Potential conflict of interest: Nothing to report.
Following prolonged hypoxia, mammalian cells invoke adaptive mechanisms to enhance oxygen delivery and promote energy conservation. We previously reported that hepatocytes subjected to prolonged moderate hypoxia (PO2 = 20-50 mmHg for >3 hours) demonstrated a reversible inhibition of cellular respiration with maintenance of cell viability, associated with a decrease in mitochondrial adenosine triphosphate (ATP) synthesis; acute hypoxia (similar PO2 for <30 minutes) did not induce a similar suppression of respiration and ATP synthesis. In the current study, using an in vitro model of primary rat hepatocytes, we measured the changes in metabolic demand for ATP during hypoxic conformance, and tested whether viability is maintained by preferentially suppressing nonessential processes while sustaining processes essential for maintaining cell homeostasis. In addition, the rate of recovery of oxygen consumption and ATP concentrations following reoxygenation after prolonged and acute hypoxia/anoxia was compared. Oxygen consumption and ATP concentrations decreased during prolonged hypoxia compared with acute hypoxia. However, ouabain-inhibitable respiration did not decrease during prolonged hypoxia, indicating that membrane Na+/K+ ATPase activity, an essential process for cell viability, was maintained. In contrast, ATP-dependent glucuronidation and sulfation of acetaminophen, deemed “non-essential” processes, were decreased significantly compared with normoxic cells. After reoxygenation, cells exposed to prolonged moderate hypoxia demonstrated a more rapid recovery of respiration compared to acute hypoxia/anoxia. Conclusion: This “hepatic hibernation” during prolonged moderate hypoxia may represent an anticipatory adaptation that seeks to maintain cell viability while delaying or preventing the onset of lethal hypoxia, and facilitates rapid recovery after the resumption of normoxia. (HEPATOLOGY 2007.)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Classical studies have shown that cellular respiration remains independent of oxygen availability until a critically low oxygen tension (PO2) of 3 to 5 mm Hg is reached.1–4 Below that point mitochondrial respiration becomes limited by O2 supply, ATP levels decrease, and hypoxic cell injury may ensue because cell processes essential for cell homeostasis cannot be sustained. However, in previous studies we found that hepatocytes exposed to moderate hypoxia (PO2 = 20–50 mm Hg) for periods of longer than 3 hours exhibited a 40% to 60% decrease in respiratory rate compared with cells exposed acutely (<30 minutes) to the same PO2.5, 6 This suppression was maintained for 24 hours without a decrease in cell viability and was rapidly reversed when normoxia (PO2 > 100 mm Hg) was restored. In addition, measurement of mitochondrial function and ATP synthesis as determined by mitochondrial membrane potential suggested a decrease in mitochondrial ATP synthesis only in chronic moderate hypoxia, with no change during acute hypoxia.6 Collectively, these findings suggest that hepatocytes can reversibly suppress ATP synthesis during prolonged moderate hypoxia.
Under steady-state conditions, ATP synthesis must match ATP utilization in the intact cell to preserve cellular ATP stores and maintain cell viability. Therefore, the decrease in mitochondrial ATP synthesis observed during prolonged moderate hypoxia should be associated with a decrease in ATP utilization or demand. Because cell viability was maintained, ATP-dependent processes that were essential for maintaining cell homeostasis must have been sustained. We therefore hypothesized that ATP utilization may have been preferentially inhibited at sites involved with nonessential cell processes. A selective inhibition of ATP utilization by cellular processes that were not essential for maintaining cell homeostasis could conceivably preserve limited ATP supplies for essential cell functions that would maintain cell viability and facilitate recovery to normal baseline activity during subsequent normoxia.
To test this hypothesis, we first sought to determine the effect of decreased oxygen consumption on cellular ATP levels. Second, we examined whether a decrease in mitochondrial ATP production and oxygen consumption during prolonged moderate hypoxia is associated with a shift toward anaerobic glycolytic ATP production, as measured by cellular lactate levels. Third, we determined whether ATP utilization is inhibited equally in reactions deemed essential for cell homeostasis compared with other reactions deemed nonessential. Cellular Na+/K+ ATPase activity was selected as an ATP-dependent function deemed essential for maintaining cell viability. By contrast, the detoxification of acetaminophen was selected as an organ-specific process that is not essential for maintaining hepatocyte viability. This activity was assessed by measuring the rates of sulfation and glucuronidation of acetaminophen in hepatocytes during prolonged hypoxia. These conjugation reactions use ATP, are dependent on the energy state of the cell, and were used to provide an indication of ATP utilization for nonessential cell function. Finally, we examined the ability of hepatocytes to restore oxygen consumption and ATP production to original normoxic values following reoxygenation.
Male Sprague-Dawley rats weighing 225 ± 25 g were provided with food and water ad libitum. Primary cultures of rat hepatocytes were obtained by collagenase digestion of livers, using methodology similar to that described by Seglen.8 Briefly, a midline laparotomy was performed, the portal vein was cannulated, and the inferior vena cava was severed. The liver was perfused in situ with Krebs Henseleit buffer (pH 7.4, 37°C) bubbled vigorously with 95% O2, 5% CO2 gas. After 10 minutes, the perfusate was switched to calcium-free Krebs Henseleit buffer (5 minutes), then calcium-free Krebs Henseleit containing ethylene glycol tetraacetic acid (EGTA) (1 mM, 5 minutes), and finally Krebs-Henseleit buffer containing 0.02% collagenase (8 minutes). The capsule of the liver was then disrupted and the cells were dispersed and filtered through a nylon mesh. The hepatocytes were then washed in buffer and purified by Percoll density gradient centrifugation at 4°C. This yielded a homogenous suspension of hepatocytes with viability that typically exceeded 90%, as determined by trypan blue (0.4%) exclusion. Cells were enumerated using a hemacytometer (Hausser Scientific).
In most studies, approximately 20 to 25 × 106 cells were seeded into spinner flasks containing 250 mL modified Dulbecco medium supplemented with HEPES buffer (25 mM) antibiotics. The media in the spinner flasks had been previously equilibrated with 21% O2, 5% CO2 passed through the head space above the media in the flask. The spinner flasks were stirred at 60 to 70 rpm in an incubator regulated at 37°C.
Oxygen Environment of Cells.
The spinner flasks were modified by the addition of a side port to accommodate a polarographic O2 electrode. When assembled, the electrode tip remained in contact with the stirred media. The electrode was calibrated in situ before adding cells to the flask, by equilibrating the media with a calibrated gas mixture delivered through the head space above the liquid. After the cells were added to the system, the electrode was used to monitor the oxygen tension in the media. Media oxygen tension was consistently lower than that of the head space as a consequence of the oxygen uptake by the cells. The PO2 in the media was controlled by adjusting the O2 concentration in the head space gas; carbon dioxide concentration in the head space was maintained at 36 mm Hg (5% CO2) throughout. Electrode calibration was confirmed periodically by tilting the flask to one side, allowing the electrode to equilibrate with head space gas, whose O2 tension was known.
Oxygen Consumption Measurements.
Respiratory rates were measured in aliquots of cells removed from the spinner flask and studied in a respirometer at 37°C as previously described.5 The respirometer consisted of a water-jacketed anaerobic chamber (2 mL volume) fitted with a polarographic O2 electrode. The electrode was calibrated using humidified gas mixtures with known O2 tensions before use. A magnetic stirrer within the chamber was used to keep the cells in suspension. Cell aliquots were transferred anaerobically to the respirometer to avoid reoxygenation that would have resulted from contact with ambient air, and the chamber was flushed with gas with a PO2 similar to that of the cells. Thus, after anaerobic transfer, the initial PO2 in the respirometer was within a few mm Hg of the media PO2 in the spinner flask. Because the respirometer chamber was functionally airtight, the oxygen tension of the suspension decreased linearly with time as the cells consumed dissolved oxygen. From the decrease in PO2 with time, the respiratory rate of the suspension in the respirometer was calculated from the following relationship:
Cellular O2 uptake rates were then calculated by dividing the above rate of O2 uptake by the measured cell concentration (in 106 cells per milliliter). This value was corrected for cell viability, determined by measuring the fraction of cells that failed to eliminate trypan blue (typically less than 10%).
Lactate concentrations were obtained from cell extracts using an enzymatic assay (Sigma). Approximately 1 × 106 hepatocytes were removed from the flask, centrifuged, and resuspended in 300 μL media with 20 μL perchloric acid (1 M). After neutralization with 5N KOH and centrifugation, a 30-μL sample of the resulting supernatant was added to 150 μL H2O plus 80 μL of a mixture containing glycine buffer, lactate dehydrogenase, and NAD+. After incubation for 15 minutes at 37°C, the absorbance (340 nm) was measured in a 96-well microplate reader (Molecular Devices). Absolute concentrations were calculated based on a calibration standard curve.
Adenine Nucleotides Analysis.
Adenine nucleotide concentrations were determined by high-performance liquid chromatography (HPLC).9 Cells were separated from media by centrifugation at 300 rpm for 5 minutes. The cell pellet was disrupted with perchloric acid (1 M) and the mixture was centrifuged at 14,000 rpm for 30 seconds. The supernatant was neutralized with 5 N KOH, and after repeat centrifugation, the resulting supernatant was assayed using a Zorbax RX C8 column. The column was sequentially equilibrated at a flow rate of 1.0 mL/min in methanol, H2O, buffer B [50 mM potassium phosphate, 8 mM tetra butyl hydrogen sulfate, 40% (v/v) acetonitrile, pH 5.8] and buffer A (50 mM potassium phosphate, 8 mM tetra butyl hydrogen sulfate, pH 5.8). An isocratic gradient of 15 minutes from 10% buffer B to 45% buffer B was employed to elute nucleotides in a total analysis time of 30 minutes. Peaks were integrated electronically (Hewlett Packard). A known quantity of xanthosine diphosphate was added to each sample as an external standard. Absolute concentrations were determined by comparing the ratio between adenine nucleotides and xanthosine diphosphate relative to a standard curve. Results were expressed in nanomoles adenine nucleotides per 106 cells. The obtained values of ATP, ADP, and adenosine monophosphate (AMP) were used to calculate a bioenergetic parameter predictive of cell viability, termed the adenylate charge of the cells, which was determined by the following relationship:
The adenylate charge of viable cells varies between 0.7 and 0.9, and falls below 0.6 during cellular injury. The adenylate charge of cells exposed to prolonged moderate hypoxia and acute hypoxia were determined to assess cell viability.
Analysis of Acetaminophen Conjugation.
Hepatocytes detoxify acetaminophen by two enzymatic conjugation pathways that are known to depend on the bioenergetic state of the cell.10 The first pathway is a multi-step reaction catalyzed by sulfotransferases, is ATP-dependent, and yields acetaminophen sulfate:
where PAPS represents phosphoadenosyl phosphosulfate and Ppi represents inorganic phosphate. The second pathway involves glucuronidation, is catalyzed by uridine diphosphate–glucuronyltransferases, is uridine triphosphate dependent, and yields acetaminophen-glucuronide:
The rates of these reactions have been shown to decrease under severe hypoxia where the [ATP] is reduced or the ATP/ADP ratio is decreased. For example, Aw and Jones found a 50% decrease in the rate of the sulfation reaction at a PO2 = 2 mm Hg, and for the glucuronidation at 2.5 mm Hg in rat hepatocytes.10 We measured the rates of these reactions in rat hepatocytes incubated under different oxygen concentrations to assess the effects of prolonged moderate hypoxia on the regulation of ATP utilization. Hepatocytes were seeded into spinner flasks at a density of 105 cells/mL. After allowing 30 minutes for equilibration, the PO2 in the hypoxic flask was reduced to approximately 20 mm Hg whereas the control flask was maintained at approximately 100 mm Hg. These conditions were maintained at 37°C for 3 hours to allow the hypoxic cells to adapt to hypoxia and reduce their rate of respiration. At t = 3 hours, acetaminophen (5 mM) was added to both flasks and allowed to incubate for 3 hours. At t = 6 hours, 2 × 106 cells were removed from each flask for analysis of acetaminophen conjugates. The cells were centrifuged (500 rpm for 5 minutes), resuspended in 300 μL media, and disrupted by the addition of 20 μL perchloric acid (7 M). The mixture was centrifuged (14,000 rpm for 30 seconds) and the supernatant was neutralized with 5N KOH and re-centrifuged. The resulting supernatant samples were analyzed by high-pressure liquid chromatography (HPLC) using the methodology outlined by Moldeus,11 using a Zorbax C14 column. The buffers used included buffer A—1% aqueous acetic acid—and buffer B—1% aqueous acetic acid, methanol, ethyl acetate (ratio 90:15:0.1). Using a flow rate of 2 mL/min, an isocratic gradient of 20 minutes from 25% buffer B to 80% buffer B was used to elute the acetaminophen conjugates in a total analysis time of 40 minutes. The HPLC peaks representing the sulfate and glucuronide conjugates were identified by running known concentrations of commercially available sulfate and glucuronide conjugates.
Reoxygenation After Hypoxia.
The effects of acute reoxygenation on the time course of recovery of oxygen consumption and adenine nucleotide concentrations were measured and compared in prolonged moderate hypoxia (PO2 of 20 mm Hg for 3 hours), and in acute hypoxia followed by 30 minutes of anoxia. After their respective exposures to hypoxic and anoxic environments, cells were transferred anaerobically from the flask into a syringe containing 5% CO2 and 21% O2. The syringe was rotated axially for 3 to 4 minutes to restore the PO2 to approximately 148 mm Hg, followed by measurement of respiratory rates and adenine nucleotides as previously described over time.
Data were analyzed using analysis of variance (ANOVA) for repeated measures. When the ANOVA indicated that a significant difference was present, individual differences were explored with the Student t test using the Bonferroni correction for multiple comparisons. Statistically significance was determined at the 0.05 level. All values are expressed as means ± standard deviation.
Oxygen Consumption Is Suppressed in the Presence of Prolonged Moderate Hypoxia, But Not in Acute Hypoxia.
Oxygen consumption as a function of decreasing PO2 was studied in two experimental groups. In one group (acute hypoxia), media PO2 was reduced from 100 to 20 mm Hg over 30 minutes by lowering the O2 concentration in the head space of the flask, and aliquots were sampled anaerobically for measurement of oxygen uptake (VO2) as the media PO2 reached 100, 50, and 20 mm Hg. In a second group (prolonged hypoxia), two subgroups were examined; (a) media PO2 was reduced from 100 to 50 mm Hg over a period of 30 minutes and then maintained at 50 mm Hg for 3 hours, and (b) media PO2 was reduced from 100 to 20 mm Hg over a period of 30 minutes and then maintained at 20 mm Hg for 3 hours. Aliquots were sampled anaerobically from the flasks for the measurement of VO2 after the media PO2 had remained at 50 and 20 mm Hg, respectively, for 3 hours. Figure 1 shows the oxygen uptake rates as a function of PO2 for the two groups. At 100 mm Hg, oxygen uptakes were not different in the acute (2.50 ± 0.02) and prolonged hypoxia groups (2.52 ± 0.03 μmol/h per 106 cells, P = NS). Oxygen consumption rates in the acute hypoxia group were not different at 100, 50, or 20 mm Hg (P > 0.07). However, respiratory rates in the prolonged hypoxia group were significantly less at 50 mm Hg (P = 0.001) and 20 mm Hg (P = 0.002) compared with the 100 mm Hg value. Respiration became unmeasurable as the O2 tension was subsequently decreased to 0 mm Hg in both groups, and is shown as zero in the figures.
Adenine Nucleotides Are Significantly Decreased in Prolonged Moderate Hypoxia, But Not in Acute Hypoxia.
Adenine nucleotide concentrations for the acute and prolonged hypoxia groups were measured as a function of PO2. As with the oxygen consumption measurements, aliquots for these measurements were obtained anaerobically. Figure 2 illustrates the changes in ATP concentrations in response to acute and prolonged moderate hypoxia. ADP and AMP data are not depicted in Figure 2 but demonstrated profiles similar to ATP. The ATP, ADP, and AMP concentrations were not different in the acute and prolonged hypoxia groups at 100 mm Hg (P = NS). Compared with the value at 100 mm Hg, ATP concentrations were not different at 50 mm Hg or 20 mm Hg in the acute hypoxia group. However, in the prolonged hypoxia group a significant decrease in ATP was found at 50 mm Hg (P = 0.002) and 20 mm Hg (P < 0.001) compared with the 100 mm Hg value. After 30 minutes of anoxia, ATP levels were significantly decreased in both groups (P < 0.01). In acute hypoxia, total adenine nucleotide concentrations were not decreased until anoxic conditions were reached. By contrast, in prolonged hypoxia the adenine nucleotide pool was significantly decreased at O2 tensions of 50 and 20 mm Hg (P < 0.01). Thus, adenine nucleotide concentrations were significantly less in the sustained hypoxia group compared with the acute group at the same PO2. Interestingly, the adenylate charge as determined from the values of AMP, ADP, and ATP was preserved above 0.7 in both groups during the duration of hypoxia and anoxia; this finding suggests that after finite anoxia, both groups could demonstrate cell viability and subsequent reversibility in oxygen consumption and ATP production after reoxygenation to normoxia.
Lactate Measurements in Normoxia and Prolonged Moderate Hypoxia Do Not Demonstrate Significant Differences.
The decrease in oxygen consumption during prolonged moderate hypoxia suggests that mitochondrial ATP synthesis is decreased. To determine whether there is a compensatory increase in anaerobic production of ATP, cellular lactate levels were measured in cells subjected to normoxia and prolonged moderate hypoxia defined as a PO2 of 20 mm Hg for 3 hours. An elevation in lactate levels would be expected if glycolytic flux had increased to maintain ATP production in the face of decreased mitochondrial respiration. As illustrated in Fig. 3, no significant differences in lactate production were detected between normoxic and hypoxic cells. At the end of the duration of prolonged moderate hypoxia, both groups of cells were exposed to PO2 = 0 mm Hg for 1 hour to confirm that the lactate assay was capable of detecting increases in lactate production; this resulted in significant increases in lactate levels in both normoxic and hypoxic cell groups. The absence of hyperlactatemia in the presence of prolonged moderate hypoxia implies that glycolytic ATP production is not triggered after decreased mitochondrial ATP synthesis, which in turn, suggests that cellular ATP demand is down-regulated.
Prolonged Moderate Hypoxia Induces Selective Suppression of Cellular ATPases.
The decrease in oxygen consumption without an increase in glycolytic flux suggested that cellular ATPase activity had decreased. A suppression of ATP utilization for non-essential activities could be protective during hypoxia, by conserving ATP supplies for obligatory cell processes that are required for survival. To determine whether the activity of the cellular ATPases was uniformly or differentially inhibited during hypoxia, we assessed ATP utilization by two independent systems. Membrane Na+/K+ ATPase activity was assessed to provide a measure of an obligatory process required for cell homeostasis. Acetaminophen conjugation rates were measured during prolonged hypoxia to provide a measure of nonessential metabolic activity related to the detoxification function of the liver.
Na+/K+ ATPase Activity.
Na+/K+ ATPase activity was assessed by examining cellular oxygen consumption rates in response to ouabain (100 μM), which inhibits its activity. Exposure to hypoxia for 3 hours produced a 35% suppression of oxygen consumption at 20 mm Hg (1.80 ± 0.13 μmol/h per million cells) compared with 100 mm Hg (2.74 ± 0.29 μmol/h per million cells). However, the ouabain-sensitive component of the oxygen consumption did not decrease during hypoxia (Fig. 4). This indicates that the ATP utilization by the Na+/K+ ATPase was not inhibited during hypoxia-induced metabolic suppression.
Acetaminophen Conjugation ATPase Activity.
Acetaminophen detoxification by hepatocytes involves ATP-dependent formation of either a sulfate or a glucuronidate conjugate,10 and these reactions were deemed non-essential or facultative in contrast to Na+/K+ ATPase activity. Acetaminophen conjugation rates decreased significantly in cells subjected to prolonged hypoxia. Synthesis of the glucuronide conjugate at 20 mm Hg was 63.9% of the rate at 100 mm Hg (Fig. 5A), whereas synthesis of the sulfate conjugate at 20 mm Hg was 20.7% of the rate at 100 mm Hg (Fig. 5B). These results suggest that hypoxia induces a selective inhibition of ATPases. Essential functions such as Na+/K+ ATPase, which is required for volume homeostasis, are maintained during prolonged hypoxia. By contrast, functions not required for cell survival such as the detoxification of acetaminophen, are suppressed and thereby contribute to the metabolic suppression.
Reoxygenation Restores Oxygen Consumption and Adenine Nucleotide Concentrations.
To determine whether prolonged moderate hypoxia influences the rate of recovery of oxygen consumption and adenine nucleotide concentrations, the time course of recovery of these parameters to normoxic values after reoxygenation was compared after prolonged moderate hypoxia, and acute hypoxia/anoxia.
Sustained exposure to 20 mm Hg was associated with a decrease in respiratory rate (2.56 ± 0.05 to 1.16 ± 0.27 μmol/h per 106 cells, P < 0.001) and ATP concentration (27.3 ± 2.5 to 14.2 ± 2.7 nmol/106 cells, P < 0.01). Reoxygenation resulted in a rapid increase in respiratory rate and ATP concentration, with a return to normoxic levels within 5 minutes. Thus, the decrease in ATP and respiratory rate induced during prolonged moderate hypoxia were found to be quickly reversible.
Media PO2 was reduced from 100 to 0 mm Hg over less than 10 minutes, maintained at 0 mm Hg for 30 minutes, and then rapidly returned to 100 mmHg. Cell aliquots were sampled at 100 mmHg, after 30 minutes of anoxia, and 20, 40, and 120 minutes after return to 100 mm Hg. Compared with the baseline values at 100 mmHg, respiratory rates remained significantly depressed after 20 (P = 0.001) and 40 minutes of reoxygenation (P < 0.002). After 120 minutes of recovery, oxygen uptake reached a value that was not different from the baseline value at 100 mmHg (P = NS). In parallel with the respiratory rate, ATP concentrations were significantly depressed during anoxia (P < 0.01) and remained depressed after 20 (P = 0.03) and 40 minutes (P < 0.03) of normoxic recovery. By 120 minutes of normoxic recovery, the ATP concentrations had returned to baseline levels. ADP and AMP values demonstrated trends similar to ATP.
Figures 6A,B and 7A,B illustrate the difference in time course of reversibility of oxygen uptake and ATP concentrations, when prolonged moderate hypoxia and acute hypoxia/anoxia are compared. While the hypoxic hibernation induced by prolonged exposure to moderate hypoxia appears to facilitate rapid reversibility, cells exposed to acute hypoxia and anoxia (30 minutes) required 2 hours of recovery before their respiratory rates and adenine nucleotide concentrations were restored to their original values.
Previous studies have demonstrated that cellular respiration remains independent of O2 supply until the oxygen concentration falls below a critical value in the range of 3 to 5 mm Hg.1–4 Likewise, cellular ATP levels and adenylate charge also appear to be maintained until the same critical O2 tension is reached. Below this critical threshold, decreases in respiration and adenine nucleotide concentrations occur because O2 availability at the mitochondria limits the rate of ATP synthesis relative to its rate of utilization. Our data confirm these observations in that cells subjected acutely to oxygen tensions of 50 and 20 mm Hg exhibited respiratory rates and adenine nucleotide levels that were not different from the values measured at 100 mm Hg.
However, our data also show that hepatocytes maintained at oxygen tensions of 20 and 50 mm Hg for several hours exhibit a reversible suppression of oxygen consumption, consistent with our previous observations.5, 6 Our previous observations based on the measurement of mitochondrial membrane potentials have suggested that this suppression of oxygen consumption is associated with a decrease in mitochondrial ATP synthesis or supply.6 This study indicates that the suppression of oxygen consumption is also associated with a decrease in cellular ATP demand or utilization, which would be anticipated in the setting of decreased ATP supply to preserve cellular ATP levels and cell viability. The decrease in ATP demand was facilitated by a selective suppression of “non-essential” ATPases, and the maintenance of essential cell functions (Na-K ATPases) required for cell viability.
Furthermore, rapid reoxygenation of these cells was associated with an immediate restoration of respiratory rate and adenine nucleotide concentrations to baseline normoxic levels (within 5 minutes). In contrast, as reported previously,7 reoxygenation of cells exposed acutely to hypoxia followed by 30 minutes of anoxia exhibited a slower recovery of oxygen consumption and ATP, with a return to baseline normoxic levels after 2 hours. With respect to these differences in recovery following reoxygenation after short-term anoxia and prolonged hypoxia, we propose that different adaptive mechanisms may be involved. As Aw et Al.7 have outlined in an earlier study, one of the mechanisms involved in mitochondrial suppression after short-term anoxia is the decrease in mitochondrial calcium concentration. The delay in recovery of mitochondrial membrane potential and oxygen consumption after reoxygenation that was observed in their study and our current study may be determined in part by the rate of recovery of mitochondrial calcium concentration. In contrast, the adaptive phenomenon of hepatic hibernation during prolonged moderate hypoxia may involve a different mechanism that facilitates a rapid recovery of mitochondrial membrane potential and oxygen consumption after reoxygenation. These observations suggest the presence of a cellular oxygen sensor that is able to detect the presence of prolonged moderate hypoxia and initiate adaptive mechanisms before the development of severe hypoxia and anoxia, which may be associated with a different set of adaptive mechanisms.
The physiological relevance of extrapolating these findings in isolated hepatocytes to the organ level is complicated by the intricate anatomy and physiology of the liver, in which there is an anatomic as well as a metabolic zonation of the hepatocytes from the periportal (zone 1) to the perivenous regions (zone 3).12, 13 The PO2 of periportal blood entering via the portal vein is approximately 60 to 65 mm Hg, and falls to 30 to 35 mm Hg in the perivenous blood, thereby predisposing the perivenous hepatocytes to a higher risk of hypoxic injury. The close proximity of the sinusoid to the hepatocyte, with a minimal separation by the space of Disse, facilitates efficient transfer of oxygen across the cell membrane, with a reported sinusoidal-to-cell PO2 gradient of approximately 5 mm Hg.14 However, other studies have shown that a substantial intracellular PO2 gradient exists between the cellular membrane and the mitochondrial inner membrane in hepatocytes,15, 16 creating a significant sinusoidal to mitochondrial PO2 gradient. It has been suggested that a sinusoidal to mitochondrial PO2 gradient of approximately 15 mm Hg must be present for optimal oxygen diffusion to facilitate mitochondrial respiration,12 thereby estimating an effective intracellular PO2 of 45 to 50 mm Hg in periportal hepatocytes and 15 to 20 mm Hg in perivenous hepatocytes. Therefore, prolonged extracellular PO2 values in the range of 20 to 50 mm Hg could represent chronic hypoxic oxygen tensions that could trigger an adaptive response in the hepatocyte via a cellular oxygen sensor which is sensitive to prolonged moderate hypoxia, especially in the perivenous zone. Interestingly, with respect to the phenomenon of hepatic zonation of metabolism, the specific function of acetaminophen conjugation is localized to the perivenous zone of hepatocytes, which are at the highest risk for hypoxia, and therefore may especially need to downregulate nonessential functions in the presence of hypoxia.
These observations suggest that hepatocytes, and possibly the liver, have the ability to suppress metabolic activity after exposure to prolonged moderate hypoxia, at PO2 values of 20 to 50 mm Hg, which are much higher than the traditional value of 3 to 5 mm Hg associated with limitation of mitochondrial electron transport and ATP synthesis.15, 17 Similar adaptive responses in the presence of moderate hypoxia have been recognized in other cell and organ systems during regional hypoxemia, including cardiac hibernation18, 19 and the downregulation of ion channels in alveolar epithelial cells.20 For example, in cardiac hibernation, the presence of chronic moderate hypoxia due to decreased coronary blood flow induces an adaptive response in cardiac myocytes that enables them to decrease their ATP supply and demand, and the decreased cellular ATP levels are selectively used to maintain cell integrity and viability at the expense of decreased contractile function. Our observations after reoxygenation mimic the restoration of contractility in hibernating myocardium, with a rapid recovery to normal metabolic activity after re-establishment of normoxia. With regard to the adaptive response in alveolar epithelial cells, one of the adaptive mechanisms involves the downregulation of energy dependent Na channels, including the Na+/K+ ATPase, which is in contrast to our current observations in hepatocytes. This downregulation in the Na+/K+ ATPase can be explained by the different roles of the protein in the two cell types. In the alveolar epithelial cell, the Na+/K+ ATPase is not predominantly responsible for cell homeostasis and viability (as in the hepatocyte), but is one of a group of proteins involved in the regulation of intra and extracellular sodium and water regulation that collectively undergo a downregulation of function in response to hypoxia.
These adaptive responses are anticipatory in nature, and are aimed at maintaining cell viability and survival in an environment characterized by persistent moderate hypoxia, and at risk of developing more severe hypoxia. The ability to detect moderate hypoxia and initiate adaptive responses is important, because responses activated only after the onset of severe hypoxia cannot protect against the development of that state. As we extrapolate these findings in isolated primary hepatocytes to a tissue and an organ level, this response may constitute a model of “hepatic hibernation,” in which the liver has the ability to down-regulate its metabolic activity in the presence of sustained moderate hypoxia, as in ischemic hepatitis. This process of hepatic hibernation may enable the liver to sustain the period of hypoxic insult better, by decreasing its oxygen consumption and ATP demand, and using these substrates selectively for essential cell functions. Furthermore, after recovery from the hypoxic insult, the process of hibernation may facilitate a rapid recovery of organ function to levels observed in normoxia.
In conclusion, our findings suggest that primary rat hepatocytes demonstrate an adaptive down-regulation of metabolism in the presence of prolonged moderate hypoxia, utilizing the limited supply of oxygen and ATP for essential cell functions required for cell survival. This process of hepatic hibernation in the presence of moderate hypoxia is an anticipatory process that seeks to prevent or delay the onset of more severe and lethal hypoxia. In addition, our findings also suggest that hepatic hibernation facilitates a very rapid recovery to normal cellular function following reoxygenation to normoxia. The ability of the hepatocyte to conform its metabolism in the presence of moderate hypoxia implies the presence of a cellular oxygen sensor that is able to detect changes in cellular PO2. The identity of this oxygen sensor remains unknown and will require further study.