Search

The liver has a multitude of functions which are necessary to maintain whole body homeostasis. This requires that various metabolic pathways can run in parallel in the most efficient manner and that futile cycles are kept to a minimum. To a large extent this is achieved due to a functional specialization of the liver parenchyma known as metabolic zonation which is often lost in liver diseases. Although this phenomenon is known for about 40 years, the underlying regulatory pathways are not yet fully elucidated. The physiologically occurring oxygen gradient was considered to be crucial for the appearance of zonation; however, a number of reports during the last decade indicating that β-catenin signaling, and the hedgehog (Hh) pathway contribute to metabolic zonation may have shifted this view. In the current review we connect these new observations with the concept that the oxygen gradient within the liver acinus is a regulator of zonation. This is underlined by a number of facts showing that the β-catenin and the Hh pathway can be modulated by the hypoxia signaling system and the hypoxia-inducible transcription factors (HIFs). Altogether, we provide a view by which the dynamic interplay between all these pathways can drive liver zonation and thus contribute to its physiological function., (Copyright © 2017 The Authors. Published by Elsevier B.V. All rights reserved.)

Oxygen has important functions as substrate for biochemical reactions and as modulator of gene expression. In the liver, the physiologically occurring oxygen gradient is a major effector of metabolic zonation. In addition, cross-talks between the O2 signaling and nutrient signaling chains initiate a dynamic zonation pattern. Under pathological situations, hypoxia appears to be a major determinant for liver diseases and cancer. Thereby transcription factors of the HIF family are activated whereas USF proteins have the potential to counteract HIFs. In addition, feedback mechanisms between hypoxia, HIF and the IGF axes appear to exist. Thus, the knowledge of these mechanisms may help to initiate new therapies in diseases with disturbed O2 availability.

Higher levels of IGF-binding protein 1 (IGFBP-1) mRNA are expressed in the less aerobic perivenous zone of the liver. Because gradients in oxygen tension (pO(2)) may contribute to zonated gene expression, the influence of arterial and venous pO(2) on IGFBP-1 biosynthesis was studied in primary cultures of rat hepatocytes. Maximal IGFBP-1 mRNA and protein levels were observed under venous pO(2), whereas less than 30% of maximal levels were observed under arterial pO(2). In contrast, the expression of IGFBP-4 was greatest under arterial pO(2), indicating that this effect of hypoxia on IGFBP-1 gene expression is specific. The response to hypoxia appears to involve reactive oxygen species, because treatment with H(2)O(2) results in a dose-dependent decrease of IGFBP-1 mRNA levels under venous pO(2), whereas IGFBP-1 mRNA expression under arterial pO(2) was not affected. Inhibition of the hypoxia-dependent IGFBP-1 mRNA induction by actinomycin D indicates that this effect is mediated at the level of gene transcription, and inhibition of IGFBP-1 mRNA by the iron chelator desferrioxamine under both venous and arterial pO(2) suggested the involvement of hypoxia-inducible transcription factors (HIF). Transfection experiments demonstrated that especially HIF-3alpha and HIF-2alpha, and to a lesser extent HIF-1alpha, contribute to the induction of IGFBP-1 mRNA expression in isolated hepatocytes, whereas experiments with vectors for the HIF prolyl hydroxylases (PHD) indicated a major role of PHD-2 in destabilization of HIFs, attenuating the induction of IGFBP-1 under venous pO(2). Reporter gene studies indicate that hypoxia stimulates IGFBP-1 expression through a putative HIF response element located approximately 250 bp upstream from the transcription initiation site. Together, these results support the concept that iron, radical oxygen species, and the HIF-2 and -3 as well as the PHD pathways play important roles in mediating effects of hypoxia on IGFBP-1 gene expression in the liver.

The substrates oxygen and glucose are important for the appropriate regulation of metabolism, angiogenesis, tumorigenesis and embryonic development. The knowledge about an interaction between these two signals is limited. We demonstrated that the regulation of glucagon receptor, insulin receptor and L-type pyruvate kinase (L-PK) gene expression in liver is dependent upon a cross-talk between oxygen and glucose. The periportal to perivenous drop in O2 tension was proposed to be an endocrine key regulator for the zonated gene expression in liver. In primary rat hepatocyte cultures, the expression of the glucagon receptor and the L-PK mRNA was maximally induced by glucose under arterial pO2 whereas the insulin receptor was maximally induced under perivenous pO2. It was demonstrated for the L-PK gene that the modulation by O2 of the glucose-dependent induction occured at the glucose-responsive element (Glc(PK)RE) in the L-PK gene promoter. The reduction of the glucose-dependent induction of the L-PK gene expression under venous pO2 appeared to be mediated via an interference between hypoxia-inducible factor 1 (HIF-1) and the glucose-responsive transcription factors at the Glc(PK)RE. The glucose response element (GlcRE) also functioned as a hypoxia response element and, vice versa, a hypoxia-responsive element was functioning as a GlcRE. Thus, our findings implicate that the cross-talk between oxygen and glucose might have a fundamental role in the regulation of several physiological and pathophysiological processes.

P-Glycoprotein transporters encoded by mdr1 (multidrug resistance) genes mediate extrusion of an array of lipophilic xenobiotics from the cell. In rat liver, mdr transcripts have been shown to be expressed mainly in hepatocytes of the periportal region. Since gradients in oxygen tension (pO(2)) may contribute towards zonated gene expression, the influence of arterial and venous pO(2) on mRNA expression of the mdr1b isoform was examined in primary rat hepatocytes cultured for up to 3 days. Maximal mdr1b mRNA levels (100%) were observed under arterial pO(2) after 72 h, whereas less than half-maximal mRNA levels (40%) were attained under venous pO(2). Accordingly, expression of mdr protein and extrusion of the mdr1 substrate rhodamine 123 were maximal under arterial pO(2) and reduced under venous pO(2). Oxygen-dependent modulation of mdr1b mRNA expression was prevented by actinomycin D, indicating transcriptional regulation. Inhibition of haem synthesis by 25 microM CoCl(2) blocked mdr1b mRNA expression under both oxygen tensions, whereas 80 microM desferrioxamine abolished modulation by O(2). Haem (10 microM) increased mdr1b mRNA levels under arterial and venous pO(2). In hepatocytes treated with 50 microM H(2)O(2), mdr1b mRNA expression was elevated by about 1.6-fold at venous pO(2) and 1.5-fold at arterial pO(2). These results support the conclusion that haem proteins are crucial for modulation of mdr1b mRNA expression by O(2) in hepatocyte cultures and that reactive oxygen species may participate in O(2)-dependent signal transduction. Furthermore, the present study suggests that oxygen might be a critical modulator for zonated secretion of mdr1 substrates into the bile.

Insulin stimulates glucose utilization in the liver, which occurs mainly in the less aerobic, perivenous, zone. Accordingly, the insulin receptor protein was predominantly expressed in this area, although the insulin receptor mRNA was homogeneously distributed. In hepatocyte cultures venous O(2) partial pressure (pO(2)) induced insulin receptor protein expression. High glucose concentrations enhanced insulin receptor protein under arterial and venous pO(2). The induction of insulin receptor protein by venous pO(2) would explain its zonated expression.

Treatment of male rats with phenobarbital (PB) results in a perivenous and mid-zonal pattern of cytochrome P-450 (CYP)2B1 mRNA expression within the liver acinus. The mechanism of this zonated induction is still poorly understood. In this study sinusoidal gradients of oxygen and epidermal growth factor (EGF) besides those of the pituitary-dependent hormones growth hormone (GH), thyroxine (T4), and triiodothyronine (T3) were considered to be possible determinants for the zonated induction of the CYP2B1 gene in liver. Moreover, heme proteins seem to play a key role in oxygen sensing. Therefore, the influence of arterial (16% O2) and venous (8% O2) oxygen tension (pO2), and of the heme synthesis inhibitors CoCl2 and desferrioxamine (DSF) on PB-dependent CYP2B1 mRNA induction as well as the repression by EGF and, for comparison, by GH, T4, and T3, of the induction under arterial and venous pO2 were investigated in primary rat hepatocytes. Within 3 days, phenobarbital induced CYP2B1 mRNA to maximal levels under arterial pO2 and to about 40% of maximal levels under venous pO2. CoCl2 annihilated induction by PB under both oxygen tensions, whereas desferrioxamine and heme abolished the positive modulation by O2, suggesting that heme is a necessary component for O2 sensing. EGF suppressed CYP2B1 mRNA induction by PB only under arterial but not under venous pO2, whereas GH, T4, and T3 inhibited induction under both arterial and venous pO2. Thus, in hepatocyte cultures, an O2 gradient in conjunction with EGF mimicked the perivenous induction by PB of the CYP2B1 gene observed in the liver in vivo.

The glucagon-stimulated transcription of the cytosolic phosphoenolpyruvate carboxykinase-1 (PCK1) gene is mediated by cAMP and positively modulated by oxygen in primary hepatocytes. Rat hepatocytes were transfected with constructs containing the first 2500, 493 or 281 bp of the PCK1 5'-flanking region in front of the chloramphenicol acetyltransferase (CAT) reporter gene. With all three constructs glucagon induced CAT activity with decreasing efficiency maximally under arterial pO2 and to about 65% under venous pO2. Rat hepatocytes were then transfected with constructs containing the first 493 bp of the PCK1 5'-flanking region in front of the luciferase (LUC) reporter gene, which were block-mutated at the CRE1 (cAMP-response element-1; -93/-86), putative CRE2 (-146/-139), promoter element (P) 1 (-118/-104), P2 (-193/-181) or P4 (-291/-273) sites. Glucagon induced LUC activity strongly when the P1 and P2 sites were mutated and weakly when the P4 site was mutated; induction of the P1, P2 and P4 mutants was positively modulated by the pO2. Glucagon also induced LUC activity strongly when the putative CRE2 site was altered; however, induction of the CRE2 mutant was not modulated by the pO2. Glucagon did not induce LUC activity when the CRE1 site was modified. These experiments suggested that the CRE1 but not the putative CRE2 was an essential site necessary for the cAMP-mediated PCK1 gene activation by glucagon and that the putative CRE2 site was involved in the oxygen-dependent modulation of PCK1 gene activation. To confirm these conclusions rat hepatocytes were transfected with simian virus 40 (SV40)-promoter-driven LUC-gene constructs containing three CRE1 sequences (-95/-84), three CRE2 sequences (-148/-137) or three CRE1 sequences plus two CRE2 sequences of the PCK1 gene in front of the SV40 promoter. Glucagon induced LUC activity markedly when the CRE1, but not when the CRE2, sites were in front of the SV40-LUC gene; however, induction of the (CRE1)3SV40-LUC constructs was not modulated by the pO2. Glucagon also induced LUC activity very strongly when the CRE1 and CRE2 sites were combined; induction of the (CRE1)3(CRE2)2SV40-LUC constructs was positively modulated by the pO2. These findings corroborated that sequences of the putative CRE2 site were responsible for the modulation by oxygen of the CRE1-dependent induction by glucagon of PCK1 gene transcription.

H2O2 mimicked the action of periportal pO2 in the modulation by O2 of the glucagon-dependent activation of the phosphoenolpyruvate carboxykinase (PCK) gene and the insulin-dependent activation of the glucokinase (GK) gene. H2O2 can be converted in the presence of Fe2+ in a Fenton reaction into hydroxyl anions and hydroxyl radicals (.OH). The hydroxyl radicals are highly reactive and might interfere locally with transcription factors. It was the aim of the present study to investigate the role of and to localize such a Fenton reaction. Hepatocytes cultured for 24 h were treated under conditions mimicking periportal or perivenous pO2 with glucagon or insulin plus the iron chelator desferrioxamine (DSF) or the hydroxyl radical scavenger dimethylthiourea (DMTU) to inhibit the Fenton reaction. PCK mRNA was induced by glucagon maximally under conditions of periportal pO2 and half-maximally under venous pO2. GK mRNA was induced by insulin with reciprocal modulation by O2. DSF and DMTU reduced the induction of PCK mRNA to about half-maximal and increased the induction of GK mRNA to maximal under both O2 tensions. Hydroxyl radical formation was maximal under arterial pO2. Perivenous pO2, DSF and DMTU each decreased the formation of .OH to about 70% of control. The Fenton reaction could be localized in a perinuclear space by confocal laser microscopy and three-dimensional reconstruction techniques. In the same compartment, iron could be detected by electron-probe X-ray microanalysis. Thus a local Fenton reaction is involved in the O2 signalling, which modulated the glucagon- and insulin-dependent PCK gene and GK gene activation.

The different endowment with key enzymes and thus different metabolic capacities of periportal and perivenous cell types led to the model of "metabolic zonation." The periportal and perivenous hepatocytes receive different signals owing to the decrease of substrate concentrations including O2 and hormone levels during passage of blood through the liver sinusoids. These different signal patterns should be important for the short-term regulation of metabolism and also for the long-term induction and maintenance of the different enzyme pathways by control of gene expression. The periportal to perivenous drop in oxygen tension was considered to be a key regulator in the zonated expression of carbohydrate-metabolizing enzymes. In primary hepatocyte cultures, glucagon activated the phosphoenolpyruvate carboxykinase (PCK) gene to higher levels under arterial than under venous oxygen. The insulin-dependent activation of the glucokinase (GK) gene was reciprocally modulated by oxygen. Exogenously added hydrogen peroxide mimicked the effects of arterial oxygen on both the glucagon-dependent PCK gene and the insulin-dependent GK activation. Therefore, the oxygen sensor could be a hydrogen peroxide-producing oxidase which could contain a heme group for "measuring" the O2 tension. This notion was corroborated by the finding that CO mimicked the positive effect of O2 on PCK gene activation. Transfection of PCK promoter-CAT gene constructs into primary hepatocytes showed that the oxygen modulation of the PCK gene activation occurred in the region -281/+69. The modulation by O2 was not mediated by isolated cAMP-responsive elements. Nuclear protein extracts prepared from hepatocytes cultured under venous PO2 as compared to arterial PO2 showed an enhanced binding activity to the promoter fragment -149/-43. Oxidative conditions such as H2O2 reduced the DNA-binding activity, thus supporting the role of H2O2 as a mediator in the O2 response of the PCK and GK genes.

The glucagon-dependent activation of the phosphoenolpyruvate carboxykinase (PCK) gene within two hours is modulated by O2 in rat hepatocytes. It was the aim of the present study to test if this short-term modulation by O2 of the glucagon induction might be influenced by long-term culture of hepatocytes for 24 hours under different O2 tensions prior to glucagon induction. Cells were precultured for 24 hours at arterial O2 (16% O2) or venous O2 (8% O2), then induced within two to four hours with 1 nM glucagon each at arterial or venous O2. In arterial O2 precultured cells PCK mRNA and activity were induced to 100% at arterial O2 and to about 60% at venous O2. In venous O2 precultured cells PCK mRNA and activity were induced only to about 70% at arterial O2 and to about 60% at venous O2. Transfected PCK promoter (-2500)-CAT constructs were activated by glucagon with the same long-term modulatory effects of oxygen as the endogenous PCK gene. Gel mobility shift assays with nuclear extracts prepared from hepatocytes and a PCK promoter fragment ranging from -149 to -42 bp revealed one complex with a higher DNA binding activity when extracts of cells precultured for 24 hours under venous O2 as compared to arterial O2 were used. Therefore, the short-term modulation by O2 of PCK gene activation by glucagon was widely lost during preculture at low O2. This diminution of O2 sensitivity of PCK induction may be due to a nuclear protein or proteins which are induced by perivenous O2 tensions and bind to the PCK promoter.

Heme proteins acting as oxidases which produce H2O2 have been proposed to function as O2 sensors. In order to find out whether the modulation by O2 of PCK gene activation and the stimulation of the ALD A gene by venous O2 operate via H2O2, the effects of different concentrations of H2O2 and catalase as H2O2 scavenger were studied in rat hepatocyte cultures under different O2 tensions. Primary hepatocytes were treated with 0.1 nM glucagon, 50 microM H2O2 and/or 100 micrograms/ml catalase each at arterial O2 or venous pO2. PCK mRNA was induced by glucagon maximally under arterial O2 and only half maximally under venous O2. ALD A mRNA was induced only by venous O2. H2O2 enhanced the induction of PCK mRNA to similar levels under venous O2 tensions and the induction of ALD A mRNA under both O2 was completely inhibited. Addition of catalase antagonized the actions of H2O2 completely. These findings support the hypothesis that an H2O2-generating heme protein is involved in the O2 sensing system regulating gluconeogenic and glycolytic gene expression in response to O2.

Hyperoxia induces the expression of the hemopexin (Hx) gene in the liver in vivo. To investigate whether the Hx gene is activated by oxygen as such or via H2O2 as an oxygen signal transmitter the effects of arterial and venous O2 tensions as well as different concentrations of H2O2 on Hx mRNA expression were studied. After preculturing primary rat hepatocytes for 24 h at arterial O2 (16%) Hx mRNA was expressed with a maximal level (= 100%), when arterial O2 tension proceeded for 2 h, and to values of approximately 50%, when venous O2 tension (8%) proceeded for 2 h. When hepatocytes were precultured for 24 h under venous O2, Hx mRNA was induced by arterial O2 to values of 60% and under venous O2 to values of approximately 35%. The expression of beta-actin remained unchanged under arterial and venous O2. Exposure of hepatocyte cultures to H2O2 decreased the expression of Hx mRNA in a dose-dependent manner after 2 h, while heme oxygenase-1 (HO-1) mRNA was induced 2.5 fold. The results suggest that O2 per se rather than the reactive oxygen intermediate H2O2 modulates Hx expression.

Oxygen modulates the glucagon-dependent activation of the phosphoenolpyruvate carboxykinase (PCK) gene. The respiratory chain or heme proteins have been proposed to function as O2-sensors. The functions of the respiratory chain are impaired by uncouplers such as 2,4-dinitrophenol (DNP); those of ferro-heme proteins are affected by carbon monoxide (CO), which locks heme in the oxy conformation. Therefore, the effects of different concentrations of CO and DNP on the glucagon-dependent induction of PCK mRNA and PCK activity were investigated at different physiological oxygen tensions in primary rat hepatocyte cultures. The cells were cultured under standard conditions from 4-24 h. After addition of fresh media PCK was induced with 1 nM glucagon. PCK mRNA and PCK activity were elevated after 2h and 3h, respectively, to 100% at 16% O2 (mimicking arterial oxygen tensions) and to about 60% at 8% O2 (mimicking venous oxygen tensions). CO counteracted the reduced induction at lower oxygen tensions: Under 8% O2 + 2% CO PCK mRNA could be elevated again to about 90% and PCK activity to about 80%. CO did not impair the induction by insulin of ornithine decarboxylase (ODC) and the incorporation of 14C-leucine into total protein. CO did not cause lactate dehydrogenase (LDH) to leak from the cells or influence the cell structures at the microscopical level. DNP (50 microM) unspecifically lowered PCK gene expression without affecting its modulation by oxygen. These results are in line with the proposal that a ferro-heme protein rather than the respiratory chain acted as an O2 sensor in the activation of the PCK gene.

The glucagon-dependent activation of the phosphoenolpyruvate carboxykinase (PCK) gene is modulated by oxygen. It was proposed that heme proteins might function as O2 sensors; their actions are impaired after replacement of the central Fe2+ ion by Co2+ and inhibition of heme synthesis by succinylacetone (SA). Therefore, the effects of CoCl2 and SA, alone and in combination, on the glucagon-dependent induction of PCK activity and PCK mRNA were investigated at different physiological oxygen tensions in primary rat hepatocyte cultures. The cells were exposed to 50 microM CoCl2 and/or 2 mM SA from 4-24 h. After addition of fresh media without CoCl2 or SA, PCK was induced with 1 nM glucagon. PCK activity and PCK mRNA were elevated to 100% at 16% O2 and to about 65% at 8% O2. CoCl2 reduced these increases to about 45% at 16% O2 and to about 35% at 8% O2. SA lowered the inductions to about 50% and 40% each at 16% and 8% O2. CoCl2 plus SA diminished the elevations to about 5% at both oxygen tensions. In the presence of CoCl2 and/or SA, ornithine decarboxylase induction by insulin was not impaired; lactate dehydrogenase did not leak from the cells, which in electron microscopical inspections had normal cell structures. These findings support the hypothesis that a heme protein is involved in the activation of the PCK gene and that it acts as an O2 sensor.