# Peroxynitrite attenuates hepatic ischemia-reperfusion injury

Peitan Liu, Baohuan Xu, John Quilley, Patrick Y.-K. Wong

## Abstract

In the present study, we examined the effects of peroxynitrite on reperfusion injury using a rat model of hepatic ischemia-reperfusion (HI/R). The left and median lobes of the liver were subjected to 30 min of ischemia, followed by 4 h of reperfusion. Groups A and B rats were sham-operated controls that received vehicle or peroxynitrite;groups C and D rats were subjected to HI/R and received peroxynitrite or vehicle, respectively. A dose of 2 μmol/kg body wt of peroxynitrite, diluted in saline (pH 9.0, 4°C), was administered as a bolus through a portal vein catheter at 0, 60, and 120 min after reperfusion. Results showed that superoxide generation in the ischemic lobes of the liver and plasma alanine aminotransferase (ALT) activity of group C were decreased by 43% and 45%, respectively, compared with group D. Leukocyte accumulations in the ischemic lobes of liver and circulating leukocytes were decreased by 40% and 27%, respectively, in group C vs.D. The ratios of mRNA of P-selectin and intercellular adhesion molecule-1 (ICAM-1) to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA extracted from the ischemic lobes of the liver of group C were decreased compared with group D. There were no differences between the groups A andB in terms of plasma ALT activity, circulating leukocytes, superoxide generation, and leukocyte infiltration in the ischemic lobes of the liver. Moreover, hemodynamic parameters (i.e., mean arterial blood pressure, cardiac index, stroke index, and systemic vascular resistance) were not significantly different among groups B,C, and D. These results suggest that administration of peroxynitrite via the portal vein only has a local effect. Exogenous peroxynitrite at physiological concentrations attenuates leukocyte-endothelial interaction and reduces leukocyte infiltration. The mechanism of the reduction of leukocyte infiltration into ischemic lobes of the liver appears because of decreased expression of mRNA of P-selectin and ICAM-1. The net effect of administration of peroxynitrite may be to reduce adhesion molecule-mediated, leukocyte-dependent reperfusion injury.

• leukocyte
• nitric oxide
• P-selectin
• reverse transcription-polymerase chain reaction.

nitric oxide (NO) is a mediator with effects that may be likened to a double-edged sword (8, 21, 25, 32). One explanation for the detrimental effects of NO is the generation of peroxynitrite formed by the reaction of NO with superoxide (2, 20, 23, 31). NO reacts with superoxide at a rate constant of 6.7 × 109 M−1 · s−1 at pH 7.4 to form peroxynitrite (2, 31). Peroxynitrite is more cytotoxic than NO or superoxide in a variety of experimental systems, since it can decompose to nitronium ion (NO2 ) and hydroxyl radical (OH·), one of the most reactive oxygen species identified (3). Peroxynitrite can result in profound cellular injury and cell death (10,12). Using the models of endotoxin and hemorrhagic shock, Szabo et al. (33) reported that peroxynitrite is formed during endotoxemia and contributes to the cellular injury. Peroxynitrite can cause protein fragmentation via oxidative stress, and inactivation of important regulatory proteins may contribute to its toxicity (12, 18). This injury could be diminished by inhibition of NO production (11, 32).

In contrast with these studies, at the physiological concentrations (nM and low μM), peroxynitrite possesses a beneficial effect similar to NO (16, 29, 30). Nanomolar concentrations of peroxynitrite have been reported to exert effects that result in vascular relaxation and inhibition of platelet aggregation (13, 38); low micromolar concentrations of peroxynitrite can cause relaxation of isolated bovine pulmonary arteries concomitant with the production of NO (36). Furthermore, peroxynitrite inhibits leukocyte-endothelial cell interactions and exerts cytoprotective effects in myocardial ischemia-reperfusion (MI/R) injury (16,29).

The importance of polymorphonuclear neutrophil (PMN) recruitment to tissue injury is demonstrated by clinical findings that PMN numbers in bronchoalveolar lavage fluid correlate with mortality in patients with adult respiratory distress syndrome (24). The organ that harbors the greatest number of migrating PMNs is also the organ first and most commonly observed to fail (1, 26, 28). Activated PMNs can produce more than 50 toxins on the plasma membrane and in the intracellular granules. During the period of reperfusion, PMNs rapidly accumulate in the microvasculature of ischemic organs. The process of migration of PMNs from the circulation to sites of injured tissue occurs in the following sequence: 1) rolling adhesion, 2) firm attachment, and 3) migration through endothelial junctions. The adherence of PMNs to endothelial cells is prerequisite for PMN migration and subsequent tissue injury.

Because the evidence for peroxynitrite formation is indirect and based on correlation analysis in in vivo study, it is necessary to develop specific peroxynitrite probes, such as relatively stable peroxynitrite donors, for studying the pathogenesis of peroxynitrite. The recent commercial availability of peroxynitrite makes it possible to evaluate the profile of peroxynitrite in vivo. The present study was designed to test the hypothesis that exogenous peroxynitrite at physiological levels can inhibit P-selectin and intercellular adhesion molecule-1 (ICAM-1) mRNA expression and leukocyte infiltration and, thereby, reduce reperfusion injury in a rat model of hepatic I/R (HI/R).

## MATERIALS AND METHODS

#### Materials.

Fischer 344 rats (male, 275–300 g body weight) were purchased from Taconic Farm (Germantown, NY). The animals were given free access to food (Purina rodent chow J001) and water. The experimental protocols followed the criteria of the NIH “Guide for the Care and use of Laboratory Animals” and were approved by the Institutional Animal Care and Use Committee.

Alanine aminotransferase (ALT) activity was measured with a Sigma diagnostics kit. Nitrate reductase (from Aspergillusspecies), o-dianisidine, β-nicotinamide adenine dinucleotide phosphate, reduced form β-NADPH, and sodium nitrite were purchased from Sigma, St. Louis, MO. The RNA Stat-60 reagent, SuperScript II, Taq DNA polymerase, and GelMarker were purchased from GIBCO BRL, Life Technologies, Gaithersburg, MD, and Tel-Test, Friendswood, TX. Aliquot of peroxynitrite and vehicle (negative control) were purchased from Alexis.

#### Experimental protocol.

The experimental protocol for partial no-flow hepatic ischemia and measurement of associated hemodynamic variables has been previously described (21, 22). Briefly, under pentobarbital anesthesia (60 mg/kg ip), the trachea was cannulated (PE-240) to maintain a patent airway. A PE-90 catheter inserted into the right external jugular vein was connected to a blood pressure transducer (model P23AC; Statham, Hatorey, Puerto Rico) for the measurement of central venous pressure (CVP) by a Grass model 7D polygraph (Quincy, MA). The catheter inserted into the external jugular vein was also used for bolus saline injection (i.e., 200 μl) for the determination of cardiac output (CO). A 1.5-French thermistor probe (Columbus Instruments) was advanced into the right common carotid artery to the arch of the aorta. The position of the carotid thermistor probe was adjusted to ensure that a change in temperature of at least 0.3°C was recorded at the aortic arch when 200 μl of room temperature normal saline was injected into the right atrium. Polyethylene catheters (PE-50) filled with heparinized 0.9% NaCl (10 U heparin/1 ml saline) were inserted into the left femoral artery and left femoral vein for measurement of mean arterial blood pressure (MABP) and drug or vehicle infusion, respectively. The blood pressure transducer and thermistor were connected to a Cardiomax II cardiac output computer (Columbus Instruments) for measurement of MABP, CO, stroke volume (SV), heart rate (HR).

Twenty minutes after all surgical procedures were completed, the baseline MABP, CO, SV, HR, and CVP were recorded. A laparotomy was performed, and a catheter (PE-10) was inserted into the portal vein, whereupon the relevant branches of the hepatic artery and portal vein, supplying the left lateral and median lobes of the liver, were occluded with an atraumatic Glover bulldog clamp for 30 min. The remaining caudal three lobes retained an intact portal and arterial blood supply, as well as venous outflow, thereby preventing the development of intestinal venous hypertension. Reperfusion was initiated by removal of the clamp. The animal received 1 ml of the sterile saline intraperitoneally, and the wound was closed with 4-0 silk and clips. Group C rats received the freshly prepared, ice-cold peroxynitrite (ONOO) (2 μmol/kg; bolus, through the portal vein catheter) at time 0 and at 1 and 2 h of reperfusion (I/R + ONOO), whereas group Drats were given vehicle (I/R + vehicle) according to the same schedule. Animals in the sham-operated control group (group B) were subjected to identical surgery without occlusion of the blood vessels and injected with freshly prepared peroxynitrite in the same way togroup C. The group A rats were subjected to sham operation but were given saline. The hemodynamic parameters (MABP, CO, SV, HR, and CVP) were recorded at the beginning of ischemia and at different time points during reperfusion. Blood samples were obtained at 4 h of reperfusion for determination of ALT activities and leukocyte counts. Biopsies of the ischemic lobes of the liver were taken following 4 h of reperfusion for extraction of total RNA and measurement of superoxide generation. Pieces of liver were stored in 4% neutral-buffered paraformaldehyde for subsequent histological study.

#### Preparation of peroxynitrite aliquots.

Peroxynitrite purchased from Alexis with purity >90% was freshly prepared for each experiment. The concentration of peroxynitrite was monitored before use in each experiment by measuring the extinction coefficient at 302 nm after the addition of 0.5 ml of peroxynitrite to 3 ml of 1 N sodium hydroxide at pH 11. The concentrations of peroxynitrite were calculated by measuring the absorbance at 302 nm (E302nm = 1.670 mM−1 · cm−1). An aliquot of peroxynitrite at 2 μM concentration was freshly prepared by diluting of an appropriate volume of 2 mM in 5 ml of ice-cold pH 9.0 saline, and 2 μmol/kg peroxynitrite was injected through portal vein catheter. At the end of last injection, the extinction coefficient of peroxynitrite was again measured, revealing concentrations that were still >94% of initial levels. Peroxynitrite injected through this route directly contacts with hepatocytes, endothelial cells, and Kupffer cells. In the pilot experiments, injection of aliquot of peroxynitrite through a peripheral vein did not effectively alter reperfusion injury in our model, probably because of the short half-life of peroxynitrite (1 s at pH 7.4), which may be metabolized by the lungs before it reaches the liver.

The vehicle used in these experiments was provided by Alexis as a negative control. The vehicle is a decomposed form of peroxynitrite prepared from the same stock as the active form and contains the same concentrations of nitrite, hydrogen peroxide, and salt. However, the vehicle has no absorbance at 302 nm under alkaline conditions.

#### ALT activity.

Plasma ALT activities were measured with a Sigma test kit (DG 159-UV) and expressed as international units per liter.

#### Determination of circulating leukocytes.

Citrate-anticoagulated blood samples were obtained after 4 h of reperfusion. Fifty microliters of each sample was diluted 20-fold with 1% acetic acid solution to lyse red cells. Leukocytes were counted by light microscopy using a hemocytometer.

#### Superoxide assay.

Superoxide anion production in the liver sample from ischemic and nonischemic lobes was measured using the method described previously (22). Briefly, tissue samples (70–180 mg) were incubated in Krebs-bicarbonate buffer (pH 7.4) consisting of (in mM) 118 NaCl, 4.7 KCl, 1.5 CaCl2, 25 NaHCO3, 1.1 MgSO4, 1.2 KH2PO4, and 5.6 glucose. The tissues were gassed with 95% O2-5% CO2 for 30 min and placed in plastic scintillation vials containing 0.25 mM lucigenin in 1 ml of Krebs-bicarbonate buffer containing HEPES (pH 7.4). The chemiluminescence elicited by superoxide in the presence of lucigenin was measured using a scintillation counter (model Mark 5303; TmAnalytic, Elk Grove Village, IL). After 3 min of dark adaptation, vials containing only the cocktail (blanks) were counted three times for 6 s each time. The tissue samples were subsequently added to vials, allowed 3 min of dark adaptation, and counted twice. Since the half-life of superoxide is very short, the results reflect the production of superoxide by activated leukocytes.

#### RT-PCR amplification of mRNA.

Liver samples were snap frozen in liquid nitrogen and stored at −70°C until analysis. Total cellular RNA was isolated by homogenizing tissues with a Polytron homogenizer in RNA Stat-60 reagent (Tel-Test). Total RNA was extracted with chloroform, and samples were centrifuged at 12,000 g for 15 min at 4°C. The RNA was precipitated by isopropanol, and the pellet was dissolved in diethyl pyrocarbonate water (Sigma). Total RNA concentration was determined by spectrophotometric analysis at 260 nm, and 4 μg of total RNA was reverse transcribed into cDNA in 30 μl of reaction mixture containing Superscript II (GIBCO BRL), dNTP, and oligo(dT)12–18primers. The cDNA was amplified using specific primers with a Perkin-Elmer DNA Thermal Cycler 480. The amplification mixture contained 1 μl of 15 μM forward primer, 1 μl of 15 μM reverse primer, 5 μl of 10× buffer, 1.5 μl of 50 mM Mg2+, 5 μl of the reverse-transcribed cDNA samples, and 1 μl ofTaq polymerase. Primers were designed from the published cDNA sequences using the Oligo Primer Detection Program. The cDNA was amplified after determining the optimal number of amplification cycles within the exponential amplification phase for each primer set. Samples were denatured at 94°C for 5 min followed by 20 cycles for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and 25 cycles for P-selectin and ICAM-1. Each cycle consisted of 94°C for 45 s, 60°C for 60 s, and 72°C for 90 s. After amplification, the sample (5 μl) was separated on a 2% agarose gel containing 0.3 μg/ml (0.003%) of ethidium bromide, and bands were visualized and photographed using ultraviolet transillumination. The size of each PCR product was determined by comparison with a standard DNA size marker. Semiquantitative assessment of gene expression was performed using the Image Master VDS program (Pharmacia Biotech). The designed primer sequences are shown belowSense primerAntisense primerP­selectin5­TGTATCCAGCCTCTTGGGCATTCC­35­TGGGACAGGAAGTGATGTTACACC­3ICAM­15­ACAGACACTAGAGGAGTGAGCAGG­35­GTGAGCGTCCATATTTAGGCATGG­3GAPDH5­GGTGAAGGTCGGTGTCAACGGATT­35­GATGCCAAAGTTGTCATGGATGACC­3

#### Statistical analysis.

Statistical significance for multigroup comparisons was determined using analysis of variance (ANOVA, Sigma Stat Program) for multiple group comparisons with repeated measurement. If a significant F value was obtained, then the group means were analyzed using the Bonferroni multiple comparison test when P < 0.05 was considered significant. The results are means ± SE.

## RESULTS

#### Plasma ALT activity.

The period of 30-min ischemia of left and median liver lobes followed by 4 h of reperfusion induced a pronounced elevation of plasma ALT activity that was 28.5-fold higher than the sham group (1,369 ± 147 vs. 48 ± 5 U/l). Injection of peroxynitrite (2 μmol/kg; through the portal vein catheter) at time 0 and at 1 and 2 h of reperfusion (group C) attenuated the reperfusion injury, reflected by the reduction of plasma ALT activity (594 ± 65 U/l) at 4 h of reperfusion (P < 0.05 relative to I/R +vehicle group) (Fig. 1). There was no statistical difference between groups A andB at 4 h of injection of saline or peroxynitrite.

Fig. 1.

#### Circulating leukocytes.

Elevation of circulating leukocytes, suggesting a systemic inflammatory response, was observed after hepatic reperfusion injury in group D rats compared with group A animals (7,027 ± 826 vs. 3,587 ± 269 cells/mm3) and reflected a 1.37-fold increase compared with group C (5,115 + 519 cells/mm3) at 4 h of reperfusion. There was no statistically significant difference between groups A andB (Fig. 3).

Fig. 3.

#### P-selectin and ICAM-1 mRNA expression.

Total mRNA was extracted from the livers of group B and from ischemic lobes of livers of groups C and D at 4 h of reperfusion. Expression of P-selectin and ICAM-1 mRNA was studied by RT-PCR amplification following separation of products by gel electrophoresis (Fig. 5). A semiquantitative analysis of RT-PCR was performed with the Image Master VDS program (Pharmacia Biotech) and normalized by assessment of the expression of GAPDH. P-selectin mRNA expression could not be detected in group B. However, expression of P-selectin and ICAM-1 genes was obtained from group D rats. Administration of peroxynitrite to rats subjected to HI/R resulted in an attenuation of gene expression for P-selectin (P = 0.108, compared with group D) and ICAM-1 (P < 0.05, compared with group D).

Fig. 5.

A: ethidium bromide stained agarose gel showing PCR products from amplified rat hepatic RNA. HI/R was carried out as described in materials and methods, and the ischemic liver lobes were obtained at 4 h of reperfusion. RNA was extracted from hepatic tissue, then reverse transcribed and amplified using primers selected from published cDNA sequences. A semiquantitative analysis of the results of RT-PCR was performed with the Image Master VDS program (Pharmacia Biotech). With this computer program, the electrophoretic bands were normalized for glyceraldehyde-3-phosphate dehydrogenase (GAPDH; lanes 2, 5, and8) gene expression. The normalized ratios for P-selectin (lanes 3, 6, and 9) and intercellular adhesion molecule-1 (ICAM-1; lanes 4, 7, and10) mRNA expression in the experiments were averaged, and the results are shown in B and C, respectively.Lane 1, marker; lanes 2–4, sham control;lanes 5–7, HI/R + ONOO; and lanes 8–10, HI/R + vehicle. Data are means ± SE of 6 animals. *P < 0.05 compared with rats of group B.# P < 0.05 compared with rats ofgroup C.

#### MABP and systemic vascular resistance index.

Systemic vascular resistance index (SVRI) was calculated as (MABP − CVP)/CI, and the results are illustrated in Fig.6 B. The initial values of MABP and SVRI were not significantly different among groups B,C, and D rats (Fig. 6, A andB). Injection of peroxynitrite (2 μmol/kg; through the portal vein catheter) to rats of groups B and Cdid not have an effect on vascular tone as evidenced by the lack of statistically significant differences in MABP and SVRI among the three experimental groups.

Fig. 6.

Time course of changes in mean arterial blood pressure (A) and systemic vascular resistance index (B) in 3 experimental groups. Data are means ± SE of 5 animals per group. There was no statistical significance among 3 experimental groups in the various time points.

#### Cardiac index and stroke volume index.

Cardiac index (CI) and stroke volume index (SVI) were calculated as CO (or stroke volume) per 100 g of body weight of rat. The initial values of CI and SVI were not significantly different amonggroups B, C, and D (Fig.7, A and B). Injection of peroxynitrite (2 μmol/kg; through the portal vein catheter) to rats of groups B and C did not affect cardiac function, as evidenced by a lack of statistically significant differences in CI and SVI among the three experimental groups.

Fig. 7.

Time course of changes in cardiac index (A) and stroke volume index (B) in 3 experimental groups. Data are means ± SE of 5 animals per group. There was no statistical significance among 3 experimental groups in the various time points.

## DISCUSSION

In the present study, we examined the hypothesis that peroxynitrite at physiological levels (nM or low μM) can inhibit mRNA expression of P-selectin and ICAM-1 (Fig. 5) and may, thereby, reduce leukocyte-dependent reperfusion injury. Peroxynitrite administered via the portal vein to the ischemic lobe of liver reduced plasma ALT activity (Fig. 1), hepatic superoxide generation (Fig. 2), and leukocyte accumulation in the ischemic lobes of the liver (Fig. 4). The attenuation of leukocyte accumulation in the ischemic lobes of the liver may be related to inhibition of P-selectin and ICAM-1 mRNA expression by peroxynitrite. In contrast, no statistically significant differences were found between sham rats + vehicle and sham rats + peroxynitrite in terms of plasma ALT activity, circulating leukocyte counts, superoxide generation and leukocyte infiltration in the liver, indicating an effect of peroxynitrite limited to rats subjected to reperfusion injury. Moreover, the beneficial effect of peroxynitrite cannot be attributed to hemodynamic changes because MABP, SVRI, CI, and SVI were not different among the various groups of rats.

NO has a number of physiological roles, including smooth muscle relaxation and cellular communication. In the vascular system, biosynthesis of NO at physiological concentrations (nM to low μM) regulates organ blood flow and leukocyte-endothelial interactions (9), inhibition of platelet aggregation (38), and inhibition of neutrophil infiltration (9,14). On the other hand, activation of inducible NO synthase (iNOS) produces a sustained elevation of NO levels (high μM) (2, 8, 31). Pathological concentrations of NO exert detrimental effects possibly via the generation of peroxynitrite formed by a reaction of NO with superoxide (19, 23, 31).

At sites of inflammation, increased production of NO is an essential compensatory response for maintaining organ blood flow (4,22) and inhibiting leukocyte adhesion and platelet aggregation (14, 15, 21, 22). The simultaneous increase in NO and superoxide enhances the formation of peroxynitrite, which induces thiol group inactivation, protein fragmentation (33), and alteration of DNA synthesis leading to cell death. Additionally, peroxynitrite is highly bactericidal toEscherichia coli (3) and can cause oxidation of sulfhydryl groups, as well as protein strand breakage (10) and cell apoptosis (18) at high micromolar concentrations.

The results of the present study are consistent with those of other studies showing that peroxynitrite at physiological concentrations (nM to low μM) has several beneficial effects similar to those of NO (16, 29, 30). Nossuli et al. (29), for example, have reported that infusion of peroxynitrite at a concentration of 1 μM reduces myocardial infarct size and preserves the coronary endothelium in a cat model of myocardial ischemia and reperfusion. Our previous work demonstrated that administration ofN ω-nitro-l-arginine methyl ester (l-NAME) attenuated peroxynitrite formation but enhanced reperfusion injury in a rat model of HI/R, supporting the idea that peroxynitrite may play an anti-inflammatory role similar to NO in acute inflammation (21). Additionally, Wu et al. (37) have demonstrated that peroxynitrite relaxes pulmonary arteries in vitro concomitant with the production of NO. Thus peroxynitrite, at low micromolar concentrations, inhibits platelet aggregation in vitro (38) and produces vasorelaxation in dog coronary arteries (19). Furthermore, peroxynitrite can produce S-nitrosothiols, which can stimulate guanylyl cyclase and release NO (5, 34, 37). Physiologically relevant concentrations of peroxynitrite significantly attenuated neutrophil-endothelium interactions and decreased the extension of necrotic tissue in an in vivo model of MI/R injury (16,30).

Peroxynitrite can influence the expression of adhesion molecules, such as P-selectin and ICAM-1. P-selectin is stored in Weibel-Palade bodies of endothelial cells, but it rapidly mobilizes to the surface of endothelial cells following exposure to complement fragments, thrombin, or histamine. The expression of P-selectin peaks within 10 min after stimulation. P-selectin recognizes sialyl Lewisx as a ligand and binds to specific carbohydrates on neutrophils. The binding results in deceleration of leukocytes flow and rolling along the endothelial cells. Our results indicate that administration of peroxynitrite (2 μmol/kg) attenuates the expression of mRNA of P-selectin and ICAM-1. These findings are in agreement with the work reported by Nossuli et al. (30) that administration of peroxynitrite (1 μmol/l) intraventricularly significantly reduced adherence of neutrophils to the ischemic-reperfused left anterior descending coronary endothelium in a cat model of MI/R. Immunochemical staining indicated that the percentage of coronary venules staining positive for P-selectin was significantly reduced in animals with MI/R + ONOO, compared with MI/R + vehicle (30). Lefer et al. (16) also reported that peroxynitrite reduced PMN adhesion to thrombin-stimulated superior mesenteric artery segments by 58% at the concentration of 100 nM, and by 63% at the concentration of 1 μM, when vs. thrombin alone in vitro. ICAM-1, a member of the immunoglobulin supergene family, is an important ligand/receptor for CD11b/CD18. ICAM-1 is expressed constitutively on endothelial cells, and upregulated in response to tumor necrosis factor-α, interleukin-1β, and lipopolysaccharide (7, 35). Antibodies to either CD18 (17) or ICAM-1 (35) in an in vivo experiment have been shown to attenuate neutrophil-dependent injury.

Our results show that hemodynamic parameters (e.g., MABP, SVRI, CI, and SVI) (Figs. 6 and 7) were not different among the various groups during the period of reperfusion. These results suggest that there was no direct or indirect effect of peroxynitrite via release of NO on the cardiovascular system. In our previous study, animals receivedS-nitroso-N-acetylpenicillamine (SNAP), an NO donor, which increased plasma NO concentrations, and therefore improved hemodynamics in the HI/R model (22). In the present studies, plasma NO levels, based on the measurement of nitrate/nitrite decomposed from NO, could not be measured, because synthesis of peroxynitrite from the reaction of superoxide with NO produces high levels of residuals of nitrite/nitrate in the aliquot of peroxynitrite.

The previously reported deleterious effects of peroxynitrite, including oxidation of sulfhydryl groups, DNA strand and protein breakage, and cell apoptosis, may reflect the use of high concentrations. Thus much of this previous work is based on the exposure of different types of cultured cells to high micromolar to millimolar concentrations of peroxynitrite (50- to 150-fold higher than pathophysiological levels). These high levels of peroxynitrite and the absence of antioxidants may have limited relevance to the in vivo situations because the formation of peroxynitrite in vivo is probably limited by several mechanisms. First, Miles et al. (23) have shown that peroxynitrite forms optimally from equimolar concentrations of NO and superoxide. The inequality of either precursor greatly limits production of peroxynitrite. NO is formed normally at 1–20 nM. Although in pathophysiological conditions higher levels of peroxynitrite can be formed (27, 31), the elevated levels were also limited to the nanomolar range. For example, Wang and Zweier (36) have demonstrated that rat hearts subjected to I/R produce <100 nmol/l peroxynitrite. Therefore, in the pathological conditions in vivo, the formation of peroxynitrite is unlikely to achieve high micromolar to millimolar concentrations. Second, many antioxidants and anti-inflammatory mediators are present in in vivo situations. The concentrations and toxicity of peroxynitrite are strongly influenced by the presence of these compounds. In the presence of plasma, protein, glucose, and glutathione, peroxynitrite will form intermediates with low toxicity. The extremely important detoxification of these chemical compounds was demonstrated by Denicola et al. (6) who reported that peroxynitrite interacts with the normal bicarbonate buffer system in human plasma and becomes much less toxic even at high micromolar concentrations. In addition, peroxynitrite can be scavenged by many compounds, including uric acid, cysteine, glutathione, ascorbic acid, deferoxamine, and vitamin E. Finally, with a half-life about 1 s under physiological conditions, peroxynitrite is unlikely to accumulate in vivo at concentrations greater than 1–5 μM (23, 31). Therefore, the concentrations of peroxynitrite in vivo would not exceed the low micromolar range (i.e., 2–5 μM).

In conclusion, peroxynitrite at nanomolar and low micromolar concentrations attenuated reperfusion injury similar to that of NO in a rat model of HI/R (21, 22). The mechanism for the effects exerted by peroxynitrite may involve its ability to modulate adhesion molecule expression in the ischemic lobes of liver and thereby attenuate leukocyte-dependent I/R injury.

## Acknowledgments

We are very grateful to Dr. Carl E. Hock for helpful discussion during the course of these experimental studies.

## Footnotes

• Address for reprint requests and other correspondence: P. Liu, Dept. of Cell Biology, UMDNJ-School of Osteopathic Medicine, 2 Medical Center Drive, Stratford, NJ 08084 (E-mail: liupe{at}umdnj.edu).

• 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.

View Abstract