Activation of poly(ADP-ribose) polymerase (PARP) mediates oxidative stress-induced cell injury. We tested the hypothesis that PARP contributes to ischemia-reperfusion (I/R) damage of the liver by triggering the mechanisms of microcirculatory failure. Leukocyte- and platelet-endothelial cell interactions as well as sinusoidal perfusion were analyzed by intravital fluorescence microscopy after lobar hepatic I/R (90 min/30 min) in C57BL/6 × 129/Sv wild-type (PARP+/+) and PARP-deficient (PARP−/−) mice. Hepatic I/R induced leukocyte/platelet-endothelial cell interactions and tissue injury in PARP+/+ mice, as indicated by impaired sinusoidal perfusion and increased alanine aminotransferase (ALT)/aspartate aminotransferase (AST) serum activities. In PARP−/− mice, however, the postischemic increase in the numbers of rolling/adherent leukocytes and platelets was significantly lower. In addition, I/R-induced translocation of CD62P as well as mRNA expression of CD62E, CD54, and CD106 were attenuated. The degree of perfusion failure was reduced and the increase in the ALT/AST activities was lower in PARP−/− mice compared with PARP+/+ mice. We conclude that PARP contributes to hepatic microvascular injury by triggering the expression/translocation of adhesion molecules and modulating leukocyte/platelet-endothelial cell interactions.
- adhesion molecules
ischemia-reperfusion(I/R) is responsible for primary liver dysfunction and failure after transplantation, liver resection, and hemorrhagic shock. There is a large body of evidence that the liver microcirculation has to be considered as a primary target in the development of hepatic I/R injury (see Ref. 20). Microcirculatory disturbances appear to be initiated by the generation of reactive oxygen species, cytokines (e.g., IL-1, TNF), and other proinflammatory mediators that activate both leukocytes and the hepatic vascular endothelium. Such activation initiates the expression of adhesion molecules and promotes microvascular leukocyte recruitment. The final consequence is the disintegration of the endothelial lining, with interstitial edema formation, sinusoidal perfusion failure, and parenchymal apoptotic/necrotic cell injury (for review, see Ref.19). Recently, we have shown in an intravital microscopic study that, in addition to leukocytes, platelets also interact with the postischemic endothelium of hepatic microvessels and play a critical role in the induction of liver damage after warm ischemia (10). The initial mechanisms triggering these microvascular phenomena remain not fully understood.
A candidate pathway of inflammation-induced tissue injury involves the nuclear enzyme poly(ADP-ribose) polymerase (PARP). Although its role as a nuclear repair enzyme is not completely understood, there is considerable evidence that PARP activation represents a cytotoxic pathway in the inflammatory response (31, 32). Reactive oxygen and nitric oxide (NO) metabolites are potent initiators of DNA single-strand breakage, which is an obligatory stimulus for the activation of PARP. This excessive and rapid PARP activation depletes the intracellular pool of NAD and high-energy phosphates, leading to the loss of membrane integrity, cell dysfunction, and finally cell death (27, 30). PARP blockade has been shown to prevent the I/R injury of brain (24), kidney (13), gut (12), and heart (39).
In the manifestation of I/R-induced liver injury, the role of PARP is not well defined and controversially discussed. Inhibitors of PARP attenuated the hepatic damage after hemorrhagic shock (17,36) but did not reduce liver injury after I/R (4). Although detection of cleaved PARP (89 kDa) is used for the assessment of apoptosis in the liver (38), hepatocellular PARP has been recently shown to be resistant to caspase-induced proteolytic cleavage (9). Although PARP inhibitors attenuated oxidative stress-induced PARP activation and energy depletion in isolated hepatocytes, this did not result in improved cell survival (11). In a recent in vivo study (7), however, PARP inhibition prevented hepatocellular I/R injury in the rat liver. Therefore, the pathway of postischemic liver injury mediated by PARP in vivo might be different from that occurring under in vitro conditions. We hypothesized that, either alternatively or in addition to the direct cytotoxic effect, activated PARP contributes to the induction of hepatic I/R injury by triggering the initial mechanisms of microcirculatory failure. Therefore, the objective of our study was to investigate the role of PARP in the mechanisms of microvascular hepatic I/R injury in vivo in mice lacking a functional gene for PARP.
MATERIALS AND METHODS
Experiments were performed in female mice at an age of 5–7 wk (weight 20–25 g). PARP-1-deficient (PARP−/−) mice were originally generated and kindly provided by Dr. Z.-Q. Wang (Lyon, France; Ref. 6). They were bred onto a C57BL/6 × 129/Sv background, and thus C57BL/6 × 129/Sv wild-type (PARP+/+) mice were used as controls for all experiments. Animals were kept under standard laboratory conditions and allowed free access to animal chow and tap water. All experiments were performed according to the German legislation on protection of animals.
The surgical procedure has been described in detail (2). Briefly, under inhalation anesthesia with isoflurane-N2O (FiO2 0.35, 0.015 L/L isoflurane; Forene; Abbott, Wiesbaden, Germany), polyethylene catheters (PE-10, ID 0.28 mm; Portex, Hythe, UK) were inserted into the left carotid artery and jugular vein for measurement of mean arterial pressure and application of fluorescence dyes. A warm reversible ischemia of the left liver lobe was induced for 90 min by clamping the supplying nerve vessel bundle with a microclip. Ninety minutes of partial hepatic ischemia were chosen in this study, since we observed in previous experiments that the microvascular hepatic I/R injury manifests itself significantly after at least 60 min of warm lobar ischemia and is most pronounced after 90 min (unpublished observation). An ischemia of 120 min causes, however, progressive exclusion of sinusoids from perfusion and, therefore, would not allow an in vivo analysis of the microcirculation (33,38).
Intravital fluorescence microscopy.
The hepatic microcirculation was analyzed by an epi-illumination technique using an intravital fluorescence microscope (Leitz, Wetzlar, Germany) from 30 to 55 min after the onset of reperfusion. The microscopic images were recorded by a charge-coupled device video camera (FK 6990, Cohu; Prospective Measurements, San Diego, CA) and transferred to a video system (S-VHS Panasonic AG 7330; Matsushita Electric Industries, Tokyo, Japan) for off-line evaluation.
Platelets were isolated from syngeneic either PARP+/+ or PARP−/− mice and labeled ex vivo by rhodamine 6G (50 μl, 0.05%/ml whole blood; Sigma-Aldrich, Deisenhofen, Germany) as described previously (15). A total of 1 × 108fluorescent-labeled platelets were infused intra-arterially at 30 min after the onset of reperfusion. Immediately after platelet infusion, platelet-endothelial cell interactions were analyzed in five to seven presinusoidal (terminal) arterioles (diameter 15–25 μm) and postsinusoidal venules (diameter 20–40 μm) as well as in all visible sinusoids of five to seven acini by using an N2 filter block (excitation 530–560 nm, emission >580 nm; Leitz). Terminal arterioles were distinguished from rarely seen final branches of portal vessels by their smaller diameter (21) and higher blood flow velocity. Then, rhodamine 6G (0.1 ml, 0.05%; Sigma-Aldrich) was administered intravenously to visualize leukocyte-endothelial cell interactions within postsinusoidal venules. Notably, an additional neutral density filter (5% transmission; Leitz) was needed for the assessment of leukocytes that were labeled through intravenous application of rhodamine 6G. The usage of this filter hampers the simultaneous visualization of previously applied fluorescent platelets because of their lower fluorescence intensity. Therefore, an analysis of the interactions of platelets with leukocytes was impossible in this study. Since it is well known that leukocytes stagnant in sinusoids neither cause injury nor affect sinusoidal perfusion (33,35), leukocyte accumulation was not quantified in microvessels of this type. After that, fluorescein isothiocyanate (FITC)-labeled dextran (molecular weight 150,000, 0.1 ml, 5%; Sigma-Aldrich) was applied intravenously and liver acini were scanned using an I2/3 filter block (excitation 450–490 nm, emission >515 nm; Leitz) for assessment of sinusoidal perfusion (2).
Quantitative assessment of platelet- and leukocyte-endothelial cell interactions was performed offline by computer-assisted analysis of the videotaped images (Capimage; Dr. Zeintl, Heidelberg, Germany). Rolling platelets and leukocytes were defined as cells crossing an imaginary perpendicular line through the vessel at a velocity markedly lower than the centerline velocity in the microvessel. Their numbers are given as cells per second per vessel cross-section. Platelets and leukocytes firmly attached to the endothelium for >20 s were counted as permanently adherent cells and expressed as number of cells per square millimeter endothelial surface (15). In sinusoids, the number of accumulated (stagnant) platelets was counted in the scanned acini and is given as 1/acinus. The sinusoidal perfusion rate was calculated as the percentage of perfused sinusoids per acinus (2).
Blood samples were taken from the carotid artery at the end of the experiment (60 min of reperfusion), immediately centrifuged at 2,000g for 10 min, and stored at −80°C. Serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) activities were measured with an automated analyzer (Hitachi 917; Roche-Boehringer, Mannheim, Germany) using standardized test systems (HiCo GOT and HiCo GPT; Roche-Boehringer) according to the recommendations of the International Federation of Clinical Chemistry.
Animals were randomly assigned to the following groups: sham-operated wild-type animals without ischemia (n = 7) and PARP+/+ (n = 7) and PARP−/− mice (n = 7) undergoing I/R (90 min/30 min). For analysis of platelet-endothelial cell interactions, PARP+/+ platelets were infused into PARP+/+ experimental animals, whereas the PARP−/− group received platelets isolated from PARP−/− animals.
The mRNA expression of the following genes was analyzed: P-selectin (CD62P), E-selectin (CD62E), ICAM-1 (CD54), VCAM-1 (CD106), and inducible NO synthase (iNOS).
Samples of frozen liver tissue were homogenized, and total RNA was extracted from the supernatants by using RNeasy spin columns (Qiagen). Total RNA was quantitated by measuring the optical density at 260 nm, and the integrity was checked by gel electrophoresis. cDNA was prepared from 2 μg of total RNA as described previously (26). An aliquot of cDNA (3 μl) was amplified by PCR on a gradient temperature cycler (RoboCycler; Stratagene, Heidelberg, Germany) as previously described (26). The oligonucleotide primers used (all from MWG-Biotech, Ebersberg, Germany) are presented in Table1. Ten-microliter samples of the amplified products at different cycle numbers were analyzed on a 1% agarose gel, stained with ethidium bromide, and visualized by UV illumination. Digital images of the agarose gels were densitometrically analyzed by using Bio-1D software (LTF-Labortechnik, Wasserburg, Germany). Net band intensity from the linear range was normalized to the housekeeping genes GAPDH and β-actin.
Immunostaining for P-selectin.
Tissue samples from the left liver lobe were taken 60 min after the onset of reperfusion. Paraffin sections were quenched with 0.5% H2O2 methanol solution to block production of endogenous peroxidase, incubated in 1.5% goat serum for 20 min to block nonspecific binding, and later incubated with a rabbit anti-mouse P-selectin antibody (Becton Dickinson, Heidelberg, Germany). Following incubation with the primary antibody, the sections were stained with commercially available peroxidase immunohistochemistry kits (Vectastain; Camon, Wiesbaden, Germany). An easily detectable reddish-brown-colored end product was obtained by development in H2O2/3-amino-9-ethylcarbazol. The sections were counterstained with Mayer's hemalaun. In each experimental group, seven sections from seven individual animals (10 observation fields per section) were examined by light microscopy (magnification ×400 and ×1,000 oil), and P-selectin expression was analyzed semiquantitatively in a blinded fashion by using a grading system of 0–2: 0, no staining; 1 weak staining; 2, strong staining.
Data analysis was performed with a statistical software package (SigmaStat for Windows; Jandel Scientific, Erkrath, Germany). The Kruskal-Wallis test followed by the Student-Newman-Keuls test was used for the estimation of stochastic probability in intergroup comparisons. Mean values ± SE are given. P values <0.05 were considered significant.
Leukocyte-endothelial cell interactions.
In postsinusoidal venules of sham-operated animals, leukocyte-endothelial cell interactions were nearly absent. After I/R in wild-type mice, the number of rolling and adherent leukocytes was significantly increased (10.1 ± 1.6 · mm−1 · s−1 and 372 ± 27/mm2) compared with the sham-operated group. In PARP−/− animals, however, both postischemic phenomena, leukocyte rolling as well as permanent adhesion, were significantly attenuated (Fig. 1).
Platelet-endothelial cell interactions.
In the sham-operated group, only a few platelets were found to be rolling on and firmly attached to the endothelium in presinusoidal arterioles (0.8 ± 0.4 · mm−1 · s−1 and 61 ± 17/mm2) and postsinusoidal venules (2.2 ± 0.3 · mm−1 · s−1 and 67 ± 23/mm2) as well as stagnant in sinusoids (2.9 ± 0.3/acinus). In contrast, I/R caused a significant increase in the number of rolling and adherent platelets in both arterioles and venules (∼6-fold), as well as in the number of platelets stagnant in sinusoids (∼3-fold) in wild-type mice compared with sham-operated animals. In PARP−/− mice, the postischemic increase in the number of rolling and adherent platelets was significantly lower, whereas the number of platelets stagnant in sinusoids did not differ between PARP+/+ and PARP−/− mice (Fig.2).
Sinusoidal perfusion was determined as a measure of microvascular I/R injury. The sinusoidal perfusion rate was 93 ± 2% in sham-operated animals, whereas it was significantly reduced to 72 ± 1% in wild-type mice after 90 min of normothermic ischemia followed by 30 min reperfusion. In PARP−/− animals, however, postischemic sinusoidal perfusion was significantly improved (Fig. 3).
Liver enzyme activities.
For assessment of hepatocellular damage in the postischemic liver, the AST/ALT serum activities were measured. After I/R, a significant increase in the liver enzyme activities was observed in wild-type mice compared with the sham-operated group. In contrast, the postischemic levels of AST/ALT serum activities were significantly lower in PARP−/− mice (Table2).
Immunostaining for P-selectin.
No staining for P-selectin was observed in sham-operated wild-type animals (score 0.2 ± 0.1). I/R induced an appearance of P-selectin on the endothelial surface of PARP+/+ mice, whereas less staining for P-selectin was found in liver tissue obtained from PARP−/− mice. The intensity of staining was 1.4 ± 0.1 for tissue sections in hepatic microvessels of wild-type mice and 0.4 ± 0.1 (P < 0.05 vs. I/R-PARP+/+) for sections of PARP−/− mice (Fig. 4). Although some of stained P-selectin could be clearly assigned to adherent platelets, most of the staining was on endothelial cells, as shown by the analysis of the tissue sections at magnification ×1,000 (Fig. 4,right) and confirmed by P-selectin immunostaining on blood-free flushed livers of PARP+/+ and PARP−/− mice in separate experiments (data not shown).
Quantification of transcript levels by competitive RT-PCR.
To investigate whether PARP affects the initial expression of adhesion molecules, RT-PCR for selected adhesion molecules was performed. As shown in Fig. 5, mRNA expression of E-selectin, P-selectin, and ICAM-1 was strongly induced after I/R in the liver of PARP+/+ mice compared with sham-operated PARP+/+ mice. In contrast, the postischemic increase in E-selectin and ICAM-1 in mRNA expression was markedly lower in the livers from PARP−/− mice than in those from PARP+/+ mice. It should, however, be noted that the RT-PCR method does not allow the differentiation between hepatocellular and endothelial expression of ICAM-1. The expression of VCAM-1 in the postischemic liver of PARP−/− mice was nearly absent compared with wild-type mice after I/R and even sham-operated animals. Interestingly, P-selectin mRNA expression did not differ between postischemic livers and PARP−/− mice. Since iNOS is suggested to mediate the expression of adhesion molecules during the inflammatory response, we also quantified the mRNA level of iNOS in the liver tissue from sham-operated animals as well as from PARP+/+ and PARP−/− animals undergoing I/R. The data show that iNOS mRNA was not expressed during I/R either in PARP+/+ or in PARP−/− mice. To confirm these results, we evaluated the baseline mRNA expression of adhesion molecules in the liver of control (nonoperated) PARP+/+ and PARP−/− mice and did not detect any difference between both types of mice (data not shown).
Activation and accumulation of inflammatory cells are the initial events of tissue injury during reperfusion (3). The present study as well as a variety of other reports have demonstrated that I/R of the liver induces leukocyte-endothelial cell interactions in the hepatic microvasculature (34). Recently, platelet-endothelial cell interactions have also been recognized as a prominent phenomenon that participates in the development of postischemic liver injury (10, 28). The analysis of platelet- and leukocyte-endothelial cell interactions in the present study shows that the postischemic increase in the number of rolling and adherent leukocytes and platelets was significantly lower in PARP−/− mice than in wild-type mice, indicating a triggering role of PARP for these cell-cell interactions. In sinusoids, however, the postischemic accumulation of platelets was not affected by targeted gene disruption of PARP. A crucial role of PARP in leukocyte infiltration was demonstrated in models of myocardial I/R as well as of mucosal colitis (40, 41). We have shown in the liver that PARP modulates both steps of I/R-induced leukocyte- and platelet-endothelial cell interactions: initial rolling and subsequent firm adhesion.
We assumed that PARP aggravates postischemic cell-cell interactions due to its involvement in the expression/translocation of adhesion molecules on the hepatic endothelium. P-selectin is stored in α-granules of platelets and Weibel-Palade bodies of endothelial cells and is translocated to the cell surface on stimulation (e.g., by hypoxia/reoxygenation), mediating leukocyte rolling and subsequent adhesion (23). Recently, we have shown that endothelial P-selectin is also essential for platelet-endothelial cell interactions in terminal arterioles and postsinusoidal venules of the postischemic liver (10). Because cell-cell interactions were affected in arterioles and venules, but not in sinusoids of PARP−/− mice (in sinusoids, P-selectin is not expressed; see Ref. 8), we suggest that the postischemic translocation of P-selectin is triggered by PARP. In fact, we observed less P-selectin staining on the endothelium of hepatic microvessels in the postischemic liver tissue from mice lacking PARP. Since the intracellular pool of P-selectin was not detectable by the method used for tissue fixation, these immunohistochemical results reflect the process of P-selectin translocation rather than de novo P-selectin synthesis. Our data are supported by investigations in the postischemic heart, which showed that P-selectin and ICAM-1 were nearly absent on the endothelial surface after 1 h of reperfusion in PARP−/− mice (40). I/R-stimulated P-selectin transcription, however, did not differ between PARP+/+ and PARP−/− mice, as shown by RT-PCR in our study. We therefore conclude that, during early reperfusion, PARP appears to influence rolling as well as subsequent firm adhesion of leukocytes and platelets through its participation in the process of P-selectin translocation.
ICAM-1 and VCAM-1 constitutively expressed on the surface of endothelial cells as well as E-selectin expressed de novo support the permanent adhesion of neutrophils during reperfusion (reviewed in Ref.14). Recently, it was shown that ICAM-1 also plays an essential role in platelet adhesion to the postischemic endothelium (16). In this study, we have semiquantitatively analyzed the gene transcription of ICAM-1, VCAM-1, and E-selectin in tissue homogenates from PARP+/+ and PARP−/− mice undergoing I/R. Although the basal expression of adhesion molecule mRNA did not differ between PARP+/+ and PARP−/− mice, the I/R-induced upregulation of adhesion molecule mRNA was markedly lower in PARP−/− mice. It is worth noting that this phenomenon cannot explain the attenuation of platelet and leukocyte adhesion observed in PARP−/− mice during early reperfusion, since 3–5 h would be required for de novo synthesis of the adhesion molecules. This fact, however, might become important for cell adhesion during later reperfusion.
Interestingly enough, the effect of PARP deficiency on cell-cell interactions does not seem to be caused by phenotypic differences in the expression of adhesion molecule counterligands, since FACS analyses of the expression of CD11b, CD18, and CD62L on blood neutrophils and lymphocytes revealed comparable expression of these receptors in both PARP+/+ and PARP−/− mice (unpublished observations).
The specific role of PARP in the regulation translocation/expression of adhesion molecules remains not fully understood. Translocation/expression of adhesion molecules can be initiated by reactive oxygen metabolites, proinflammatory mediators, and thrombin. Recently, PARP inhibition was shown to completely abrogate NO and cytokine release (TNF, IL-1) in the postischemic rat liver (7) as well as to suppress peroxynitrite- and TNF-α-induced surface expression of CD62P and ICAM-1 in human umbilical vein endothelial cells (40). Although PARP can potentially mediate the expression of adhesion molecules via stimulation of iNOS-derived NO production, iNOS mRNA expression during early reperfusion was not detected in our study with the highly sensitive RT-PCR method.
The transcription of adhesion molecules also might be enhanced by PARP. Although PARP itself is neither an integral part of the basal transcription machinery nor essential for transcription, this enzyme was recently shown to be an active component of the positive cofactor 1 and has the capacity to stimulate basal and activator-dependent transcription with respect to its DNA-binding properties (18). The expression of adhesion molecules may be also regulated by PARP through changes in cellular energy balance or directly through induction of chromatin relaxation, with the consequence that DNA becomes accessible to RNA polymerase (1, 18,25). PARP can poly-ADP ribosylate transcription factors and thereby modify their ability to bind to DNA. Furthermore, poly-ADP ribosylation may initiate the repair of DNA strand breaks, which, in turn, may interfere with transcription of damaged genes (29).
We have shown that PARP activation mediates the hepatic I/R damage not only by a direct impact through dramatic alterations in the levels of metabolic intermediates (NAD+ and ATP) but also due to aggravation of microcirculatory derangements. It became apparent that leukocytes and platelets already interact with the postischemic endothelium during the first minutes of reperfusion and, on activation, release reactive oxygen radicals, NO, and proinflammatory mediators (5, 22, 37). These mediators cause an increase in microvascular permeability, followed by interstitial edema and impairment of blood flow in sinusoids. Perfusion failure augments postischemic damage because focal ischemia and tissue hypoxia continue to persist despite reperfusion (33). Confirming this concept, the hepatic I/R injury was characterized in our experiments by sinusoidal perfusion failure and increased activities of the liver enzymes ALT/AST. After I/R in PARP-deficient animals, however, sinusoidal perfusion was ameliorated and ALT/AST activities were significantly lower than in PARP+/+ mice, demonstrating that PARP deficiency attenuated both microvascular and hepatocellular injury in the postischemic liver.
In summary, our in vivo results indicate that PARP contributes to hepatic microvascular injury by triggering the expression/translocation of adhesion molecules and modulating leukocyte/platelet-endothelial cell interactions. PARP deficiency prevents hepatic microvascular/cellular I/R damage. The concept that PARP regulates the microvascular mechanisms of I/R injury provides new insights into the pathogenesis of hepatic I/R.
We thank B. Lorenz, S. Münzing, and A. Schropp for technical assistance, Dr. C. Engelschalk (Institute of Clinical Chemistry, University of Munich) for measuring ALT/AST activities, and Dr. Z.-Q. Wang for providing PARP−/− mice.
The study was supported by the Deutsche Forschungsgemeinschaft (FOR 440).
Address for reprint requests and other correspondence: F. Krombach, Institute for Surgical Research, Univ. of Munich, Marchioninistr. 27, D-81377 Munich, Germany (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
May 22, 2002;10.1152/ajpgi.00085.2002
- Copyright © 2002 the American Physiological Society