Introduction

In the liver, IL6, IL1β and TNFα mediate the acute phase reaction, which is characterised by the increase in positive acute phase proteins on the expense of the expression of negative acute phase proteins (Ramadori and Christ 1999; Papanicolaou et al. 1998; Heinrich et al. 1990). Yet, due to the wide systemic cytokine actions the prolonged acute phase reaction may result in the septic state, which even today carries a high mortality risk for patients in intensive care. One major complication in the pathogenesis of the septic shock is the impairment of body glucose homeostasis comprising increased glucose uptake and utilization at the site of tissue disturbance, inhibition of hepatic glycogen synthesis, glucose intolerance and insulin resistance. During initial sepsis, blood glucose concentrations are elevated as a consequence of increased hepatic glucose output from glycogen, which during prolonged sepsis results in the depletion of glycogen stores, because glycogen breakdown exceeds glycogen storage (Mizock 1995). This may not be compensated by the increase in hepatic glucose formation via gluconeogenesis, because the expression of the key control enzyme of gluconeogenesis, the cytosolic form of phosphoenolpyruvate carboxykinase (PCK1), is impaired by proinflammatory cytokines (Memon et al. 1994; Christ et al. 1994; Christ and Nath 1996).

So far, these changes in the hepatic carbohydrate metabolism have not been shown on the tissue level. The present study provides evidence that the levels of α2M mRNA are elevated by IL6 in the periportal zone of the liver lobule on the expense of PCK1 mRNA expression. Glycogen depletion is enhanced by IL6 while PCK1 enzyme activity is not affected during the time period investigated. Thus, the attenuation of PCK1 gene expression by IL6 may provide nucleotides for the vast synthesis of positive acute phase proteins. During the initial acute phase reaction, the hepatic glucose providing capacity may be accomplished by the depletion of hepatic glycogen pools.

Materials and methods

Animals and chemicals

Male Wistar rats (200–250 g) (Winkelmann, Borchen, Germany) were kept under a 12 h night/12 h day rhythm (light from 7.00 a.m. to 7.00 p.m.) with free access to water and food (“Sniff”, Spezialitäten GmbH, Soest, Germany). Animal housing and experimental procedures were according to the German legislation on animal protection. Chemicals were of the highest purity grade available and purchased from commercial sources. Recombinant human (rh) IL6 was from Strathmann Biotec (Hamburg, Germany).

Experimental setting

At 6.00 a.m., rhIL6 (in 0.1% rat serum albumin in 0.9% NaCl) was injected intraperitoneally at a concentration of 0.02 μg/kg body weight. Untreated control animals were kept separately from IL6 injected animals. At the times indicated, animals were anaesthesised by the intraperitoneal injection of pentobarbital (60 mg/kg body weight), and livers were removed after perfusion with PBS (phosphate-buffered saline: 140 mM NaCl, 2.7 mM KCl, 9 mM Na2HPO4, 1.5 mM KH2PO4, pH 7.4). Organs were cut in pieces of 500 mg each, frozen in −30°C isopentane and stored at −70°C for further processing for ISH (in situ hybridisation), isolation of total RNA, and determination of enzyme activities and glycogen content. Time course experiments revealed that PCK1 mRNA levels reached a maximum at noon and glycogen levels were lowest at 2 p.m. during the normal feeding rhythm. α2M mRNA was maximally elevated at 2 p.m. after IL6 injection (not shown). Therefore, these points in time were chosen to analyse livers for mRNA levels and glycogen content, respectively.

In situ hybridization (ISH)

Hybridization probes were generated by in vitro transcription using labelled digoxigenin-UTP according to the manufacturer’s protocol (Roche, Mannheim, Germany). Antisense transcripts to detect PCK1 mRNA were generated from the HindIII-linearized plasmid pBS-PCK by using the T3 RNA polymerase promoter flanking the 1.2-kb PCK1 cDNA fragment in the Bluescript (pBS) vector (Bartels et al. 1989). To synthesise antisense transcripts for the detection of α2M (α2-macroglobulin) mRNA, a PCR product was generated from the plasmid pBR-α2M (pBR322 vector containing a 0.6 kb cDNA fragment of rat α2M; a gift of Dr. Elke Roeb, Giessen, Germany) by the use of a primer pair carrying the T7 RNA polymerase promoter sequence at the 5′-end of the reverse primer. The resulting PCR product was then used in combination with T7 RNA polymerase in the in vitro RNA synthesis reaction as described above. To improve hybridization efficacy, 1,200 and 600 nucleotides long PCK1 and α2M transcripts were shortened to 200 nucleotides in length by alkaline hydrolysis for 38 and 30 min, respectively, according to standard procedures (Ausubel et al. 1993). Specific hybridisation of the probes to PCK1 and α2M mRNA was checked by Northern blotting using 15 μg of total RNA isolated from whole rat livers.

PCK1 and α2M sense transcripts were generated by in vitro RNA synthesis using T7 RNA polymerase in combination with pBS-PCK and a α2M cDNA PCR product, which was gained by the use of a primer pair carrying the T7 RNA polymerase promoter sequence at the 5′-end of the forward primer in combination with T7 RNA polymerase, respectively. These transcripts did not hybridise to PCK1 and α2M mRNA as confirmed by Northern blot analysis and were used as control probes in the ISH reactions.

Frozen liver sections of 10 μm thickness were mounted onto poly-l-lysine-coated slides, air-dried for 1 h, fixed in 4% formaldehyde in PBS for 10 min, rinsed with 3×-concentrated PBS for 5 min and finally washed 2 times in 1×-concentrated PBS. Sections were treated for 10 min with 0.3% Triton X-100 to solubilise membranes, then washed 3 times in PBS for 5 min each and further incubated for 15 min at 37°C with proteinase K (5 μg/ml). Proteinase K was inactivated by washing for 5 min in 0.1 M glycine. Sections were again fixed in 4% paraformaldehyde and then washed once with 3×-concentrated PBS for 5 min followed by 2 washes in 1×-concentrated PBS for 5 min each. Subsequently, they were incubated for 10 min in 0.25% acetic anhydride in 0.1 M triethanolamine buffer (pH 8.0) and again washed 3 times in PBS for 5 min each (Eilers et al. 1995). To minimise unspecific binding of hybridisation probes, sections were pre-incubated in 100 μl for 1 h at room temperature in pre-hybridisation solution (1 ml is made of 500 μl deionised formamide, 200 μl of 50% dextrane sulfate, 20 μl of 50×-concentrated Denhardt´s solution (1% bovine serum albumin, 1% polyvinylpyrrolidone, 1% Ficoll in water), 50 μl of salmon sperm DNA (10 mg/ml), 200 μl of 20×-concentrated SSC (3 M NaCl, 0.3 M sodium citrate in water, pH 7.0) and 1 μl of 20% SDS. Then, the pre-hybridisation solution was changed against 100 μl of hybridisation solution, which contained either the digoxygenin-labeled PCK1 or α2M hybridisation probe in pre-hybridisation solution at 25 ng/100 μl. Slides were covered with parafilm and incubated in a humid chamber at 42°C overnight. After the hybridisation, sections were washed 3 times with 4×-concentrated SSC for 10 min each at 50°C followed by the incubation with 100 μl of RNase A (20 μg of RNase A in 1 ml of 10 mM Tris, 1 mM EDTA, 0.5 M NaCl, pH 8.0) at 37°C. Finally, sections were washed for 10 min each in 2×-, 0.1×-, and 0.05×-concentrated SSC at 50°C. The detection of digoxygenin-labelled hybrids followed the protocol of Roche (Mannheim, Germany). Incubation times for the staining reaction ranged from 2 to up to 24 h. Under these conditions, no binding of non-relevant sense transcripts was observed.

Histochemistry

In frozen liver sections (10 μm), glycogen was visualised with a periodic acid-Schiff (PAS) stain and SDH (succinate dehydrogenase) was detected as a periportal marker enzyme using succinate and tetrazolium as substrates as described (Lojda et al. 1976).

Tools and assays

Total RNA from rat livers was isolated by the use of the RNeasy kit according to the protocol of the supplier (Qiagen, Hilden, Germany). For Northern blot analyses, 5 μg of total RNA each were separated in a denaturing agarose gel, and PCK1 or α2M mRNA detected by the use of the same probes as for ISH with the exception that the probes were not subjected to alkaline hydrolysis. Hybrids were identified essentially as described (Christ et al. 1989).

PCK1 enzyme activity was determined as described (Seubert and Huth 1965). Phosphorylase a enzyme activity and glycogen content were measured according to standard procedures (Bergmeyer 1974).

Results

Induction of the hepatic acute phase reaction by IL6

An increase in acute phase protein expression is a major feature of the hepatic acute phase response. In the rat, α2M is the prominent acute phase protein elevated maximally 8 h after injection of IL6, the main mediator of the acute phase reaction (Geiger et al. 1988). Hence, in the present study, we injected IL6 into rats to trigger a hepatic acute phase response at 6 a.m. in the morning. After 8 h, levels of α2M mRNA were increased in untreated animals from nearly undetectable levels at 6 a.m. to 27% of the increase in the presence of IL6, which was set to 100% in each single experiment (Fig. 1a). In order to demonstrate whether all hepatocytes contributed to this increase, the tissue distribution of α2M mRNA was investigated by ISH in liver sections prepared from livers before injection of IL6 at 6 a.m. and 8 h after injection at 2 p.m. To discriminate between periportal and perivenous areas in the liver lobule, sections were stained histochemically for succinate dehydrogenase, which is a periportal marker enzyme. α2M mRNA expression was predominant in periportal over perivenous hepatocytes. While in untreated animals expression was barely increased at 2 p.m. as compared to 6 a.m., α2M mRNA levels were clearly elevated 8 h after injection of IL6. This increase was mainly due to the enhancement of α2M mRNA in periportal areas of the liver, which, however, extended into midzonal regions (Fig. 1b).

Fig. 1
figure 1

IL6-induced increase in α2M mRNA in periportal hepatocytes. Rats, kept under the normal feeding rhythm were left either untreated or injected with rhIL6 (0.02 μg/kg body weight) intraperitoneally at 6 a.m. and livers excised after another 8 h (2 p.m.) for the analyses of α2M mRNA by Northern blot hybridisation (a) or ISH (b) by the use of digoxygenin-labeled RNA probes complementary to α2M mRNA. In b succinate dehydrogenase (SDH) was stained histochemically to identify periportal regions in liver sections adjacent to sections shown in the row above. In a values are means ± SEM from 5 to 7 separate experiments. * Values are different from untreated controls at the P ≤ 0.01 level (Student’s t test for unpaired values). pv Perivenous; pp periportal

Full size image

Changes in the hepatic carbohydrate metabolic capacity by IL6

In rat liver, PCK mRNA levels are low at 6 a.m. and increase during the light period reaching highest levels between noon and 6 p.m. (Eilers et al. 1995). In the present study, this finding was confirmed. Yet, in the presence of IL6 this increase was attenuated and levels were only in the range of 30% of that in untreated animals (Fig. 2a). At 6 a.m., periportal hepatocytes contained higher amounts of PCK mRNA than hepatocytes surrounding the central vein as demonstrated by co-localization with succinate dehydrogenase in serial sections. The amount further increased during the fasting period until noon. This increase was attenuated in animals treated with IL6 (Fig. 2b). During that period in time, i.e. from 6 a.m. to noon, PCK1 enzyme activity was not affected by IL6 (data not shown).

Fig. 2
figure 2

IL6-induced inhibition of PCK1 mRNA increase in periportal hepatocytes. Rats were treated as described in the legend to Fig. 1. Livers were excised 6 h after injection of rhIL6 and analysed for PCK1 mRNA by Northern blot hybridisation (a) or ISH (b) by the use of digoxygenin-labeled RNA probes complementary to PCK1 mRNA. In b succinate dehydrogenase (SDH) was stained histochemically to identify periportal regions in liver sections adjacent to sections shown in the row above. In a values are means ± SEM from 5 to 7 separate experiments. * Values are different from untreated controls at the P ≤ 0.025 level (Student’s t test for unpaired values). pv Perivenous, pp periportal

Full size image

At the onset of the fasting period, from 6 to 8 a.m., glycogen levels in livers of untreated animals increased very likely due to the uptake of carbohydrates during the feeding period at night. This increase was attenuated in animals treated with IL6. From 8 a.m. to 2 p.m., glycogen levels decreased again to 75% of the starting levels at 6 a.m. in untreated animals, but to only 58% in animals treated with IL6 (Fig. 3a). Thus, glycogen depletion in IL6-treated rat livers was significantly faster (5,25%/h) as compared to untreated control livers (3,10%/h). These results were corroborated by histochemical staining of glycogen in liver sections of rats at 6 a.m. and 2 p.m. While at 6 a.m. glycogen levels were higher in periportal areas, at 2 p.m. they were lower in the periportal as compared to the perivenous areas in untreated animals but even lower in animals treated with IL6 (Fig. 3b). The accelerated glycogen depletion in livers of animals treated with IL6 could be due to the stimulation of phosphorylase a enzyme activity by IL6. However, phosphorylase a activity was not significantly different in IL6-treated from untreated animals (758 ± 15 vs. 785 ± 53 μmol × h−1 × g wet wt.−1).

Fig. 3
figure 3

IL6-induced depletion of hepatic glycogen in periportal hepatocytes. Rats were treated as described in the legend to Fig. 1. Livers were excised 2 (8 a.m.) and 8 h (2 p.m.) after injection of rhIL6 and analysed for glycogen (Gg) content biochemically (a) or histochemically (b). In b succinate dehydrogenase (SDH) was stained histochemically to identify periportal regions in liver sections adjacent to sections shown in the row above. In a values are means ± SEM from 5 to 7 separate experiments. * Values are different from untreated controls at the P ≤ 0.01 level (Student’s t test for unpaired values). pv Perivenous, pp periportal

Full size image

Discussion

Mechansim of IL6 inhibition of PCK gene expression

The PCK1 gene in the liver is expressed to higher levels in periportal as compared to perivenous hepatocytes, which is also true for other gluconeogenic enzymes such as fructose 1,6-bisphosphatase, and glucose 6-phosphatase. Reciprocally, glycolytic enzymes like glucokinase and pyruvate kinase are predominant in the perivenous over the periportal zone. This situation is a basic feature of the concept of “metabolic zonation” saying that gluconeogenesis is preferentially located in periportal and glycolysis in perivenous hepatocytes of the liver lobule (Jungermann and Katz 1989; Jungermann and Thurman 1992; Jungermann 1988; Gumucio 1989; Quistorff 1990). The expression of enzymes of carbohydrate metabolism changes during the daily feeding rhythm in rats. In the feeding (dark) period, mRNA abundance and enzyme activity of PCK1 is low and glucokinase mRNA and enzyme activity is high, while conversely, in the fasting (light) period PCK1 is high and glucokinase expression is low, respectively (Bartels et al. 1989; Eilers et al. 1995, 1993). In the present study, IL6 inhibited the increase in PCK1 mRNA during the initial phase of fasting, from 6 a.m. to noon in the periportal zone of the liver (Fig. 2). Yet, enzyme activity was not affected. This, however, is not surprising because the increase in enzyme activity follows the physiological increase in mRNA with a time lag of 6 h at least (Eilers et al. 1993), a period in time, which was not investigated in the present study.

A few studies only exist describing the impact of the acute phase reaction on PCK gene expression in vivo. In mouse liver, the induction by cortisone of PCK1 enzyme activity was inhibited by bacterial endotoxin (Rippe and Berry 1972; Berry et al. 1968). The proinflammatory cytokines IL6 and TNFα decreased PCK1 gene transcription in rat and mouse liver, respectively (Hill and McCallum 1991, 1992). In vitro, in cultured rat hepatocytes, the glucagon-dependent induction of PCK1 gene transcription was inhibited and PCK1 mRNA degradation was accelerated by rhIL6, rhIL1β and rhTNFα (Christ et al. 1994, 1997; Christ and Nath 1996). Thus, the major action of cytokines on PCK1 gene expression is operative at the transcriptional and posttranscriptional level leading to the provision of nucleotides by inhibition of gene transcrition and acceleration of mRNA degradation.

Concept of the “Molecular economy” of the hepatic acute phase reaction

During the acute phase reaction, the liver is challenged by proinflammatory cytokines to synthesize positive acute phase proteins requiring an increased demand in amino acids and nucleotides. The demand in amino acids can be accomplished by external support from the cytokine-stimulated increase in muscle proteolysis and hepatic uptake (Andus et al. 1991, 1993; Klasing 1988). Yet, the demand in nucleotides can only be met by hepatocellular resources, which may either be the increase in nucleotide synthesis, again demanding for hepatocyte metabolic capacity, or the reduction of expression of genes not required in this acute situation. Therefore, the inhibition of the expression of the negative acute phase proteins is regarded to serve this demand (Schreiber 1987), which, however, may be not sufficient. Hence, in addition the expression of metabolic enzymes is reduced to accomplish the required demand in nucleotides. These comprise enzymes of gluconeogenesis and glycolysis (Christ et al. 1994, 1997; Christ and Nath 1996; Hill and McCallum 1991, 1992), cholesterol synthesis (Feingold et al. 1995), xenobiotics metabolism (Sewer et al. 1996; Morgan et al. 1994), which might be essential for the maintenance of the burst synthesis of acute phase proteins on the expense of negative acute phase proteins and down-regulation of hepatocyte-specific functions. This can be regarded as a kind of economy on the molecular level (molecular economy) providing vast synthetic capacity on sudden demand. To the present knowledge, it can be assumed that on the molecular level an intensive crosstalk between glucoregulatory hormone signal chains and the signal chains of inflammatory cytokines is operative (Ramadori and Christ 1999). Hence, it may be supposed that the increase in α2M mRNA induced by IL6 and the simultaneous reduction in PCK mRNA may serve this molecular economy. During the initial acute phase reaction, this might be beneficial for the host defence reaction. However, during prolonged sepsis this is crucial and may cause the dysregulation of metabolic homeostasis.

Metabolic implications of IL6 inhibition of PCK gene expression

During prolonged sepsis, the increased utilization of glucose at the site of inflammation, trauma or tissue lesions is favoured by a hormonal milieu to deplete body glucose stores comprising elevated catecholamine and glucagon serum levels with simultaneous insulin resistance (Mizock 1995; Memon et al. 1994; Goto et al. 1994). From the present results it may be hypothesised that the inhibition of PCK1 mRNA increase during the normal feeding rhythm serves to provide nucleotides for the acute phase protein synthesis. This might not impair gluconeogenesis, which is supported by the fact that PCK1 enzyme activity was not affected by IL6 in the time period investigated. Indeed, in cultured rat hepatocytes, the gluconeogenic rate from lactate was 0.63 ± 0.09 μmole glucose × g−1 w wt. × h−1 in the absence and 0.64 ± 0.31 in the presence of IL6 (unpublished results). Hence, the acceleration of hepatic glycogen degradation alone may contribute to the increased glucose demand during IL6-induced initial acute phase reaction. Since phosphorylase a activity was not affected, this was very likely due to the inhibition of glycogen synthesis corroborating previous findings (Kanemaki et al. 1998). This situation may not disturb the hepatocyte glucogenic capacity as long as PCK1 enzyme activity is not decreasing. Yet, the half-life of PCK1 protein to be expected in vivo was in the range of 5–13 h (Hopgood et al. 1973). Thus, during the prolonged acute phase reaction, the exhaustion of hepatic glycogen pools and the inhibition of PCK1 gene expression results in the lowered capacity of the liver to supply glucose, which may contribute to the emergence of hypoglycemia, one major complication in the pathogenesis of the septic shock (Filkins 1985).