Resuscitation with Ringer's Ethyl Pyruvate Solution Prolongs Survival and Modulates Plasma Cytokine and Nitrite/Nitrate Concentrations in a Rat Model of Lipopolysaccharide-Induced Shock

Abstract

The glycolytic intermediate, pyruvate, is capable of scavenging reactive oxygen species (ROS). However, this compound is relatively unstable and hence is not useful as a therapeutic agent. Ethyl pyruvate, a simple derivative of pyruvate, appears to be more stable, and when formulated in a calcium-containing Ringer's-type balanced salt solution (REPS), has been shown to be salutary in rat models of two pathophysiological conditions—mesenteric ischemia/reperfusion and hemorrhagic shock/resuscitation—that are thought to be mediated, at least in part, by ROS. Because ROS also have been implicated in the pathogenesis of lipopolysaccharide (LPS)-induced shock, we carried out a series of experiments to determine if REPS is beneficial in this condition. Anesthetized rats were challenged with intravenous LPS (20 mg/kg). When mean arterial pressure (MAP) decreased to 60 mmHg, 3- to 5-mL boluses of either REPS (n = 10) or Ringer's lactate solution (RLS; n = 10) were infused as needed to prevent MAP from decreasing further. By design, the maximal volume of fluid infused was 7 mL/kg. Resuscitation with REPS as compared with RLS prolonged survival time (498 ± 48 min vs. 362 ± 30 min;P = 0.0014). Resuscitation with REPS as compared with RLS also was associated with significantly lower circulating concentrations of nitrite/nitrate and interleukin (IL)-6 and higher plasma levels of IL-10. These data support the view that delayed treatment with REPS modulates the inflammatory response to LPS, and prolongs survival time in a lethal model of endotoxic shock.

INTRODUCTION

Reactive species of oxygen (ROS) have been implicated as being important mediators in a variety of pathological conditions, including thermal injury (¹), hemorrhagic shock/resuscitation (²), and mesenteric ischemia/reperfusion (³). Examples of ROS of biological importance include superoxide radical anion (O₂^−.), hydrogen peroxide (H₂O₂), hydroxyl radical (OH^.), and peroxynitrite (ONOO⁻). Although O₂^−. is only moderately reactive, it can be converted by the enzyme superoxide dismutase to the more reactive species, H₂O₂. Both O₂^−. and H₂O₂ are capable of reacting with nitric oxide (NO^.) to form the highly reactive moiety, ONOO⁻ (⁴). In addition, in a series of reactions catalyzed by ionized iron, O₂^−. and H₂O₂ can interact to form another extremely reactive species, OH^. (⁵).

Pyruvate, a key intermediate in cellular metabolism, is also an effective scavenger of ROS (⁶). Pyruvic acid rapidly undergoes nonenzymatic decarboxylation in the presence of H₂O₂ to form acetate, carbon dioxide, and water. Recently, pyruvate also has been shown to be capable of scavenging OH^. (⁷). Administration of exogenous pyruvate has been shown to be salutary in numerous animal models of redox-mediated tissue or organ injury (^7–9).

However, the usefulness of pyruvate as a therapeutic agent is limited by its very poor stability in solution (¹⁰). When dissolved in an aqueous solvent, pyruvate spontaneously undergoes a series of condensation and cyclization reactions that yield a variety of products, some of which may be toxic (¹¹). To circumvent this issue, Sims et al. (¹²) formulated a more stable derivative of pyruvic acid, namely ethyl pyruvate, in a calcium- and potassium-containing balanced salt solution. These investigators showed that treatment with this fluid, called Ringer's ethyl pyruvate solution (REPS), ameliorates structural and functional damage to the intestinal mucosa caused by mesenteric ischemia and reperfusion in rats (¹²). Subsequently, Tawadrous et al. (¹³) showed that resuscitation with REPS instead of Ringer's lactate solution (RLS) prolongs survival and decreases hepatic lipid peroxidation, a marker of redox stress, in rats subjected to hemorrhagic shock.

Redox-mediated phenomena have been implicated in the pathogenesis of sepsis and septic shock. Treatment with agents that scavenge ROS has been shown to be beneficial in various animal models of sepsis (^14,15), lipopolysaccharide (LPS)-induced acute lung injury (¹⁶), and LPS-induced shock, organ injury, and mortality (^17–20). In some of these studies, treatment with ROS scavengers was shown to limit LPS-induced production of tumor necrosis factor (TNF) or other cytokines (^18,19). Prompted by these observations, we sought to test the hypothesis that treatment with REPS would ameliorate shock and modulate the release of inflammatory mediators in a rat model of LPS-induced hypotension.

MATERIALS AND METHODS

This research protocol complied with the regulations regarding animal care as published by the National Institutes of Health and was approved by the Institutional Animal Use and Care Committee of the University of Pittsburgh.

Surgical preparation

Adult male Sprague-Dawley rats weighing 450 to 550 g were purchased from Hilltop farms (Somerset, PA) and were maintained at the University of Pittsburgh Animal Research Center with a 12-h light-dark cycle and free access to standard laboratory chow and water. The rats (450–550 g) were anesthetized with intraperitoneal sodium pentobarbital (40 mg/kg). The animals were endotracheally intubated with a beveled 16-guage angiocatheter and were ventilated with room air using a Harvard rodent ventilator (Holliston, MA) with a tidal volume of 10 mL/kg and a frequency sufficient to maintain an arterial PCO₂ between 35 and 45 mmHg. Arterial blood gases were monitored frequently using an ABL-725 blood gas analyzer (Randiometer, Copenhagen, Denmark). The right femoral vein and right carotid artery were isolated by dissection and were cannulated with PE 90 catheters. The arterial catheter was used for blood sampling and continuous monitoring of arterial pressure. The venous catheter was used for administration of LPS and fluid resuscitation.

Experimental protocol

Before the administration of LPS, the animals were maintained in a steady state as defined by stable mean arterial pressure (MAP) and arterial blood gas values for at least 30 min. Arterial pressure was measured continuously and was recorded in real time on a strip-chart recorder (Gould, Cleveland, OH). At T = 0 min, Escherichia coli LPS (serotype O111:B4; Sigma, St. Louis, MO; 20 mg/kg) was administered intravenously. When MAP decreased to <60 mmHg, the rats were randomly assigned to one of two treatment groups (n = 10 each) designated RLS and REPS. Animals in the RLS group were resuscitated with 3- to 5-mL boluses of RLS as needed to maintain an MAP of ≥60 mmHg until a total volume equal to 7% of the rat's body weight was administered. No further fluid therapy was provided after this point. Treatment in the other group was exactly the same, except that REPS was used as the resuscitation fluid. RLS was obtained from a commercial source (Baxter, Deerfield, IL) and contained 109 mM NaCl, 4 mM KCl, 2.7 mM CaCl₂, and 28 mM sodium L-lactate. pH was not adjusted and ranged from 6.0 to 7.5. REPS contained 130 mM NaCl, 4 mM KCl, 2.7 mM CaCl₂, and 28 mM ethyl pyruvate (pH 7.0–7.2). Blood samples for determinations of arterial blood gases, hemoglobin, ionized calcium, and measurements of circulating concentrations of inflammatory mediators were obtained at baseline (i.e., after surgical preparation but prior to injection of LPS) and at T = 3 h and T = 6 h after injection of LPS. Blood samples for measurements of inflammatory mediators were collected into iced tubes. The samples were centrifuged (2000 g for 10 min). The plasma was aspirated and frozen at −80°C until assayed.

Cytokine and nitrite/nitrate measurements

Plasma concentrations of immunoreactive TNF, interleukin (IL)-6, and IL-10 were determined using ELISA kits from R&D Systems (Minneapolis, MN). All cytokine assays were performed using the methods recommended by the manufactures. Plasma concentrations of total nitrite/nitrate were determined using the cadmium-reduction method followed by Greiss-reaction (Fluka Chemicals, Milwaukee, WI) (²¹). The plates for the nitrate/nitrite assay were read using a MRX microplate reader. (Dynex Technologies, Chantilly, VA) at 550 nm. NaNO₂ was used to generate a standard curve.

Statistical methods

Results are expressed as means ± SD. Differences in survival times between the RLS and REPS groups were analyzed using the log-rank test. Data from the measurements of plasma concentrations of TNF, IL-6, IL-10, nitrite/nitrate, rate of resuscitation, and MAP were analyzed using two-way analysis of variance (ANOVA) for repeated measures followed by post hoc testing using the Student-Neumann-Keuls test. Differences were considered significant for P < 0.05.

RESULTS

Resuscitation with REPS prolongs survival

Administration of LPS resulted in the development of sustained hypotension, beginning typically about 15 min after the injection of endotoxin. Resuscitation with RLS or REPS was initiated as soon as MAP decreased to less than 60 mmHg. Although all rats died in both treatment groups, survival times were significantly longer in the in the group resuscitated with REPS compared with the group resuscitated with RLS (498 ± 48 min vs. 362 ± 30 min, respectively;P = 0.0014;Fig. 1). The groups received similar volumes of resuscitation fluid (36 ± 1.7 mL in the RLS group vs. 33 ± 4.4 mL in the REPS group). Also, the timing of completion of fluid resuscitation (150 ± 107 min in the RLS group vs. 133 ± 96 min in the REPS group) and the rate of resuscitation (Fig. 2, bottom panel) were not different between groups. Although there was no difference between the groups in the baseline MAP, there was a statistically significant difference in MAP over time in both groups. MAP was also significantly higher in the REPS group as compared with the RLS group at T = 5 and T = 6 h (Fig. 2, top panel). There were no differences over time between the two groups with respect to blood hemoglobin concentration or plasma ionized calcium concentration (data not shown).

Fig. 1: Comparison of Kaplan-Meier survival curves for endotoxic rats resuscitated with RLS (n = 10; solid line) or REPS (n = 10; dashed line). The significance of the difference in survival times was assessed using the log-rank test.

Fig. 2: Comparison of MAP (top panel) and volume of intravenous fluid administered over 30- to 60-min intervals (bottom panel) in endotoxic rats resuscitated with either RLS (n = 10; •) or REPS (n = 10; ▪). Each rat that received fluid resuscitation during a given 30- or 60-min period is indicated by a circle (RLS group) or a square (REPS group) at the bottom of the figure in the top panel. In the bottom panel, open bars indicate the volume received by the animals in the RLS group and the closed bars indicate the volume received by the animals in the REPS group. There was no difference between the groups in the rate of resuscitation. One animal in each group died by T = 4 h and one additional animal died in each group by T = 5 h. The values in the figure indicate mean ± SD. ^† P < 0.05 for the time-matched contrast between groups. There was no difference between the groups in the mean time at which fluid resuscitation was completed (150 ± 107 min in the RLS group vs. 133 ± 96 min in the REPS group).

Resuscitation with REPS modulates production of inflammatory mediators

Plasma concentrations of TNF increased significantly following administration of LPS in both groups (Fig. 3A). The circulating concentrations of this cytokine were similar in both groups at all time points examined. Baseline plasma concentrations of IL-10 were similar in the two groups (Fig. 3B). After injection of LPS, circulating levels of IL-10 increased significantly in both groups. However, the concentration of IL-10 was significantly greater at T = 3 h in the REPS group as compared with the RLS group. A similar trend was seen at T = 6 h, although the between-group difference for IL-10 concentration at this time point did not achieve statistical significance. Baseline IL-6 levels were similar in the two groups (Fig. 3C). After the injection of LPS, IL-6 concentrations increased significantly in both groups. However, compared with the RLS group, the group resuscitated with REPS had significantly lower IL-6 levels at T = 3 and 6 h. Circulating concentrations of nitrite/nitrate were similar in the two groups at baseline (Fig. 3D). Although injection of LPS significantly increased plasma nitrite/nitrate levels in both groups, resuscitation with REPS instead of RLS was associated with significantly lower nitrite/nitrate concentrations at both T = 3 and T = 6 h.

Fig. 3: Comparison of plasma concentrations of TNF (a), IL-10 (b), IL-6 (c), and nitrite/nitrate (d) in endotoxic rats resuscitated with either RLS (n = 10; open bars) or REPS (n = 10; solid bars). An asterisk indicates P < 0.05 vs. the baseline value for the same group. ^† P < 0.05 for the time-matched between group contrast. The values in the figure indicate mean ± SD.

Resuscitation with REPS ameliorates the development of hyperlactatemia

Blood lactate concentrations were similar in the two groups at baseline (Fig. 4). After the injection of LPS, blood lactate concentrations increased significantly in both treatment groups, albeit to significantly lesser extent in the REPS group as compared with the RLS group at both T = 3 and T = 6 h. There were no significant differences in blood pH pCO₂ or calculated base deficit between the two treatment groups at any time points (Table 1).

Table 1: The various arterial blood gas parameters of the animals in both groups at times 0, 3, and 6 h. All values represent the mean ± SD

Fig. 4: Comparison of blood lactate concentrations in endotoxic rats resuscitated with either RLS (n = 10; open bars) or REPS (n = 10; solid bars). An asterisk indicates P < 0.05 vs. the baseline value for the same group. ^† P < 0.05 for the time-matched between group contrast. The values in the figure indicate mean ± SD.

DISCUSSION

Numerous prior studies have documented that pyruvate can protect against oxidant-mediated damage to cells and tissues (^8,22). Basic principles of organic chemistry suggest that ethyl pyruvate should be more stable in solution than the parent α-ketoacid, and preliminary studies suggest that this prediction holds true (A. Ajami, personal communication). In addition, ethyl pyruvate, being considerably more lipophilic than pyruvate anion, should enter cells much more readily than the parent compound. Published studies using a related compound, methyl pyruvate, support this view (²³).

This study used an overwhelmingly lethal model of acute endotoxemia in rats. Our primary objective was to determine whether delayed treatment with a balanced salt solution containing ethyl pyruvate (i.e., REPS) instead of RLS could prolong survival in a model of endotoxemia that is associated with profound hypotension such as occurs commonly in patients with septic shock. Treatment with the control or experimental resuscitation fluid (i.e., RLS or REPS, respectively) was delayed until well after the onset of hypotension, mimicking the way fluid therapy is often used in the clinical setting. Using this experimental design, we showed that treatment with REPS significantly prolonged survival time.

Treatment with REPS had no discernable effect on circulating concentrations of the “alarm” cytokine, TNF, in endotoxemic rats. In view of our experimental design, this result was to be expected. In most experimental models of acute endotoxemia, TNF is released into the circulation as a monophasic spike that becomes detectable soon after injection of LPS and typically peaks about 60 min later. In the present study, treatment with REPS or RLS was not initiated until about 15 min after the injection of LPS. Thus, infusion of the experimental agent was delayed until after the initiation of the release TNF. Accordingly, it is not surprising that TNF release was not affected.

Despite the absence of an effect on circulating concentrations of TNF, treatment with REPS modulated plasma levels of several other inflammatory mediators, namely IL-6, IL-10, and NO as assessed by circulating levels of nitrite/nitrate. The observed effect of REPS on IL-6 release was somewhat surprising because some studies suggest that production of this cytokine is triggered by the prior release of TNF. For example, Ghezzi et al. (²⁴) reported that treatment with an anti-TNF antibody inhibited IL-6 production by LPS-stimulated macrophages in vitro and by mice challenged with intraperitoneal LPS in vivo. Similar findings have been reported by other investigators using a variety of animal models (²⁵). Other investigators, however, have obtained data suggesting that LPS-induced IL-6 release occurs via mechanisms that are largely independent of TNF production. Thus, Vasilescu et al. (²⁶) reported that adding recombinant TNF to samples of human whole blood failed to trigger the release of IL-6, whereas addition of LPS to whole blood samples resulted in high levels of IL-6 even in the presence of an anti-TNF monoclonal antibody. More recently, in a series of in vitro studies, Panesar et al. (²⁷) showed that LPS stimulates IL-6 production by hepatocytes irrespective of the presence or absence of an anti-TNF antibody in the cultures. The data obtained in the present study tend to support the view that at least some of the IL-6 that is produced in rats after injection of LPS is triggered by events that are both independent of TNF production and are capable of being inhibited by ethyl pyruvate.

What pharmacological property of ethyl pyruvate might account for the observed modulation of LPS-induced IL-6, IL-10 and NO release that was observed in the group of rats treated with REPS? Although the data presented herein are insufficient to directly address this question, a plausible hypothesis can be put forward. In a previous study, we showed that resuscitation with REPS prevented hepatic lipid peroxidation in rats subjected to hemorrhagic shock (¹³). Previously, Varma and colleagues (²⁸) showed that ethyl pyruvate is capable of protecting the lens of the eye from redox-mediated stress caused by exogenous administration of compounds, such as menadione, that foster the formation of ROS. Together, these data support the view that ethyl pyruvate is capable of scavenging ROS. Other studies have shown that other ROS scavengers, such as N-acetylcysteine (NAC), dimethyl sulfoxide, quercetin, and phenyl-N-tert-butyl-nitrone (PBN), are capable of modulating LPS-induced cytokine or chemokine production and/or inducible NO synthase (iNOS) induction in various in vitro or in vivo systems (^19,29–31). Thus, it seems plausible that scavenging of ROS is responsible for the beneficial actions of REPS observed in our study.

Particularly pertinent to the results obtained in the present study are the observations by Yoshida and colleagues (³²) that ROS can trigger IL-6 release by certain cell types, and that pretreatment with the ROS scavenger, PBN, down-regulates hepatic LPS-IL-6 mRNA expression 3 h after endotoxin administration in rats (³¹). These data support the view that oxidant stress is capable of triggering IL-6 release and that using pharmacological agents to scavenge ROS can inhibit IL-6 production induced by LPS. The data obtained herein using REPS are consistent with this view.

Also noteworthy in the context of the results presented here is the recent observation by Bergamini et al. (²⁹) that treatment with NAC down-regulates iNOS induction and plasma nitrite/nitrate levels in rats challenged with LPS. These findings are similar to those reported here. Interestingly, Leach et al. (²⁰) showed that another ROS scavenger, tempol, failed to affect circulating nitrite/levels in a rodent model of acute endotoxemia. Thus, the pharmacological anti-inflammatory effects of various ROS scavengers are not entirely consistent across this broad class of agents, but seem to be at least somewhat specific for each individual compound.

Whereas treatment with REPS decreased plasma levels of IL-6 and nitrite/nitrate, resuscitation with this agent had the opposite effect on circulating concentrations of IL-10. This finding is of considerable interest because the effects of IL-10 on the innate immune response are largely counterregulatory; i.e., the predominant actions of IL-10 are to down-regulate the inflammatory response (³³). Our findings are consistent with a recent report by Aihara et al. (³⁴) who showed that incubation with NAC enhanced IL-10 secretion by human alveolar macrophages. Our findings are also consistent with data recently reported by Sang et al. (³¹) who showed that pretreatment with PBN significantly upregulated LPS-induced hepatic IL-10 mRNA in rats. However, in animal models of infection, there is evidence that elevated plasma IL-10 levels are associated with increased mortality (³⁵). The implications of elevated IL-10 levels in our study using a model of inflammation without infection are still unclear. The IL-10 levels in our study were quite variable. However, these immunomodulatory effects may be at least partly responsible for the improved survival time and they warrant further investigation.

The data presented here are insufficient to explain how ethyl pyruvate treatment modulates expression of IL-6, IL-10, and iNOS. However, it is known that ROS can activate or modulate important intracellular signaling pathways. Thus, NF-κB, a transcription factor that is important in regulating expression of many proinflammatory mediators, is activated by the release of ROS and/or alterations in the redox potential of the cell (³⁶). Activation of activator protein (AP)-1, another important transcription factor involved in regulation of the inflammatory response, is also at least partially redox dependent, possibly because of H₂O₂-mediated activation of the upstream mitogen-activated protein kinases, ERK1/2 (³⁷). In results not reported here, our laboratory has shown that incubation with ethyl pyruvate can block NF-κB activation in LPS-stimulated murine macrophages and cytokine-stimulated cultured human enterocytes (R. Yang, unpublished observations). Thus, it seems plausible that REPS modulates the inflammatory response in vivo by down-regulating the activation of NF-κB or, perhaps, other signaling pathways.

We observed that resuscitation of endotoxemic rats with REPS instead of RLS was associated with lower blood lactate concentrations. In the absence of any measures of tissue perfusion, the mechanism responsible for this finding is at best speculative. It is possible that elevated circulating lactate concentrations in the RLS group were merely a reflection of the exogenous lactate load administered to these animals. However, it is also plausible that resuscitation with REPS was associated with lower lactate levels on the basis of improved tissue perfusion. Increasingly, however, it is becoming apparent that hyperlactatemia in sepsis or endotoxemia is less a reflection of impaired oxygen delivery than it is of profound alterations in intermediary metabolism that favor a marked increase in glucose-to-lactate flux, independent of tissue oxygenation (³⁸). Therefore, it is conceivable that the anti-inflammatory effects of REPS modulated LPS-induced aerobic glycolysis in various cell types and thereby reduced circulating levels of lactate.

In summary, we have shown that delayed resuscitation with REPS significantly prolonged survival in a rodent model of shock induced by a lethal dose of LPS. Treatment with REPS modulated the inflammatory response, as evidenced by changes in the circulating levels of nitrite/nitrate, IL-6, and IL-10. Based on these data, we believe that REPS warrants further evaluation as a resuscitation fluid for the treatment of sepsis and septic shock.