2003), probably via the activation of one of the most important protective kinase cascades, the PI-3 kinase/Akt pathway, provides a novel possibility of prevention of multiorgan failure in septic shock. Although Akt activation seems to occur in all studied tissues (Fig. 4), the regulation of transcription factor activation seems to have some tissue specificity, indicating that there are some tissue-specific components between the MAP kinase pathway and the transcription factors regulated by this pathway. It can explain the observation that Akt activation suppressed NF-κB activation in lung and in liver but not in spleen (Fig. 8), whereas c-Fos activation was suppressed in all studied tissues (Fig. 9).

Isoquinasolines such as 4-HQN inhibit PARP-1 activity by competitive binding to the NAD binding site (Banasik et al., 1992). Thus, it is likely that all PARP isoforms are inhibited by 4-HQN because the catalytic site is highly conserved among the various PARP isoforms. In fact, no effects of 4-HQN on enzymes other than the PARP have been reported. The high potency of the compound on PARP together with the fact that our findings on 4-HQN-treated animals were very similar to those that were reported in PARP-1-deficient mice makes it likely that the principal action of the drug was mediated via PARP inhibition. The various classes of recently emerging potent, nontoxic PARP inhibitors will help to further clarify this question.

Together, our data show that LPS induces a different extent of MAP kinase activation in different organs, which can activate AP-1 and NF-κB transcription factors in a tissue-specific manner. Activation of these transcription factors induces activation of proinflammatory genes that are most likely responsible for the tissue damage during septic shock. PARP-1 inhibitors beside their well known effect of inhibiting NAD⁺ and ATP depletion, influence LPS-induced transcription factor activation and gene expression. These effects of PARP-1 inhibitors are mediated by the activation of the PI3-kinase/Akt pathway, which can inhibit MAP kinase activation and can attenuate transcription factor activation and inflammatory tissue damage (Fig. 10) in a tissue-specific manner.