We studied the metabolic rate, cellular energetic state, hypoxia-inducible factor-1 (HIF-1) activation, and expression of enzymes involved in energy metabolism using rainbow trout (Oncorhynchus mykiss) hepatocytes over the oxygen range from 21 to 1 kPa. Oxygen dependence of these factors was assessed by gradually reducing oxygen supply to cells from 21 kPa to 10, 5, 2, and 1 kPa. Moreover, time course experiments for up to 20 h at oxygen tensions of 1 and 2 kPa were carried out. Reduction of oxygen from 21 kPa to 10, 5, 2, and 1 kPa decreased metabolic rate of the cells by 14, 24, 37, and 46%, respectively. This response was instantaneous and fully reversible upon reoxygenation. Cellular ATP content and the expression of all mRNAs studied decreased when oxygen was reduced from 21 to 5 and 2 kPa. The lowest ATP levels, ∼43% of the initial value, were measured at 5 kPa of oxygen, whereas the reduction in mRNA amounts was most pronounced at 2 kPa. At 1 kPa oxygen tension, both ATP content and mRNA amounts returned to normoxic (21 kPa) levels with a concomitant activation of HIF-1, indicating reorganization of energy metabolism in adaptation of cells to low oxygen supply. These results show that oxygen has a direct regulatory effect on metabolism of trout hepatocyte cultures, supporting the view that oxygen has a profound role in metabolic regulation in cells.
- metabolic regulation
- energy metabolism
- hypoxia-inducible factor-1
cells can survive periods of inadequate oxygenation by a number of adaptive responses. These responses, comprising rearrangements in metabolism, membrane transport, and gene expression, adjust cellular energy demand to lowered capacity for aerobic energy production (6, 19). The most potent defense against hypoxia is metabolic suppression, also known as oxygen conformance. In this phenomenon, cellular energy demand and supply are coordinately downregulated at oxygen levels well above those limiting aerobic energy metabolism (6, 19). Although metabolic suppression is considered as a hallmark of hypoxia tolerance, it also takes place, to some degree, in cells of hypoxia-intolerant animals (2, 17, 29, 48). In these cells, respiration is typically suppressed at oxygen tensions of 1–3 kPa, which is at least one order of magnitude higher than the oxygen tensions that limit oxidative phosphorylation (44).
Gene expression in hypoxia is modulated both transcriptionally and posttranscriptionally. The major transcriptional regulator of hypoxia-inducible gene expression is hypoxia-inducible factor-1, HIF-1 (56). The oxygen sensitivity of HIF-1 function is primarily caused by oxygen-dependent hydroxylation of conserved proline residues in the α-subunit, which targets the protein for proteasomal degradation and thus prevents its accumulation in normoxia (21, 22). Accumulation of HIF-1α, typically observed at oxygen tensions of 1–2 kPa in mammalian cells, leads to binding of the active HIF-1 dimer to hypoxia-responsive elements (HRE) in its target genes and subsequent enhancement of gene transcription (51). More than 70 HIF-1-regulated genes, including many genes of glycolytic enzymes, have been identified in mammals (56). Yet, more genes are transcriptionally repressed than induced in hypoxia (14). If the oxygen level is in the range of 0.02–1 kPa, this downregulation of gene expression is enhanced by global arrest of translation, which is mediated by inhibition of eukaryotic initiation factor 2α (26).
However, cells in animal tissues routinely encounter oxygen tensions at a range from under 1 kPa to ∼15 kPa, depending on the tissue, its blood supply, and the distance of the cell from the blood capillaries. Little is known about the oxygen-dependent responses of cells to these higher oxygen tensions. Interestingly, in skeletal muscle, metabolic rate appears to be determined by oxygen availability, even in hyperoxic oxygen tensions (20, 57). Moreover, when oxygen supply was slowly decreased from 16 kPa over several hours, isolated rat hepatocytes suppressed respiration significantly at oxygen tensions as high as 9.3 kPa (48). These studies suggest that changes in oxygen availability may regulate metabolism and gene expression in cells well above the oxygen levels that are traditionally considered hypoxic (i.e., associated with inadequate oxygenation).
To shed light on this, we studied the oxygen dependence of metabolic rate, energetic state of the cells, and expression of enzymes involved in energy metabolism in rainbow trout (Oncorhynchus mykiss) hepatocytes over the oxygen tension range from 21 to 1 kPa (21–1% oxygen in the gas phase). We also assessed the oxygen dependence of HIF-1α accumulation and DNA-binding activity in these cells. Moreover, to separate immediate and more sustained responses, time course experiments for up to 20 h at reduced oxygen tension were carried out. Hepatocytes represent a metabolically active cell type. Moreover, trout hepatocytes cultured as monolayers maintain their differentiated structures and functions for several days after isolation (50). With a few exceptions, the oxygen dependence of cellular energetics has been studied with suspended cells or mitochondria using closed chamber respirometry. To provide more physiological conditions, we used monolayer cultures of cells, and with these, we conducted long-term real-time measurement of overall metabolic rate, i.e., both aerobic and anaerobic metabolism, by use of direct calorimetry with constant flowthrough of perfusion medium.
MATERIALS AND METHODS
Rainbow trout (O. mykiss, wt 70–160 g) were obtained from the Finnish Institute for Fisheries and Environment (Parainen, Finland) and were maintained at 12°C with a 12:12-h light-dark cycle. The fish were held in laboratory conditions for at least 1 wk before cell isolations. The fish were fed trout pellets (TESS, Raisio, Finland) regularly but were fasted for 24 h before cell isolation. All procedures with animals were done with approval of the local committee for animal experimentation.
Isolation and culture of hepatocytes.
Hepatocyte isolation was carried out by the two-step collagenase perfusion method (49) as described previously (45). After isolation, hepatocytes were suspended in Leibovitz L-15 culture medium containing 10% FBS, 100 IU/ml penicillin, and 100 μg/ml streptomycin (all from GIBCO-BRL Life Technologies, Paisley, UK). Cells at a density of 2–3 × 106 cells/ml, viability exceeding 90%, were plated on poly-l-lysine (Sigma Chemical, St. Louis, MO)-coated culture bottles or glass microplates (Thermometric, Järfalla, Sweden) and were cultured under atmospheric air (21% oxygen) at 20°C for 2–3 days before experiments. Because a confluent monolayer of cells may consist of several cell layers, there is a potential for oxygen gradients within the layer. Thus the cells were grown to subconfluence.
The metabolic rate of hepatocytes under different oxygen levels was measured using a twin heat-conduction microcalorimeter TAM 2277 (Thermometric) equipped with flowthrough titration/perfusion vessels and stainless steel reference ampoules with volumes of 4 ml (36). The thermostatted water bath of the calorimeter was kept at 20°C, and the calorimeter was electrically calibrated before the experiments (36). The time constants of the two measuring cylinders used were 92–102 s, and the baseline stabilities >8 h ± 0.1 μW.
For the experiments, two microplates of subconfluent monolayers of cells were placed in a plateholder in a measuring ampoule filled with serum-free MEM (GIBCO), pH 7.6, supplemented with 100 IU/ml penicillin, 100 μg/ml streptomycin, 2 mmol/l l-glutamine, and 25 mmol/l HEPES. The reference ampoule was filled with the same medium, and both ampoules were lowered in measuring position by a three-step procedure (36). Already during this prethermostation stage, cells were perfused with Eagle’s MEM and supplements as described above at the rate of 20 ml/h. The different oxygen levels were obtained by changing the gas composition of the perfusate with oxygen-nitrogen mixtures using a gas mixing flowmeter (model GF-3; Cameron, Port Aransas, TX). Equilibration of desired oxygen tensions in the perfusate reservoir was monitored with an oxygen electrode and oxygen meter (model 781; Strathkelvin Instruments, Glasgow, UK). To maintain the equilibrated oxygen tensions all of the way into the measuring ampoule, the flow tubings for perfusate were made of viton and stainless steel, i.e., materials with minimum permeability to oxygen. The perfusate in the measuring ampoule was mixed by motorized slow rotation of the plateholder.
The cells were first perfused with medium equilibrated with 21% oxygen-79% nitrogen. After stabilization of the heat signal, the heat production of the cells was recorded at 1-s intervals for ∼30 min. Thereafter, the perfusate was equilibrated with 10, 5, 2, and 1% of oxygen for ∼1-h periods, and thereafter the oxygen level was stepwisely (2, 5, and 10% of oxygen) increased back to 21%. The heat production of the cells was continuously recorded as above. In some experiments, after measurement of heat production rate at 21% of oxygen, the oxygen concentration of the perfusate was directly decreased to 1%. The heat production was then recorded for up to 4 h, whereafter the oxygen level of the perfusate was returned back to 21%. The heat production level at each oxygen concentration was calculated from the values over the 30-min period preceding the next oxygenation change.
Blank determinations (no cells in the measuring ampoule) were run in parallel with all experiments. At the end of each experiment, cells were lysed in RIPA buffer (10 mM Tris·HCl, 150 mM NaCl, 1% sodium deoxycholate, 1% Triton X-100, 1 mM Na3VO4, and 10 mM NaF, pH 7.4), frozen in liquid nitrogen, and stored at −80°C. Protein amounts were measured from the cell lysates using the BCA protein assay kit (Pierce, Rockford, IL).
Exposure of cells to different oxygen concentrations for determination of gene expression and ATP content.
Before treatments, the similarity of subconfluence levels in cell cultures was checked, and culture medium was changed in serum-free MEM, pH 7.6, with above-described supplements. Cells were exposed to reduced oxygen concentrations in a multigas incubator (Sanyo MCO-175M). Atmospheric gas concentration was made up with nitrogen at 20°C. In all experiments, equilibration of desired oxygen tension in the culture medium was monitored with an oxygen electrode and oxygen meter (model 781; Strathkelvin Instruments).
To study the oxygen dependence of gene expression and ATP content, oxygen concentration of culture medium was reduced stepwisely from 21% (atmospheric air) to 10, 5, 2, and 1% for 2-h periods, and one culture bottle was withdrawn and sampled for each oxygen concentration. To study the time course of changes in gene expression and ATP content during hypoxia, oxygen concentration of the culture medium was decreased from 21% to 1 or 2%. Thereafter, culture bottles were withdrawn and sampled at 1, 2, 4, 6, 8, and 18 or 20 h. Opening of the incubator for sampling did not cause any significant increase in oxygen concentration of the culture medium. For controls, culture bottles were maintained under atmospheric air (21% oxygen) at 20°C and sampled at randomized times.
ATP content and cell viability.
For measurement of ATP content, cells were immediately washed with ice-cold PBS and gently scraped off. An aliquot was taken for measurement of protein content, the cells were centrifuged (500 g for 2 min at 4°C) and deproteinized by adding ice-cold 0.5 M perchloric acid, and the precipitated proteins were pelleted by centrifugation (10,000 g for 2 min at 4°C). The supernatant was neutralized with 1.5 M KOH, and ATP content was determined from the neutralized extract with the luciferase-luciferin method using a bioluminescense assay kit (CLS II; Roche, Mannheim, Germany) according to the manufacturer’s instructions. Protein contents were determined using the Bio-Rad protein assay according to the manufacturer’s instructions (Bio-Rad). ATP contents were related to protein content of each culture bottle. Cell viability was assessed fluorometrically measuring lactate dehydrogenase released in culture medium using the CytoTox-ONE Homogenous Membrane Integrity Assay (Promega, Madison, WI).
Preparation of whole cell protein extracts and RNA isolation.
For protein extracts, cells were washed with ice-cold PBS, scraped off, and pelleted by centrifugation (500 g for 2 min at +4°C). The cell pellets were frozen in liquid nitrogen and stored at −80°C. Whole cell extracts were prepared as previously described (51). Briefly, cell pellets were resuspended in buffer containing 25% glycerol (vol/vol), 420 mmol/l NaCl, 1.5 mmol/l MgCl2, 0.2 mmol/l EDTA, 20 mmol/l HEPES, 0.5 mmol/l phenylmethylsulfonyl fluoride, 0.5 mmol/l dithiothreitol, 1 mmol/l Na3Vo4, 2 μg/ml apoprotein, 2 μg/ml leupeptin, 2 μg/ml antipain, and 2 μg/ml pepstatin and centrifuged at 135,00 g for 30 min at +4°C. The supernatants were collected and stored at −80°C. The protein concentrations of the extracts were determined using the Bio-Rad protein assay according to the manufacturer’s instructions. Total RNA was isolated from the cells using Sigma Tri Reagent according to the manufacturer’s instructions.
Northern blot analysis.
For Northern blot analysis, 20 μg of glyoxylated total RNA was separated on a 1.25% agarose gel at 100 volts for 2 h and transferred to a nylon membrane (Hybond-N; Amersham). Membranes were hybridized at 42°C with [α-32P]dCTP-labeled DNA probes labeled by random priming (Prime-a-gene labeling system; Promega) recognizing mRNAs of rainbow trout glucokinase (38), pyruvate kinase (39), cytochrome-c oxidase subunit 1 (COX1; see Ref. 4), F1F0-ATP synthase subunit 6 (ATP6; see Ref. 4), and β-actin (Genbank accession no. AJ438158). After stringent washing [5× saline-sodium citrate (SSC)-0.1% SDS, 2× SSC-0.1% SDS, 1× SSC-0.1% SDS, and 0.2× SSC-0.1% SDS for 10 min each, all at room temperature], the membranes were visualized by autoradiography (Kodak X-OMAT AR) and analyzed with Chemi-imager (Alpha Innotech). Data were normalized to the 18S rRNA stained from the membrane with methylene blue after transfer.
Whole cell protein (20 μg) was separated on 8% SDS-polyacrylamide gel and transferred to a nitrocellulose membrane (Schleicher & Schuell, Keene, NH). Membranes were blocked for 1.5 h in 3% nonfat dry milk in PBS with 0.3% Tween 20 at room temperature and incubated with primary antibodies overnight at 4°C. Primary antibodies and the dilutions used were the following: polyclonal COOH-terminal rainbow trout HIF-1α (52) 1:200, monoclonal anti-OxPhos complex IV subunit I against human COX1 (Molecular Probes, Eugene, OR) 1:1,000, polyclonal GCK-(H-88) against the COOH terminal of human glucokinase (Santa Cruz Biotechnologies, Santa Cruz, CA) 1:200, and anti-α-actin 1:2,000 (Sigma Chemical). Horseradish peroxidase-conjugated anti-rabbit (HIF-1α and glucokinase) or anti-mouse (COX1 and α-actin) antibodies (Amersham Pharmacia Biotechnology) were used as secondary antibodies. The proteins were detected using enhanced chemiluminescense according to the manufacturer’s instructions (enhanced chemiluminescence; Amersham Pharmacia Biotechnology). The signals were captured on X-ray film, and the images were analyzed with a Chemi-imager (Alpha Innotech). Equal loading was confirmed by staining gels with Coomassie brilliant blue.
Electrophoretic mobility shift assay.
The binding activity of HIF-1 protein on conserved HIF-binding sites in DNA (HRE) was analyzed on whole cell protein extracts by electrophoretic mobility shift assay carried out as previously described (30, 52). Because no fish sequences were available, the sense and antisense strands for the HIF-1-binding sites in the promoter region of human transferrin gene (46) were used for generation of a γ-32P-labeled oligonucleotide probe. Reaction mixtures containing 10 μg of protein, 0.1 μg of poly(dI-dC) and γ-32P-labeled oligonucleotide in 10 mM Tris (pH 7.5), 50 mM NaCl, 50 mM KCl, 1 mM MgCl2, 1 mM EDTA, 5% (vol/vol) glycerol, and 5 mM dithiothreitol were incubated for 30 min at room temperature. The samples were run in a native 4% polyacrylamide gel at 150 volts at room temperature for 1.5 h. The gel was dried under vacuum, and the protein-DNA complexes were visualized by autoradiography (Fuji Super RX).
Data are expressed as means ± SE of the number of independent preparations from different fish. Statistical significance of the data was tested by repeated-measures ANOVA followed by appropriate posttests or by Student’s t-test using SigmaStat software.
Effect of oxygen availability on metabolic rate of hepatocytes.
The metabolic heat production of trout hepatocytes at atmospheric 21% oxygen was 206 ± 31.1 μW/mg protein (n = 13), approximately corresponding to 61.18 ± 9.2 pW/cell (106 cells contain 0.297 mg protein, ∼3 × 105 cells in 2 microplates). When the oxygen level of the perfusate was decreased to 10, 5, 2, and 1% of oxygen, metabolic rate of hepatocytes decreased to 86.1 ± 2.0, 76.3 ± 2.63, 63.4 ± 2.92, and 54.2 ± 4.24% of the initial rate at 21% oxygen, respectively (n = 4, Fig. 1).
Initially, the cells responded 8.1 ± 0.69 min after each decrease in the gas composition of the perfusate in the reservoir. The following steep decrease in heat production of the cells took 10.3 ± 0.51 min, whereafter heat production stabilized at a new level. With the flow rate used, this corresponds to the time required for the perfusate to reach the ampoule (6 min) and to the time required for the perfusate to replace the volume of the medium in the ampoule (12 min). Thus the metabolic rate of the cells decreased immediately upon the decrease in ambient oxygen tension. When oxygen level of the perfusate was returned to 21% in a stepwise manner, metabolic rate of the cells again immediately increased back to the levels initially measured at corresponding oxygen levels (Fig. 1). In experiments where the initial metabolic rate of the cells at 21% oxygen showed upward drifts (2 of 4 experiments, Fig. 1), metabolic rate after reoxygenation to 5, 10, and 20% oxygen was higher than during the stepwise decrease at these oxygen levels (Fig. 1). When oxygen concentration was directly reduced from 21 to 1%, cells responded immediately by decreasing metabolic rate to 68 ± 5.5% of the initial rate (n = 5). Although this response was less pronounced than upon a graded decrease of oxygen to 1% (54.2 ± 4.24%), the difference between these treatments did not reach statistical significance (Student’s t-test). No further changes occurred in metabolic rate of the cells for up to 4 h of incubation at 1% oxygen, and the metabolic response was fully reversible for at least after 4 h at 1% oxygen.
It is important to note that, because oxygen tension was monitored from the perfusate reservoir rather than from the measuring ampoule, we cannot state with absolute certainty that the equilibrated oxygen tensions were perfectly maintained throughout the system. Therefore, some diffusion of oxygen from the environment could have increased the oxygen tension, and thus the oxygen levels that affect metabolic rate could actually be slightly higher than designated.
ATP content and cell viability.
ATP content of trout hepatocytes under atmospheric air (21% of oxygen) was 7.1 ± 0.84 nmol/mg protein−1 (n = 18). This corresponds to ∼1.8 mM concentration of ATP in the cells (fresh weight 1.432 mg/10−6 cells, 79.5% water). When hepatocytes were subjected to a graded reduction in oxygen concentration of the medium, ATP content of the cells declined (Fig. 2A). The lowest ATP content, ∼43% of the level at 21% oxygen, was measured at 5% of oxygen, and at 2% oxygen the level was still significantly lower, ∼55% of that at 21% oxygen (n = 8). At 1% of oxygen, however, ATP content of the hepatocytes appeared to return to initial levels observed at 21% of oxygen (Fig. 2A). Similarly, when oxygen concentration was directly reduced from 21 to 1%, there was no significant change in cellular ATP content after either 6 or 20 h (n = 10, Fig. 2B).
Incubation at 1% oxygen did not affect cell viability. Cell viabilities were 96.1 ± 0.35, 98.3 ± 0.47, and 97.1 ± 0.39%, at 21% oxygen, and after 6 and 20 h at 1% oxygen, respectively (n = 10).
Oxygen dependence of gene expression.
When oxygen supply to cells was gradually reduced from atmospheric 21% oxygen to 10, 5, and 2% for 2-h periods, amounts of all studied mRNAs decreased (Fig. 3). The response was most pronounced in the mRNAs of mitochondrial enzymes of electron transfer chain, subunit 1 of COX1, and ATP6. For these, the decrease was significant already at 5% oxygen (P < 0.05 and P < 0.001 for COX1 and ATP6, respectively, n = 9, Fig. 3). After the 2-h period at 2% of oxygen, mRNA amounts of COX1 and ATP6 were further decreased to <50% of the levels at 21% of oxygen (P < 0.01 and P < 0.001 for COX1 and ATP6, respectively). In addition to mitochondrial enzyme transcripts, the response at 2% oxygen was significant also for the mRNA of pyruvate kinase (P < 0.05). Notably, after the 2-h period at 1% oxygen, all studied mRNAs returned to initial levels observed at 21% oxygen and were significantly elevated compared with the values seen after 2 h at 2% oxygen (P < 0.01, 1% compared with 2% for all mRNAs, Fig. 3).
To find out whether protein amounts of the enzymes were oxygen dependent, the amounts of COX1, glucokinase, and actin protein at different oxygen levels were analyzed. The amounts of all these proteins decreased slightly in response to reduced oxygen supply, although the response was significant only for actin at 2% of oxygen (P < 0.01, n = 4, Fig. 4B). Moreover, an identical pattern of changes in protein level and mRNA expression was seen in some individual experiments (Fig. 4A). HIF-1α protein started to accumulate at 2% of oxygen, and maximal HIF-1α levels were detected at ∼2–1% of oxygen (Fig. 4A). It should be noted that, because HIF-1α translocates into the nucleus upon its activation, HIF-1α protein accumulation and HIF-1 DNA binding are usually assessed in nuclear extracts, rather than whole cell extracts as used here. However, the accumulation of HIF-1α at the same oxygen levels in trout hepatocytes has also been observed using nuclear extracts (A. J. Soitamo and M. J. Nikinmaa, unpublished observations).
Time course of gene expression responses to low oxygen tension.
The time course of gene expression responses to low oxygen was studied at 2 and 1% oxygen levels. At these oxygen concentrations, HIF-1α protein amount and its DNA-binding activity increased to maximal levels in 1–2 h and thereafter declined but remained above the initial level determined at 21% oxygen during the subsequent 16 h of hypoxia (Fig. 5). At 1% oxygen, HIF-1 activation was accompanied by a concomitant induction of glycolytic enzyme expression (glucokinase and pyruvate kinase), whereas this response was absent at 2% oxygen (n = 5 or 6; Fig. 6).
There was no significant change in mRNA and protein expression of electron transfer chain enzyme subunits COX1 and ATP6 or actin upon 20 h exposure to 1 or 2% oxygen (Figs. 6 and 7). Yet a trend toward a decrease in COX1 protein amounts was seen after 2 h at 2% of oxygen (Fig. 7B). The existence of a threshold oxygen level close to 2% oxygen was further supported by the observed variation in the levels of all studied mRNAs upon long-term exposure to 2% oxygen; in some individual experiments, mRNAs clearly increased, whereas a decline in mRNA amounts was seen in other experiments (data not shown).
Oxygen dependence of metabolism.
In the present study, the response of cultured trout hepatocytes to changes in oxygen availability over the oxygen range from 21 to 1% was immediate and fully reversible upon reoxygenation. Previously, similar immediate regulation of metabolic rate over a wide range of oxygen tensions has been shown in monolayer cultures of fetal sheep skeletal muscle cells (7) and in isolated skeletal muscles of dog (20) and frog (57). The phenomenon has been linked to the fact that contracting tissues, especially skeletal muscle, exhibit large changes in metabolic rate during changes in muscle work (20). Thus it has been suggested that oxygen plays a key role in regulation of changes in metabolic rate in muscle tissue (18, 43). Yet the existence of a comparable mechanism in other cell types has remained unclear. Our results indicate that, although changes in metabolic rate in the liver tissue are modest (only ∼1.5- to 2-fold changes compared with up to 100-fold changes in skeletal muscle), a mechanism exists in trout hepatocytes by which oxygen availability contributes to the control of metabolism.
Generally, ATP concentration remains fairly constant in tissues during changes in metabolic rate (18). This energetic steady state is maintained by coordinated balancing of ATP demand and production. Accordingly, oxygen-dependent changes of metabolic rate in muscle occur at stable ATP concentrations, indicating strict coupling of ATP demand and supply (20, 57). In contrast, ATP content in trout hepatocyte cultures declined significantly already when oxygen decreased from atmospheric 21% to 5 and 2%. We did not determine ATP levels or cell viability upon prolonged exposure of cells to these oxygen levels. Yet, the facts that metabolic depression was fully reversible and ATP content of the cells seemed to increase back to original values when the oxygen level was further reduced to 1% indicate that cell viability was not affected. Furthermore, cell viability did not decrease during the 20 h of exposure to 1% oxygen. A similar reversible decrease in ATP has been previously observed in trout hepatocytes under complete anoxia (28) and in mammalian hepatocytes subjected to changes in oxygen availability (48). In the latter study, an oxygen-dependent decrease of ATP content amounted to ∼50% in suspended rat hepatocytes in response to reduction of oxygen supply from 13 to 2%. Moreover, in rat hepatocytes cultured under 12% oxygen atmosphere for 24 h, a decrease of ATP content by 47% was reported, and yet cell viability was still not affected (16). ATP supply and demand in hepatocytes thus appear to be less tightly coupled compared with muscle cells and may be tentatively linked to the absence of a phosophocreatine/creatine system, which buffers ATP levels in the muscle tissue.
An estimation according to measurements of ATP and ADP levels in mammalian tissues (8) would indicate that a 57% reduction in ATP concentration from an initial concentration of 1.8 mM upon change of oxygen supply from 21 to 5% would increase the intracellular free ADP concentration by up to 40-fold. Because ATP depletion is reversible, the ADP formed is apparently retained in the cells, which may dramatically alter the ATP-to-ADP mass action ratio (35). However, different from classical models, which relay on regulation of ATP-producing pathways by changes in adenylate phosphate concentrations during changes in metabolic rate (13, 31), more recent views suggest that ATP-producing and ATP-utilizing pathways would be controlled in parallel by a complex cytosolic signaling network (3, 12, 18, 25). The function of this signaling network would be modulated by regulatory information, such as oxygen supply (18). Rather than being significant in the control of metabolic rate, mass effect and feedback regulation by adenylate phosphates have been proposed to act as a homeostatic fine-tuning mechanism ensuring rapid stabilization of metabolic pathways after perturbation in the cells (3, 12, 18, 25).
Unexpectedly, ATP contents of trout hepatocytes recovered at 1% of oxygen. Because the metabolic rate of the cells decreased further at this oxygen level, the reduction of ATP utilization must have transiently exceeded that of ATP production. In line with this, this oxygen level has been shown to trigger oxygen conformance with coordinated downregulation of ATP production and ATP-utilizing functions in suspended trout hepatocytes and several other cell types from different species (2, 9, 29, 48). Thus, whereas the observed oxygen dependence of metabolism at higher oxygen tensions possibly indicates the role of oxygen in the regulation of metabolic rate changes, the responses evoked at 1–2% oxygen levels likely represent rearrangement of cellular metabolism in adaptation to low oxygen supply. In accordance with that, as in cultured mammalian cells (51, 56), these oxygen levels also trigger HIF-1α stabilization and HIF-1 activation in trout hepatocytes.
The time course of HIF-1α protein accumulation and HIF-1 DNA binding was similar to that observed in mammalian cells (54) and can be explained by HIF-1-induced upregulation of oxygen-sensitive prolyl hydroxylases during long-term hypoxia (1). In mammals, target genes of HIF-1 include glycolytic enzymes. In line with this, in trout hepatocytes, glucokinase and pyruvate kinase were induced concomitantly with HIF-1 activation when oxygen was directly reduced from 21 to 1%. Moreover, enhanced lactate production has been previously reported in trout hepatocytes at 1% oxygen (29). Yet, in contrast to mammalian cells, where an increase of glycolytic enzyme expression is still observed after 16 h of hypoxia (11), induction of these enzymes was only transient, lasting 2 h, in trout hepatocytes. Moreover, we saw no induction of these enzymes after a stepwise decrease of oxygen to 1%. Therefore, it remains unclear if an induction of glycolytic enzymes, possibly enhancing glycolytic ATP production, contributed to recovery of cellular ATP at low oxygen.
Interestingly, HIF-1 has recently been shown to also regulate aerobic energy metabolism in mammalian cells (24, 40). Kim et al. (24) and Papandreou et al. (40) demonstrated that HIF-1 actively represses mitochondrial respiration by inducing pyruvate dehydrogenase kinase 1 (PDK1). PDK1 inhibits pyruvate dehydrogenase enzyme, which in turn limits entry of pyruvate in the tricarboxylic acid cycle and thereby suppresses both the tricarboxylic acid cycle activity and respiration in the cells.
Possible contributions of energy-requiring processes.
In the present study, we did not directly address the oxygen dependence of specific ATP-consuming functions. However, responses of two of the major energy-consuming processes in trout hepatocytes, protein synthesis and Na+-K+-ATPase, to changes in oxygen supply have previously been studied in suspended trout hepatocytes. A reduction of oxygen from atmospheric level decreased protein synthesis at 1–2% of oxygen (29, 37) and inhibited Na+-K+-ATPase activity at 10 and 5% oxygen (5) as well as at 1% (5, 29) in these cells. In the present study, the amounts of some proteins tended to decrease, whereas that of others remained unaltered or increased slightly during long-term exposure to 1 and 2% of oxygen. Because it is plausible that protein synthesis is suppressed at these oxygen tensions also in hepatocytes cultured in monolayers, this suggests that alterations in protein synthesis do not directly translate into protein amounts. On the other hand, the expressions of mitochondrial and glycolytic enzymes, as well as of actin mRNAs, were strikingly dependent on oxygen supply in trout hepatocyte cultures. Notably, the oxygen-dependent changes in all transcript amounts followed cellular ATP levels. Previously, a reversible decrease in both mitochondrial-encoded and nuclear-encoded mRNAs of electron transport chain together with depletion of cellular ATP has been observed in mammalian cell lines exposed to prolonged (6–24 h) hypoxia (1–2% O2; see Refs. 23 and 55). In the study of Vijayasarathy et al. (55), a marked (30–70%) decrease in transcript amounts of COX1 and depletion of cellular ATP was accompanied with a reduction of same magnitude in mitochondrial transcription rate. Particularly because mitochondrial RNA polymerase requires a high concentration of ATP for maximal activity, high concentrations of ATP may have regulatory function on mitochondrial transcription rate (34). Thus, rather than representing specific oxygen-dependent gene responses, the observed decline in transcript amounts is likely caused by an overall decrease in transcription in the cells.
Finally, it is possible that the oxygen-regulated changes in hepatocyte metabolism are connected to proliferation of the cells. We omitted serum from the experimental medium as a potential cause of variation, but trout hepatocytes do not require components of serum for proliferation (41). Low oxygen availability (1–2%) has been shown to both inhibit cell proliferation and to stimulate it, depending on cell type (15, 27, 42). Notably, a complete suppression of DNA synthesis has been reported in rat hepatocytes cultured at 8 or 5% oxygen (16).
It is important to note that, in the present study, the dose responses of metabolism and gene expression to oxygenation changes were obtained primarily by gradually reducing oxygen supply to the cells. Consequently, the responses observed at different oxygen levels were obtained after different times of exposure to reduced oxygen and may thus not be directly comparable. In addition to duration of exposure at a specific oxygen level, the metabolic responses may also depend on the magnitude of oxygenation change (47) or the rate of change in oxygenation (48). However, in the present study, the responses of cells in experiments where oxygen was directly reduced to 1 or 2% were mainly similar to those obtained by gradual reduction of oxygen to the same oxygen levels, suggesting that the oxygen level as such is of primary importance for the cellular response. Moreover, the fact that reoxygenation reversed metabolic rate of the cells to levels corresponding to those during the graded decrease of oxygen suggests that metabolic rate of the cells would be characteristic for each oxygen level. Notably, the existence of a threshold oxygen level for gene expression and ATP content at 1–2% oxygen was evident from both types of experiments.
The reason why oxygen dependence of metabolic rate at high oxygen levels has not been observed in previous studies with trout hepatocytes is probably methodological. Pannevis and Houlihan (37) and Krumschnabel et al. (29) studied oxygen dependence of respiration in trout hepatocytes using closed-chamber respirometry. Studies by Arthur et al. (2) and Casey and Arthur (10) have shown that the degree of oxygen conformance in mouse muscle cells and in neonatal rat cardiac myocytes is markedly reduced when a closed chamber respirometry is used instead of a perfusion system. The perfusion setup more closely resembles the in vivo situation, since constant delivery of substrates is provided and compounds produced by the cells are gradually diluted and swept away. Furthermore, because stirring is required for Clark-type oxygen electrodes, earlier studies on the oxygen dependence of cellular energetics of trout hepatocytes have been carried out with cells in suspension (29, 37). Cells in suspension lack cell-cell interactions, which enable communication between cells and are thus important for growth and function of cells in tissues (32, 33). Trout hepatocytes in monolayer cultures readily aggregate and reestablish junctional complexes and gap junctions between cells. Cells that fail to aggregate show loss of functional properties (50). Because all studies so far demonstrating immediate regulation of metabolic rate by oxygen supply have been carried out either in isolated tissues (in situ or in vitro) or in monolayer cultures of cells using perfusion setup (7, 20, 57), it seems plausible that the presence of cell-cell interactions and flowthrough perfusion are important in determining the responses of cells to changes in oxygen supply. Clearly, further studies are needed to reveal the details of the regulation of cellular metabolism by oxygen.
The present study was carried out to investigate whether monolayer cultures of primary cells (hepatocytes) respond to changes in oxygen availability over the range from 21 to 1% oxygen. The normal oxygen tensions in arterial and venous blood of rainbow trout are ∼15 and 5 kPa (15 and 5% in the gas phase), respectively (53). Thus, for trout hepatocytes, as for most other cell types, the 21% oxygen level is in fact hyperoxic, suggesting that metabolism is upregulated in cells cultured under atmospheric air.
Our results clearly show that oxygen has a direct regulative effect on metabolism of trout hepatocyte cultures. In addition to metabolic rate, both gene expression and energetic state of the cells are strikingly dependent on oxygen supply. These findings support the view that oxygen has a profound role in metabolic regulation in cells independent of constraints placed by mitochondrial energy production.
This study was supported by the Academy of Finland, Biological Interactions graduate school, and the Emil Aaltonen Foundation.
We thank Susanna Airaksinen and Minna Vainio for helpful advice, Nina Vuori for technical assistance, and Gerhard Krumschnabel for critical comments on the manuscript.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 the American Physiological Society