Temporal changes in the involvement of pyruvate dehydrogenase complex in muscle lactate accumulation during lipopolysaccharide infusion in rats

Abstract

A characteristic manifestation of sepsis is muscle lactate accumulation. This study examined any putative (causative) association between pyruvate dehydrogenase complex (PDC) inhibition and lactate accumulation in the extensor digitorum longus (EDL) muscle of rats infused with lipopolysaccharide (LPS), and explored the involvement of increased transcription of muscle-specific pyruvate dehydrogenase kinase (PDK) isoenzymes. Conscious, male Sprague–Dawley rats were infused i.v. with saline (0.4 ml h⁻¹, control) or LPS (150 μg kg⁻¹ h⁻¹) for 2 h, 6 h or 24 h (n = 6–8). Muscle lactate concentration was elevated after 2, 6 and 24 h LPS infusion. Muscle PDC activity was the same at 2 h and 6 h, but was 65% lower after 24 h of LPS infusion (P < 0.01), when there was a 47% decrease in acetylcarnitine concentration (P < 0.05), and a 24-fold increase in PDK4 mRNA expression (P < 0.001). These changes were preceded by marked increases in tumour necrosis factor-α and interleukin-6 mRNA expression at 2 h. The findings indicate that the early (2 and 6 h) elevation in muscle lactate concentration during LPS infusion was not attributable to limited muscle oxygen availability or ATP production (evidenced by unchanged ATP and phosphocreatine (PCr) concentrations) or to PDC inhibition, whereas after 24 h, muscle lactate accumulation appears to have resulted from PDC activation status limiting pyruvate flux, most probably due to cytokine-mediated up-regulation of PDK4 transcription.

Classically, the rise in blood lactate concentration in sepsis has been attributed to an increased rate of anaerobic glycolysis (mainly in skeletal muscle) due to poor tissue perfusion and insufficient O₂ delivery and/or mitochondrial dysfunction (Mizock & Falk, 1992), and therefore, a major therapeutic aim has been to increase muscle oxygen delivery under such conditions. However, it is now apparent that muscle oxygen delivery may be adequate, as evidenced by adequate tissue blood flow and perfusion (Lang et al. 1985; Fish et al. 1986) and maintenance of a normal muscle high-energy phosphate pool (Vary et al. 1986; Hotchkiss & Karl, 1992). Thus, sepsis-induced lactate accumulation seems to result from mechanisms other than a deficit in muscle oxygen availability (for review see Gladden, 2004).

Altered regulation of skeletal muscle pyruvate dehydrogenase complex (PDC) has been implicated in the development of lactic acidosis during sepsis (Vary, 1996). In skeletal muscle, the PDC is central to the control of carbohydrate oxidation, since it catalyses the irreversible oxidative decarboxylation of pyruvate to acetyl-CoA, thereby linking glycolysis to the tricarboxylic acid (TCA) cycle. The activity of PDC is subject to regulation by a bimodal mechanism of competing pyruvate dehydrogenase kinase (PDK) and pyruvate dehydrogenase phosphatase (PDP), which together control a phosphorylation (inactivation) and dephosphorylation (activation) cycle (Wieland, 1983). At rest, increased PDK activity plays the predominant role in PDC regulation, thereby decreasing flux through PDC and limiting skeletal muscle carbohydrate oxidation (Wieland, 1983). Four, dedicated PDK isoenzymes have been identified, with different regulatory properties and tissue distributions; PDK2 and PDK4 isoforms are most prevalent in skeletal muscle (Bowker-Kinley et al. 1998).

Mitochondria isolated from septic rats have been shown to have impaired oxidation of pyruvate, but not other substrates (Mela-Riker et al. 1992), suggesting a specific defect in their ability to oxidize pyruvate, but not generalized mitochondrial dysfunction. In keeping with this, Vary et al. have reported that PDC activity was approximately 50% lower in hindlimb muscle from septic rats (for review see Vary, 1998). Furthermore, the same group reported a negative linear correlation between blood lactate concentrations and skeletal muscle PDC activity, 3 or 7 days following the induction of sepsis (Vary, 1996). In support of these observations, pharmacological activation of the PDC by administration of dichloroactetate (DCA), a non-competitive inhibitor of PDK, has been shown to normalize muscle and blood lactate concentrations during sepsis (Vary, 1998). Vary et al. also reported a 2-fold increase in hindlimb skeletal muscle PDK activity during sepsis (Vary, 1998), which they concluded could be responsible for the sepsis-induced inhibition of PDC and the lactic acidosis they observed. Several cytokines have been implicated in the pathogenesis of sepsis, most notably tumour necrosis factor-α (TNF-α), and it has been shown that sequestering TNF-α with a TNF binding protein (TNFbp) normalized muscle and blood lactate concentrations in sepsis, and prevented the inhibition of PDC and increased PDK activity (Vary et al. 1998). Collectively, these investigations clearly point to inhibition of the PDC as being responsible for the lactic acidosis in sepsis, with PDK activity being central to this inhibition. Furthermore, these studies implicate TNF-α in mediating the effects of sepsis on these regulatory enzymes.

Vary et al. utilized an intra-abdominal abscess model of sepsis, which developed over 3–7 days; they provided no information on changes occurring in the first 24 h of the septic challenge (Vary et al. 1998). We have previously assessed the haemodynamic changes produced by continuous infusion of lipopolysaccharide (LPS) in conscious rats, and have shown that in this experimental protocol, the hyperdynamic circulatory state mimics that seen in the early stages of clinical sepsis (Waller et al. 1995). Therefore, in the present study, we used the LPS infusion model to examine changes in fast-twitch muscle, which is known to be highly susceptible to sepsis (Tiao et al. 1997; Wray et al. 2003). The first aim was to establish the temporal relationship between muscle lactate accumulation and muscle PDC activity and flux, our hypothesis being that if muscle lactate accumulation was entirely due to inhibition of PDC, changes in muscle lactate concentration would be accompanied by a reduction in PDC activity, with evidence of inhibition of flux via the PDC reaction.

Inhibition of PDC in other inflammatory conditions has been shown to be due to increased PDK activity (Fuller & Randle, 1984; Feldhoff et al. 1993), with evidence for selective up-regulation of the muscle-specific PDK isoform, PDK4, at both a transcriptional and protein level (Sugden et al. 2000; Wu et al. 2000). Therefore, the second aim was to investigate the changes in PDK2 and PDK4 mRNA expression during LPS infusion, to determine whether an increase in PDK transcription could be responsible for any inhibition of PDC, and if so, whether or not it was isoform selective.

Finally, since TNF-α and interleukin-6 (IL-6) have been implicated in the pathogenesis of sepsis, and specifically in the inhibition of PDC (Vary et al. 1998), changes in the transcription of both cytokines were also evaluated to test the hypothesis that increases would occur prior to any increase in PDK transcription and inhibition of PDC activity and flux.

Methods

Ethical approval

All procedures were approved by the University of Nottingham local ethical review committee and were performed under the Home Office License authority UK (1986).

Animals

Male Sprague–Dawley rats (400–450 g) were anaesthetized (fentanyl; Janssen-Cilag, UK), and medetomidine (Pfizer, UK; 300 μg kg⁻¹ of each i.p.) and had intravenous (right jugular vein) catheters implanted. Anaesthesia was reversed with atipamezole (Antisedan; Pfizer, UK), and analgesia was provided with nalbuphine (Bristol-Myers Squibb, UK), 1 mg kg⁻¹ of each s.c.). Rats were allowed to recover overnight in their home cages and were given free access to both food and water. A counterbalanced tether system connected to a harness fitted to the rat carried the catheter, and allowed relatively unrestricted movement within the home cage. Venous catheters were connected to fluid-filled swivels (Blair et al. 1980) to allow continuous administration of saline or LPS over a 24 h period.

Prepared rats were divided into six groups: three groups were infused with sterile saline (control, 0.4 ml h⁻¹) for either 2 h (n = 8), 6 h (n = 7) or 24 h (n = 8), and the three other groups received a continuous infusion of LPS (E. coli, serotype 0127:B8; Sigma, UK; 150 μg kg⁻¹ h⁻¹ dissolved in sterile isotonic saline), again for either 2 h (n = 8), 6 h (n = 8) or 24 h (n = 6). At the specified time after onset of saline or LPS infusion, animals were terminally anaesthetized (sodium pentobarbital i.v., Sagatal; Rhône Mérieux). The extensor digitorum longus (EDL) muscle was removed from the left hind-limb and was immediately snap frozen in liquid nitrogen less than 5 s after removal. One portion for each muscle sample was subsequently freeze-dried and stored at −80°C, and the remainder was stored ‘wet’ in liquid nitrogen.

Muscle metabolite analysis

After removal of all visible blood and connective tissue, freeze-dried muscle samples were powdered and extracted with 0.5 mol l⁻¹ perchloric acid (containing 1 mmol l⁻¹ EDTA), and neutralized with 2.2 mmol l⁻¹ KHCO₃. Neutralized extracts were used for determination of ATP and PCr (as markers of muscle metabolic stress), glycogen and lactate concentrations, using modified spectrophotometric methods of Harris et al. (1974). Acetylcarnitine was measured by a radioisotope enzymatic assay as previously described (Cederblad et al. 1990), as an indicator of the magnitude of flux through the PDC reaction.

Muscle PDC activity

A small portion of frozen ‘wet’ muscle was used to determine PDC activity as previously described by Constantin-Teodosiu et al. (1991). Briefly, the activity of PDC in its dephosphorylated active form (PDCa) was assayed in a buffer containing NaF and dichloroacetate (DCA), and was expressed as a rate of acetyl-CoA formation (mmol (min⁻¹ (kg muscle wt)⁻¹) at 37°C.

Muscle mRNA analysis

Probe and primer sets were purchased for PDK2, PDK4, TNF-α, IL-6 and hydroxymethylbilane synthase (HMBS) (Applied Biosystems, USA), and the detection of mRNA was performed using a ABI-PRISM 7700 Sequence Detector (Applied Biosystems). HMBS was selected as an endogenous control to correct for potential variations in RNA loading and efficiency of reverse transcription. The expression of HMBS was previously found to be unaffected by LPS infusion (data not shown). Samples were run in triplicate, and were interpreted using the 2^−ΔΔCt method for relative quantification of gene expression. The threshold cycle (Ct) for HMBS was subtracted from the Ct of the target gene to adjust for variations in mRNA/cDNA generation efficiency; this was carried out for all samples. Treatment groups were normalized to the average of the control group for each time point.

Statistics

All data are reported as mean ± s.e.m. Comparisons of time-points within each treatment and between treatments were performed using one-way ANOVA, with the LSD post hoc test being used to locate any significant differences. Statistical significance was declared at P < 0.05.

Results

Muscle metabolites

Skeletal muscle metabolite data are presented in Table 1. Muscle ATP, PCr and glycogen concentrations did not change between treatment groups or within each treatment group. Similarly, muscle acetylcarnitine concentration did not change over time within the control group. Muscle acetylcarnitine concentration was 47% lower after 24 h of LPS infusion compared to the corresponding control value (P < 0.05; Table 1), and was lower at 24 h compared to 2 h (P < 0.05) in the LPS-treated group. Muscle lactate concentration did not change over time in the control group (Fig. 1). Muscle lactate concentration was significantly greater in the LPS-treated group compared to control at 2 h (12.1 ± 1.4 versus 6.6 ± 1.7 mmol (kg dry muscle (dm))⁻¹, P < 0.01), and this difference persisted at 6 h (9.3 ± 1.4 versus 5.0 ± 1.1 mmol (kg dm)⁻¹, P < 0.05) and 24 h (7.1 ± 1.2 versus 3.7 ± 0.7 mmol (kg dm)⁻¹, P < 0.05; Fig. 1) of infusion. Furthermore, in the LPS-treated group, muscle lactate concentration was lower at 24 h compared to 2 h (7.1 ± 1.2 versus 12.1 ± 1.4 mmol (kg dm)⁻¹, respectively; P < 0.05).

Table 1.

Extensor digitorum longus muscle ATP, phosphocreatine (PCr), glycogen and acetylcarnitine concentrations during saline (control) and LPS infusion

	Time (h)
	2		6		24
	Control	LPS	Control	LPS	Control	LPS
ATP	34.3 ± 1.4	33.4 ± 2.6	33.3 ± 1.7	31.7 ± 2.3	35.6 ± 2.4	36.0 ± 2.1
PCr	78.7 ± 4.8	68.9 ± 7.4	71.4 ± 7.5	77.3 ± 5.7	80.1 ± 4.6	68.8 ± 14.1
Glycogen	136 ± 6	121 ± 11	129 ± 9	134 ± 12	151 ± 12	116 ± 18
Acetylcarnitine	0.38 ± 0.05	0.33 ± 0.66	0.29 ± 0.04	0.37 ± 0.04	0.36 ± 0.05	0.19 ± 0.03‡*

All concentrations are expressed as mmol (kg dry muscle)⁻¹. Values represent mean ± s.e.m.

^‡P < 0.05, significantly different from time-matched control.

^*P < 0.05, significantly different from 2 h LPS-treated muscle. n = 6–8 in each group.

Figure 1. Extensor digitorum longus muscle lactate concentrations during saline (control) and LPS infusion.

Values represent mean ± s.e.m.‡P < 0.05, significantly different from time-matched control. *P < 0.05, significantly different from 2 h LPS-treated muscle. n = 6–8 in each group.

Muscle PDC activity

Muscle PDC activity did not change over time within the control group. Muscle PDC activity was significantly lower after 24 h of LPS infusion compared to control (0.32 ± 0.05 versus 0.80 ± 0.08 mmol min⁻¹ (kg wet muscle (wm))⁻¹, respectively; P < 0.01; Fig. 2). Furthermore, in the LPS-treated group, muscle PDC activity was lower at 24 h (0.32 ± 0.05 mmol min⁻¹ (kg wm)⁻¹) compared to 2 h (0.85 ± 0.17 mmol min⁻¹ (kg wm)⁻¹, P < 0.05) and 6 h (0.91 ± 0.32 mmol min⁻¹ (kg wm)⁻¹, P < 0.05; Fig. 2).

Figure 2. Extensor digitorum longus muscle PDC activity during saline and LPS infusion.

Values represent mean ± s.e.m.‡‡‡P < 0.001, significantly different from time-matched control. *P < 0.05, significantly different from 2 h LPS-treated muscle. †P < 0.05, significantly different from 6 h LPS treated muscle. n = 6–8 in each group.

Muscle PDK2 and PDK4 mRNA expression

Figure 3 shows fold change in PDK2 and PDK4 mRNA expression from corresponding control values after LPS infusion. Muscle PDK2 mRNA expression did not change significantly from the expression levels recorded in the control group, although there was a small decrease in PDK2 expression at 24 h compared to 6 h (Fig. 3A). Muscle PDK4 mRNA expression was no different from control at 2 h, but was 9.6-fold greater after 6 h (P < 0.001) and 23.5-fold greater (P < 0.001) after 24 h of LPS infusion (Fig. 3B).

Figure 3. Fold changes in PDK2 (A) and PDK4 (B) mRNA expression from corresponding control value after LPS infusion within extensor digitorum longus.

A value > 1 indicates greater than control mRNA expression and < 1 is lower than control mRNA expression. Values represent mean ± s.e.m.‡‡‡P < 0.001, significantly different from time-matched control. †P < 0.05, significantly different from 6 h LPS-treated muscle. n = 6–8 in each group.

Muscle TNF-α and IL-6 mRNA expression

Figure 4 shows fold change in TNF-α and IL-6 mRNA expression from corresponding control values during LPS infusion. TNF-α mRNA expression was 11-fold greater than the corresponding control value after 2 h of LPS infusion (P < 0.01; Fig. 4A). Thereafter, TNF-α mRNA expression declined toward control values, but was still 4.4-fold greater than control at 24 h (P < 0.05). IL-6 mRNA expression was elevated 176-fold above control after 2 h of LPS infusion (P < 0.01), declining thereafter, but remaining significantly elevated above corresponding control values after 6 h (P < 0.01) and 24 h of LPS infusion (P = 0.07).

Figure 4. Fold changes in TNF-α (A) and IL-6 (B) mRNA expression from corresponding control value after LPS infusion within extensor digitorum longus.

A value > 1 indicates greater than control mRNA expression and < 1 is lower than control mRNA expression. Values represent mean ± s.e.m.‡P < 0.05, ‡‡P < 0.01, significantly different from time-matched control. n = 6–8 in each group.

Discussion

To our knowledge, this is the first study to establish the temporal relationship, if any, between muscle lactate accumulation and muscle PDC activity and flux in rat EDL muscle during in vivo LPS-induced sepsis. The results demonstrate that the initial (at 2 and 6 h) increase in muscle lactate concentration during LPS infusion preceded any measurable inhibition of PDC, whereas after 24 h of LPS infusion, the muscle lactate accumulation was accompanied by inhibition of PDC activity and flux. Since there were marked early changes in muscle cytokine mRNA levels (at 2 h), which were followed by equally marked increases in PDK4 mRNA expression (from 6 h onwards), a logical interpretation of the results is that the increase in muscle lactate concentration in EDL after 24 h of LPS infusion was due to inhibition of PDC activity attributable to a cytokine-mediated increase in PDK4 transcription.

The first notable finding of the present study was the elevation of skeletal muscle lactate concentration (Fig. 1) in the absence of a change in muscle PDC activity (Fig. 2) during the early stages of LPS infusion (2 and 6 h). The mechanism underlying this early increase in skeletal muscle lactate concentration cannot be determined from the present study, but it is clear that it was not due to inhibition of PDC activity, or a decline in muscle PDC flux, since muscle acetylcarnitine concentration was unchanged (Table 1). Perhaps the most likely explanation for this early rise in muscle lactate concentration is an adrenaline-mediated activation of muscle glycolysis, and thereby lactate production. Sepsis is known to markedly increase the adrenaline concentration of blood, in a time- and dose-dependent manner (Jones & Romano, 1989). Furthermore, although not measured, it is likely that circulating adrenaline increased in the present study, as LPS-mediated hindquarter vasodilatation is blocked in the same model by non-selective and β2 selective adrenoceptor blockade (Gardiner et al. 2005). Adrenaline binds to 2-adrenoreceptors triggering a classical cascade of events that results in cAMP formation, which in turn catalyses the transformation of glycogen phosphorylase b to its active a form, through the cAMP-dependent protein kinase (Cohen, 1981). The extent of glycogen phosphorylase b to a conversion is believed to be the rate-limiting step of muscle glycogenolysis during muscle contraction. In sepsis, a drive for non-aerobic ATP formation, and therefore lactate formation, seems to be linked to an adrenaline-mediated stimulation of the sarcolemmal Na⁺–K⁺-ATPase pumps (James et al. 1999). Irrespective of the mechanism, it seems unlikely that the LPS-mediated early rise in muscle lactate concentration in the present setting can be attributed to a major increase in glycolysis, since muscle glycogen concentration did not change significantly from control (Table 1), and the difference in muscle lactate concentration between control and LPS-treated groups at 2 and 6 h equates to only about 2–3 mmol (kg dry muscle)⁻¹ of glycogen, which is a value within the acceptable range of the coefficient of variation for glycogen determination (2–3%; Harris et al. 1974) and therefore difficult to detect. Furthermore, it seems clear that the early increase in EDL lactate concentration is unlikely to have been a result of muscle hypoperfusion, since we have previously shown increased hindquarter blood flow in this model during the early stages of LPS infusion (Waller et al. 1995). Finally, one possibility is that the early onset of muscle lactate accumulation may be due to a specific defect in sarcolemmal lactate transport. For example, studies have demonstrated a decrease in skeletal muscle membrane potential and increased intracellular Na⁺ concentrations in a variety of models of sepsis and septic shock (Gibson et al. 1977; Trunkey et al. 1979; Illner & Shires, 1981; Shiono et al. 1989; Hannon & Boston, 1990), and the fall in muscle membrane potential during shock has been shown to correlate with the rise in muscle lactate concentration in dogs (Jennische et al. 1978), primates (Peitzman et al. 1985) and humans (Liaw et al. 1980). However, in pilot experiments, we have measured markedly elevated arterial blood lactate concentration at 2 and 6 h of LPS infusion (N. Alamdari, A. J. Murton, D. Constantin-Teodosiu, S. M. Gardiner, T. Bennett, R. Layfield and P. L. Greenhaff, unpublished observations). Therefore, it is possible that in this instance the early increase in muscle lactate concentration may have at least partly been due to a reduced lactate concentration gradient across the sarcolemma (resulting in lower muscle lactate efflux) because of a sepsis-mediated impairment of blood lactate clearance.

A second major finding of the present study was that, in contrast to the observations at 2 and 6 h, the elevation in skeletal muscle lactate concentration after 24 h of LPS infusion (Fig. 1), was accompanied by decreases in muscle PDCa activity (Fig. 2) and acetylcarnitine concentration (Table 1), and a marked elevation in PDK4 mRNA expression (Fig. 3B), clearly demonstrating that PDC activity and flux were simultaneously inhibited at this later stage. Hence, inhibition of muscle PDC could have been responsible for muscle lactate accumulation at this time-point, possibly mediated by a marked increase in muscle PDK4 transcription. Vary et al. demonstrated that sepsis (induced by an intra-abdominal abscess), allowed to develop over 5 days, increased PDK activity in skeletal muscle although the specific PDK isoform was not identified (Vary, 1998). Increases in muscle PDK activity with starvation and experimentally induced diabetes have also been shown to be associated with a selective up-regulation of PDK4 mRNA expression (Sugden et al. 2000; Wu et al. 2000), but whether sepsis induces similar shifts in the abundance of the PDK isoenzymes was hitherto unknown. We demonstrate here, for the first time, a marked elevation of PDK4, but not PDK2 mRNA after 6 h of LPS infusion, which increased further at 24 h to 24-fold above control. Clearly, therefore, LPS infusion can rapidly, and selectively, up-regulate PDK4 transcription, and this is the most likely reason for the inhibition of PDC we observed after infusion of LPS for 24 h, and for the increase in total PDK activity reported by others in sepsis (Vary, 1998).

The time-course of change in cytokine mRNA expression in the present study is consistent with the observations of Waller et al. (1995), who, in the same LPS-infusion model, saw early (1–2 h) elevations in plasma TNF-α levels, with a return to basal thereafter. Since there is evidence that TNF-α may be responsible for the inhibition of PDC activity, our observations suggest that the increased cytokine levels seen at 2 h (Fig. 4) may have led to the increased transcription of PDK4 seen from 6 h (Fig. 3B), which resulted in the inhibition of PDC seen at 24 h (Fig. 2). It is plausible that the mechanism of PDK4-induced inhibition of PDC at 24 h of LPS infusion was attributable to the direct binding of fork-head box O transcription factors (FOXO) to the promoter region of the PDK4 gene (Furuyama et al. 2003) Under normal circumstances FOXO transcription factors are phosphorylated by the serine-threonine kinase Akt, and as a consequence leave the nucleus rendering them incapable of activating transcription, including PDK4 transcription (Rena et al. 1999; Birkenkamp & Coffer, 2003). One important function of Akt therefore is the regulation of gene transcription through inactivation (phosphorylation) of FOXOs. In the context of cytokines, it has recently been demonstrated that TNF-α decreases Akt protein levels and/or signalling in apidocytes (Medina et al. 2005) and human skeletal muscle (Plomgaard et al. 2005) thereby presumably activating (dephosphorylating) the family of FOXO transcription factors. It is plausible, therefore, that cytokine-mediated impairment of Akt function in the present experimental setting increased FOXO-mediated transcriptional activity and resulted in the marked up-regulation of PDK4 transcription observed after 24 h of LPS infusion (Fig. 5). In support of this conclusion, the muscle-specific atrogins, atrogin-1 (MAFbx) and MURF-1, that are known to be transcriptionally regulated by the Akt–FOXO signalling axis, are also dramatically up-regulated in sepsis (Wray et al. 2003) and other inflammatory conditions (Lecker et al. 2004). Multiple previous studies support the existence of a regulatory role for these atrogins in proteosome-mediated muscle atrophy in a variety of catabolic states (Gomes et al. 2001; Bodine et al. 2001), including sepsis (Glass, 2003), and the mRNA levels for atrogin-1 and MURF-1 are frequently used as molecular markers of muscle wasting (Lecker et al. 2004).

Figure 5. Mechanism of TNF and IL-6 inhibition of PDC activity and lactate accumulation.

LDH, lactate dehydrogenase; PDC, pyruvate dehydrogenase complex; PDK4, pyruvate dehydrogenase kinase isoform 4; CAT, carnitine acetyltransferase; TCA, tricarboxylic acid cycle; FOXO, forkhead transcription factor; MURF-1, muscle ring finger 1; MAFbx, muscle atrophy F-box/Atrogin-1. ↑ denotes up-regulation ↓ denotes down-regulation.

It is well established that PCr is a sensitive marker of muscle oxygen availability and mitochondrial function (for review see Greenhaff & Timmons, 1998). However, there was clearly no change in the concentration of muscle ATP and PCr at any time-point during LPS infusion in the present experiment (Table 1), which, as outlined earlier, is in accordance with reports of adequate muscle blood flow and perfusion (Lang et al. 1985; Fish et al. 1986) and the maintenance of a normal muscle energy charge (Vary et al. 1986; Hotchkiss & Karl, 1992) in sepsis. Thus, in contrast to several animal models of severe sepsis and septic shock (Schumer et al. 1971; Crouser et al. 2002; Brealey et al. 2004), our results indicate no decrease in muscle oxygen availability or impairment of mitochondrial ATP production in an animal model, which shows haemodynamic changes representative of the early stages of clinical sepsis (Waller et al. 1995), probably because the challenge is less severe and does not induce circulatory shock. Indeed, the haemodynamic profile of the animal model utilized here has been characterized over the same time-course, and does not constitute a model of endotoxaemic shock, but represents the hyperdynamic circulatory status seen in early clinical sepsis (Parrillo, 1993). It has been shown that after 2 h of LPS infusion (Waller et al. 1995), there was an apparent modest hypotension and vasodilatation in the internal carotid, renal and hindquarters vascular beds. After 6 h, although arterial blood pressure had returned to baseline, there was still vasodilatation in the renal vascular bed, but vasoconstriction in the internal and common carotids and the hindquarters. After 24 h, there was apparent hypotension, tachycardia and generalized vasodilatation.

In conclusion, the present study demonstrates that the early increase in muscle lactate concentration during LPS infusion was not accompanied by a reduction of muscle PDC activity or flux, and might have been due to an adrenaline-mediated increase in muscle glycolysis acting in conjunction with decreased muscle lactate efflux. In contrast, after 24 h of LPS infusion, muscle lactate accumulation was accompanied by inhibition of muscle PDC activity and flux, which was probably due to a rapid and dramatic increase in PDK4 transcription, probably mediated by skeletal muscle cytokines. Finally, we could find no evidence for an impairment of muscle oxygen availability or mitochondrial ATP production in an animal model which represents the hyperdynamic phase of early clinical sepsis.