INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: glycolysis; lactate; mitochondrial function; sepsis; skeletal muscle

Sepsis remains a major cause of death in intensive care patients, and considerable effort has been made in recent years to clarify the hemodynamic changes that occur, and which are presumed to induce tissue hypoxia and/or energy production failure. But even though a set of different pathophysiological changes such as hyperlactatemia and increased muscle glucose uptake are generally observed during sepsis (1), energy metabolism disturbances are still not clearly understood. In fact, the two energetic pathways (glycolysis and mitochondrial respiration) have never been studied concomitantly, which sometimes leads to a misinterpretation of individual changes in substrates, intermediates, or products, within a sole metabolic pathway. For example, it is assumed that lactic acidosis is merely due to tissue hypoxia (4). In contrast, a recent study has shown that lactate turnover may increase in septic patients, under normoxic conditions (i.e., hemodynamically stable patients) (5). In fact, hyperlactatemia can be induced by several but not exclusive factors, the main two being an activation of the glycolytic pathway and/or an alteration in the mitochondrial structure and function (6). As skeletal muscle represents about 40% of the body mass, and provides between 50 and 100% of the substrates for oxidative phosphorylation during stress in humans, it is interesting to study this tissue in regard to the metabolic consequences of sepsis.

The aim of the present study was to globally assess the energy drive in septic rat myocytes, studying both glycolysis and mitochondrial maximal activities together.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experimental Animals

Ten male Sprague-Dawley rats, weighing 400-500 g, were assigned to a sham-control or septic group. After intraperitoneal anesthesia using ketamine (11 mg/100 g) and xylazine (0.5 mg/100 g), the animals were laid out on a thermo-regulated mattress to keep core temperature constant. Core temperature was controlled using a tele-thermometer connected to a rectal probe. Sepsis was induced using the cecal ligation-perforation (CLP) model with a 0.5-cm incision (9). The animals were resuscitated with 10 ml/100 g saline solution administered subcutaneously. Sham-operated control animals underwent the same procedures, but the cecum was neither ligated nor incised. Recovery was performed while breathing room air. Sepsis was defined by temperature variation (<36° C or >38° C) and/or decreased sensitivity to touch, diarrhea, eye hemorrhage, and piloerection. Tissue and blood sampling were performed at baseline (H0) and 4 h after (H4).

Blood and Muscle Sampling

A venous blood sample was taken from the internal jugular vein at H0 and H4, and centrifuged. The lateral gastrocnemius (a mixed fiber I and II muscle) was removed from the left leg at H0 and from the right leg at H4. One part was stored at 80° C for kinetic measurements, and the other was used for mitochondrial function assessment.

Glucose and Lactate Measurements

These were performed using a specific electrode (YSI 2300 Stat plus, Yellow Spring Instruments, Ohio). The results are expressed as mmol/L in plasma and µmol/g in muscle.

Glycolysis Assessment

The technique has previously been described in detail (12). The tissues were homogenized in a HEPES buffer and centrifuged, and the supernatants were used for experiments. The functioning of the system was assayed in a spectrophotometer by a continuous recording of the NADH decay. The system was fed successively with glucose and glucose 6-phosphate (G6P), and their conversion into glycerol-phosphate was shortened by means of the enzymes of the hexose part of glycolysis and added auxiliary ones. After glucose addition, the system reached a steady state, which corresponds to the aerobic flux (J_A). Subsequently, G6P was added to the system, which reached another steady state, which represents the anaerobic flux (J_B). The metabolic response time (t₉₉) represents the transition between both steady states. The experiments were performed in 37° C-thermostated cuvettes. The results are expressed as nmol or µmol glucose/min/g tissue for fluxes and seconds for t₉₉.

Mitochondrial Function Assessment

Respiratory parameters were determined using a previously established method (13), providing permeabilized muscle fibers with functionally intact mitochondria. The muscle fibers were isolated and incubated in a relaxing solution A (K⁺MES/high-energy phosphate buffer), with the addition of saponin. After rinsing, oxygen consumption rates were measured at 37° C using Clark-type electrodes (1302 model, Stratkelvin, UK) in closed-thermostated oxygraph cuvettes filled with solution B (solution A/bovine serum albumin instead of high- energy phosphates). Fibers, substrates, and inhibitors were added as follows: pyruvate-malate, fibers, ADP, rotenone, succinate, myxothiazol + antimycin, and ascorbate + tetramethyl-p-phenylenediamine. Respiratory activities are expressed as µmol O₂/h/mg of muscle.

Data Analysis

All measurements were performed in triplicate. The results are expressed as mean ± SE and compared using Student's paired and unpaired t test. A p < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Typical absorbance time graphs for sham and septic animals are presented in Figure 1.

View larger version (14K): [in this window] [in a new window]   Figure 1.   Typical composite absorbance time graphs for both sham-control and septic groups, at H0 and H4. This figure was obtained by an ASCII transformation of the spectrophotometer signal. An increased anaerobic glycolytic flux (JB) was observed at H4 in all septic animals, compared with sham-control animals. The arrow indicates the G6P injection; before arrow: aerobic glycolytic flux (JA); after arrow: anaerobic glycolytic flux (JB). Half tone line: absorbance time graph at H0; black line: absorbance time graph at H4.

Effects of Surgical Procedures (Except CLP) and Anesthesia

These effects are estimated comparing sham-control rats at H0 and H4. No clinical signs of sepsis were observed. No significant changes in either lactate or glucose plasma and muscle content, or mitochondrial function disturbances, were induced by surgical procedures and anesthesia (see Tables 1 and 3).

Metabolic Disturbances Induced by Sepsis

At H4, all animals in the septic group presented clinical signs of sepsis and hypothermia (temperature = 36 ± 0.5° C). No plasma glucose content variation was observed, whereas muscle glucose content (see Table 1) did significantly decrease at H4 ([G_m] = 2.8 ± 0.1 versus 1.6 ± 0.1 µmol/g; p < 0.001). A very significant increase in plasma lactate content (see also Table 1) was observed in septic rats ([lactate] = 1.2 ± 0.3 versus 3.3 ± 0.6 mmol/L; p < 0.001).

We observed a significant increase in both aerobic (J_A = 7.2 ± 0.9 versus 18.2 ± 4.1 nmol glucose/min/g; p < 0.05) and anaerobic (J_B = 7.5 ± 1.2 versus 15.4 ± 3.4 µmol glucose/min/g; p < 0.05) glycolytic fluxes in septic rats, associated with a metabolic response time decrease (t₉₉ = 0.13 ± 0.02 versus 0.19 ± 0.01 s; p < 0.05) (see Table 2). Mitochondrial function assessment showed a significant pyruvate-malate-dependent oxygen consumption decrease (O₂-PM = 0.113 ± 0.007 versus 0.144 ± 0.008 µmol O₂/h/mg; p < 0.05) in septic animals (see Table 3). No other mitochondrial function disturbances were observed, either for succinate or ascorbate-TMPD-dependent oxygen consumption.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Sepsis is characterized by a number of metabolic abnormalities, including increased glucose uptake and hyperlactatemia (16). If the search for the causes of metabolic abnormalities during sepsis have so far been largely unsuccessful, two main hypotheses have been put forward: (1) cellular hypoxia, resulting from abnormal microcirculatory flow, or (2) defect(s) in energy-producing aerobic metabolic pathways. Although other tissues such as heart, liver, and leukocytes may exhibit similar metabolic abnormalities, skeletal muscle, which forms a large fraction of the body cell mass (40%), is a major contributor to the overall lactate metabolism. Until now, traditional techniques for assessing metabolism have included measurements of enzyme activities, concentration of substrates, intermediates, and products, with a view to gathering evidence of the metabolic state of the tissue. Although those measurements may provide insights into the actions of individual components of the glycolytic pathways, the snapshot provided by the data may not tell the whole story and tends to provide little information about the metabolic flow through the cascade of enzymatic reactions (18). Moreover, most studies investigated a specific metabolic pathway (i.e., aerobic or anaerobic glycolysis), but never assessed the global functioning of the energetic system (glycolysis, both aerobic and anaerobic and mitochondrial function). To our knowledge, this is the first study that has measured both aerobic and anaerobic metabolism concomitantly, using recent but validated in vitro techniques (12, 18).

In our study, hyperlactatemia was observed in all animals in the septic group (3-fold), whereas no changes were observed in the sham-control group. Apart from hyperlactatemia and decreased muscle glucose content, as already observed by numerous authors (2, 3), the clearest and most striking effects of sepsis were an enhanced glycolytic activity, both aerobic and anaerobic (150 and 140% increase, respectively). The trend toward decreased muscle lactate content in septic rats is clearly dependent on sample-size power, and a more pertinent muscle lactate content measurement might need tissue space and water content assessments. Hyperlactatemia is thought to be a reliable predictor of outcome in critically ill patients (1, 19), and its variation over time has been highlighted as a useful tool to assess patient response to treatment (20, 21). It has already been taken for granted that cells derive no metabolic benefits by using the glycolytic pathway when oxidative metabolism is possible, and thus as a corollary, that rapid glycolysis by cells that have oxidative metabolism capabilities provides evidence of some cellular defect. Hyperlactatemia during sepsis was therefore considered to represent inadequate tissue perfusion (19, 22). This reductionist approach to whole-body metabolism contrasts starkly with the complex kinetics of lactate production and use by the tissues (2). It now appears unlikely that cellular hypoxia is the main abnormality during sepsis (4, 23, 24), and that it is responsible for the enhanced glycolysis (25). It has been shown that glycolysis is independent of the oxygen level, that many tissues seem to generate lactate under aerobic conditions (so-called aerobic glycolysis) (5), and that high rates of oxidative phosphorylation and glycolysis are not mutually exclusive (25). In fact, cellular glucose uptake and glycolytic flux may well be stimulated by mediators produced during the systemic inflammatory response.

What role does this apparently "excessive" glycolytic flux play in cellular energetics? According to our results, increased plasma lactate levels during sepsis may indeed represent a "normal" adaptive response, that is, increased glycolysis, to balance an eventual aerobic pathway dysfunction. In fact, we observed a significant decrease in pyruvate-malate- (a complex I electron donor) dependent oxygen consumption, whereas in contrast, oxygen consumption by permeabilized myocytes supplemented with succinate (a complex II electron donor) and ascorbate-TMPD (a complex IV electron donor) did not decrease in the septic compared with the sham-control group. This mitochondrial function disturbance can be interpreted either as a selective complex I (which catalyzes the first irreversible reaction in the mitochondrial oxidation of glucose and determines whether pyruvate is oxidized via the Krebs cycle or is converted to lactate via lactate dehydrogenase or to alanine transaminase) and/or a complex I-ubiquinone relation alteration. It has been suggested that sepsis could promote the conversion of pyruvate dehydrogenase to the inactive isozyme, the consequence of which is the accumulation of pyruvate and a subsequent formation of lactate. The use of dichloroacetate (DCA) has thus been proposed as a therapy for lactic acidosis during sepsis. However, the results are controversial (2) and show that DCA does not seem to affect oxygen consumption (26). These observations together with our results suggest that the accumulation of pyruvate and increased lactate level are undoubtedly related to a defect in the first part of the respiratory chain (complex I-ubiquinone) rather than to a decrease in pyruvate dehydrogenase activity.

These respiration measurements are in agreement with previous clinical and experimental data, suggesting sepsis- induced mitochondrial structure and function disturbances. In an experimental model, Welty-Wolf and coworkers observed early structural damage to mitochondrial membranes and matrix during sepsis (7). In hemodynamically unstable septic patients, Levy and coworkers observed lactic acidosis associated with elevated lactate to pyruvate and arterial ketone ratios, suggesting both a decrease in the cytoplasmic and mitochondrial redox state (27). Changes in the redox state of cytochrome- c-oxidase were detected in septic baboons (28), whereas a link between mitochondrial enzyme changes and the severity of sepsis was established by other investigators (6). This complex I decreased activity has also been observed in a previous experimental study (29). The data are thus conflicting, showing that mitochondrial function might not be affected, or even increased, during sepsis (30). However, Singer and Brealey (8) have recently published some results obtained from human or nonhuman cell preparations, and there seems to be a depression in mitochondrial function whereby the production of free radicals and nitric oxide could have considerable importance.

Our results clearly show a mitochondrial respiration dysfunction, and thus a decreased oxidative phosphorylation. This may have two main consequences: first, as glycolysis is a more rapid energetic pathway, pyruvate cannot be totally integrated in the Krebs cycle and its overproduction contributes to hyperlactatemia. However, the fact that the respiratory chain continues to function, although at a lower level, means that oxygen is sufficient but misused (mitochondrial P₅₀ is thought to be  0.05 mm Hg (34). Second, the decreased oxidative phosphorylation thus induces a decreased adenosine triphosphate (ATP) production, which can be estimated to be about 100 µmol ATP/h/g muscle (see Table 3 for calculation). The decrease in oxidative phosphorylation may then explain a glycolysis activation, the results of which can be estimated to be about 400 µmol ATP/h/g muscle production (see Table 2, and considering the variations observed in the sham-control animals). It thus appears that glycolysis not only compensates for the aerobic energy production failure, but also produces excess energy, concomitantly with lactate overproduction. Is this due to an inefficient glycolysis control, and/or to a particularly high energy requirement during sepsis (in order to maintain homeothermy for example)? These observations should induce further investigation.

In conclusion, mitochondrial function assessment on permeabilized muscle fibers, associated with glycolytic flux measurement, provides a complete evaluation of the energy metabolism, as well as an insight to the influence of different situations. Recent findings support the view that mitochondrial dysfunction and ultrastructural modifications are of importance. Our results are consistent with the hypothesis that an altered mitochondrial function is responsible, at least in part, for the hyperlactatemia characteristic of septic hypermetabolism and that hyperlactatemia is thus mainly a consequence of an increased glycolysis rate.