Journal of Applied Physiology

O2 availability modulates transmembrane Ca2+ flux via second-messenger pathways in anoxia-tolerant hepatocytes

S. C. Land, R. H. Sanger, P. J. S. Smith


Land, S. C., R. H. Sanger, and P. J. S. Smith.O2 availability modulates transmembrane Ca2+ flux via second-messenger pathways in anoxia-tolerant hepatocytes.J. Appl. Physiol. 82(3): 776–783, 1997.—Transmembrane Ca2+-flux was studied from single isolated turtle hepatocytes by using a noninvasive Ca2+-selective self-referencing microelectrode. Cells in Ca2+-reduced culture medium demonstrated a vanadate-and lanthanum-inhibitable Ca2+-efflux of 4 × 10−17 mol Ca2+ ⋅ μm−2 ⋅ s−1continuously over 170 h. This flux diminished with 50 nM phorbol 12-myristate 13-acetate, a protein kinase C (PKC) activator, and was reinstated on PKC deactivation with sphingosine. Progressive hypoxia resulted in a reversible suppression of Ca2+ efflux to 90% of normoxic controls with an apparent Michaelis constant for oxygen of 145 μM. PKC activation was critical in this suppression, as anaerobic administration of sphingosine caused a Ca2+ influx and cell rupture. Hypoxia was also associated with an altered pattern of adenosine-mediated control over Ca2+ efflux. Adenosine (100 μM) elevated Ca2+ efflux twofold in normoxia, but neither adenosine nor the A1-purinoreceptor antagonist 8-phenyltheophylline altered the observed anaerobic suppression. Aerobic administration of 2–10 mM KCN failed to reproduce the anaerobic suppression; however, in conjunction with 10 mM iodoacetate, complete metabolic blockade caused a Ca2+ influx and cell rupture. These observations suggest modulatory control by oxygen over transmembrane Ca2+ efflux involving second-messenger systems in the hypoxic transition.

  • oxygen sensing
  • hypometabolism
  • adenosine
  • protein kinase C
  • calcium homeostasis
  • calcium-selective self-referencing probe

cell death during severe hypoxia begins with the failure of anaerobic ATP production to meet the energetic demands of membrane ion pumps. As the membrane begins to depolarize, free cytosolic intracellular Ca2+ concentrations ([Ca2+]i) rise, leading to a generalized, poorly controlled activation of Ca2+-dependent phospholipases and proteases. This enhances the rate of depolarization, leading to disruption of cell function, osmotic swelling, and, ultimately, cell death (11, 19). Clearly, the success with which tissues survive hypoxia rests greatly on the ability to maintain the cell membrane potential and, in particular, control Ca2+fluxes between cellular compartments and the external environment.

The cellular basis for anoxia tolerance has been studied by using hepatocytes isolated from a well-characterized vertebrate anaerobe, the Western painted turtle (Chrysemys picta bellii; Refs. 3-5, 14-18). In response to hypoxia, these cells actively suppress ATP synthesis synchronously with demand from major ATP sinks (ion pumping, protein turnover, urea synthesis, glucose release), resulting in an overall tenfold suppression of metabolic rate in anoxia. This large-scale metabolic reorganization is tightly controlled; it is rapid in onset, occurs without perturbations in adenylate concentrations or the cellular membrane potential (despite a reduction in ATP-demand by membrane ion pumps); it involves oxygen-dependent expression and suppression of specific genes; and it sustains a new, lower rate of flux through specific biochemical pathways until reoxygenation (see Ref. 17 and references therein). The effect dramatically spares on-board fermentable substrate, limits rates of metabolic waste accumulation, and results in a profound extension of survival time in anoxia.

This kind of metabolic dormancy clearly occurs in response to oxygen limitation, but the question remains how, and at what level of organization, does the lack of oxygen act as a unifying signal? This direction of inquiry has recently indicated the potential for oxygen to act as a modulator of at least one of the major turtle hepatocyte ATP sinks (protein expression) through an oxygen-receptor mechanism (17). In particular, this study emphasized the capacity for the oxygen signal to override, or modulate, existing modes of control for a complex cellular process during a hypoxic challenge. Given that tight control over the membrane potential and [Ca2+]iis critical in the early stages of anoxia, we reasoned that survival of anoxia would require sensitive detection of oxygen availability, oxygen-dependent readjustment of Ca2+ fluxes, and an intracellular control mechanism capable of coordinating the oxygen signal among channel, pump, and ionic-exchange components of Ca2+ homeostasis.

The present study examines the relationship among oxygen availability, protein kinase C (PKC) activity, and adenosine purinoceptor activation in the control of turtle hepatocyte transmembrane Ca2+ flux (the net balance between plasma membrane Ca2+-adenosinetriphosphatase (ATPase) activity, Na+/Ca2+exchanger activity, and Ca2+channel influx) measured with a noninvasive, Ca2+-selective, self-referencing electrode (24). Our results show that during a transition toward hypoxia Ca2+ flux is suppressed, possessing a high apparent Michaelis constant (K m) for oxygen and a dependency for PKC activation. Furthermore, the oxygen signal demonstrated a capacity to modulate existing purinoceptor-mediated pathways of Ca2+ flux.



Ca2+ ionophore (cocktail A) was purchased from FLUKA Chemical (Ronkonkoma, NY). Minimum essential medium amino acids were from GIBCO/BRL (Grand Island, NY). 12-O-tetradecanoylphorbol 13-acetate (TPA; phorbol 12-myristate 13-acetate) was from Calbiochem (La Jolla, CA). Sphingosine, iodoacetate, KCN, and all other bench and cell-culture chemicals were purchased from Sigma Chemical (St. Louis, MO).

Hepatocyte Isolation and Culture

Adult female Western painted turtles were purchased from Lemberger (Oshkosh, WI) and maintained at 25°C in freshwater aquaria at the Marine Resources Center, Marine Biological Laboratory. Animals were fed commercial turtle food ad libitum. Hepatocytes were isolated as described previously (5), and stock primary cultures were maintained at 25°C in turtle hepatocyte primary culture medium containing minimum essential medium amino acids (GIBCO) in 50-ml Falcon culture vessels (18). An aliquot of cells was washed several times in Ca2+-depleted turtle saline (MgCl supplemented to 10 mM) 24 h before an experiment, then plated at low density (<1,000 cells/dish) onto a circular 22-mm laminin-coated glass coverslip that formed the base of a center well drilled into a 35 × 10-mm Falcon sterile culture dish. To ensure Ca2+ fluxes were in a steady-state condition before an experiment, cultures were allowed to adjust to a Ca2+-depleted/Mg2+-supplemented medium (remaining Ca2+concentration varied between 6 and 30 μM) overnight at 25°C.

Measurement of Extracellular Ca2+ Flux

Extracellular Ca2+flux was measured by using a self-referencing Ca2+-selective microelectrode and employed the same system described by Smith et al. (24) and Kuhtreiber and Jaffe (13). Briefly, silanized “bee-stinger”-tipped microelectrodes (∼3-μm tip diameter) were backfilled with 100 mM CaCl2, then front filled with a 30-μm column length of Ca2+-selective ionophore in a liquid ion-exchanger cocktail (FLUKA,cocktail A). The electrode was then connected to a motion-controlled head-stage via a Ag/AgCl wire, and the circuit was completed by placing a salt bridge (3M KCl in agar connected to a Ag/AgCl wire) submerged in culture medium at the edge of the dish. The entire assembly was mounted on the stage of a Zeiss Axiovert inverted microscope housed in a Faraday cage. Electrode movements relative to the cell during data collection were viewed with the aid of a video monitor. Mode of data collection and movement control are as described previously (24). Nernstian calibration of each electrode was conducted by recording the millivolt signal obtained in 0.1, 1, and 10 mM solutions of CaCl2; typical calibration yielded a slope of 28 mV per decade Ca2+ concentration. The selectivity of the Ca2+ ionophore used in the tip of the electrode precludes the detection of other physiologically common ion species.

The self-referencing Ca2+-selective electrode was moved to within 1 μm of the cell surface, and care was taken to avoid contact with the plasma membrane. Amplitude of oscillation was 10 μm perpendicular to the surface of the cell with a frequency of 0.3 Hz. Ca2+-specific voltage differences obtained over the distance of electrode movement (ΔμV) were converted to flux by substitution into the Fick equation, as described previously (24). As the diffusion constant for Ca2+, we used 8 × 10−6cm2 ⋅ s−1, and all results treated this way were corrected for background Ca2+ concentration. Note that the combined time required for data processing and signal averaging through DVIS10 (5 s/data point) is such that the Ca2+-selective self-referencing probe will not resolve rapid (i.e., in ms) events. Throughout each experiment, cells were deemed viable and healthy if control Ca2+ flux reflected a steady-state condition, and experiments were terminated if cells demonstrated visible signs of bleb formation (usually the result of the electrode having come in contact with the cell membrane). Ca2+ flux measured from cells maintained in Ca2+-reduced culture medium for 1 wk remained steady at −4 × 10−17mol ⋅ μm−2 ⋅ s−1. A time period of 24 h postisolation in Ca2+-depleted medium was selected as an adequate period of adjustment before each experiment.

Experimental Design

The experiments were divided into four parts.

Part A: characterization of steady-state Ca2+ flux from turtle hepatocytes.

To demonstrate the presence of an extracellular Ca2+ gradient, the electrode was placed as close to the cell as possible without touching the plasma membrane and oscillated over a 10-μm excursion relative to the equatorial centerline. The electrode was then moved outward, maintaining the same excursion amplitude, along an axis perpendicular to the cell membrane, and the change in flux was subsequently recorded. In a second experiment, the electrode was maintained at a position close to the cell, and the excursion amplitude was increased over a range from 1 to 10 μm. A third experiment investigated the effect of changing extracellular Ca2+concentrations ([Ca2+]e) on the Ca2+ gradient and Ca2+ flux over a 10-μm excursion amplitude. Ion-substitution experiments examined the response of self-referencing electrode-Ca2+ efflux toward vanadate (10) and La3+ (7), both implicated in the inhibition of Ca2+ ATPase activity at micromolar concentrations. A final series of experiments involved the substitution of Li+ for Na+ in the media to determine the presence and contribution of a Na+/Ca2+exchanger to transmembrane Ca2+flux. Li+ will enter the cell through Na+ channels but substitutes less efficiently for Na+ in the exchanger.

Part B: oxygen dependence of Ca2+ flux.

Cells were plated in Ca2+-reduced, Mg2+-supplemented turtle saline as described above and placed on the microscope stage under a gas-perfused chamber with access ports for microelectrode entry. Silicon grease was used to firmly seal the base of the culture dish and the gassed chamber to the microscope stage. The chamber was then gently perfused with 95% air-5% CO2 to establish a normoxic environment. Oxygen depletion was achieved by perfusing 95% N2-5%CO2over the surface of the medium, followed by 95% Ar-5% CO2 to achieve extreme hypoxia. Care was taken at all times to avoid gas-induced vortexing of the culture medium, and this was greatly facilitated by ensuring that the point of gas entry into the chamber was below and away from the inside edge of the culture dish. Oxygen concentrations were monitored during each experiment by using a Diamond General Development (Ann Arbor, MI) model 723 polarographic oxygen microelectrode connected to a chart recorder. In all experiments, the oxygen microelectrode was within 100 μm of the point of Ca2+ measurement and approximately on the same plane. Oxygen microelectrodes were calibrated before each experiment by placing the tip of the electrode to a depth of 3 cm into either air- or N2-Ar-equilibrated, Ca2+-reduced turtle saline. Conversion of oxygen tension to oxygen concentration was as described previously (6, 15).

Part C: oxygen availability and PKC activation and inhibition.

To determine whether the extracellular Ca2+ flux was under second-messenger control, cells were exposed to sequential additions of 50, 100 (n = 5), or 150 nM (n = 2) of TPA, an activator of PKC. The reversibility of the TPA effect was investigated by addition of 50 μM of the PKC inhibitor sphingosine under aerobic and anaerobic conditions. In each case, stock solutions were made up in a solvent of 80% ethanol, and, in the case of sphingosine, the stock was sonicated immediately before addition. Each drug was administered by pipette directly into the culture chamber, followed by gentle agitation to ensure a homogenous concentration in the medium around the cell. In all experiments, final solvent concentration was <0.1% and was determined as having no measurable effect on either Ca2+ flux or the stability of ionophore in the electrode tip.

Part D: effect of adenosine and the A1-receptor antagonist 8-phenyltheophylline (8-PT).

To examine the interplay between oxygen availability and adenosine purinoceptor-mediated control over Ca2+ flux, cells were plated as above and exposed to normoxia, normoxia + 100 μM adenosine, normoxia + 10 μM 8-PT, anoxia + 100 μM adenosine, and anoxia + 10 μM 8-PT. Changes in transmembrane Ca2+ flux were followed as before. Final vehicle concentrations for adenosine and 8-PT solvents (70% ethanol and dimethyl sulfoxide, respectively) were <0.1% after addition.

Part E: effect of KCN addition and glycolytic blockade on normoxic Ca2+ flux.

To assess the underlying oxygen- and energy-dependent mechanisms regulating Ca2+ flux, cells were plated as above and then exposed to 10 mM KCN followed by 10 mM iodoacetate (IAA). Previous experiments have established these concentrations as sufficient to inhibit oxidative phosphorylation and glycolysis, respectively (4, 17). Ca2+ self-referencing electrode measures were continued to the point of cell rupture.

Statistics and Data Presentation

Data were collected as described before (24) by using DVIS 10 software (BRC, MBL, Woods Hole, MA). Note that the computer is configured such that a negative flux signal represents a net Ca2+ efflux from the cell, whereas a positive signal represents an influx. Unless otherwise noted, significance was assessed by using one-way analysis of variance with Dunnett’s t-test, with confidence limits at 95% and value of nreflecting the number of individual hepatocyte preparations used in each experiment. Data from representative single experiments show either single data points or the means ± SD of every 100 data points collected.


Steady-State Ca2+ Flux

An extracellular Ca2+ gradient can be measured around a single turtle hepatocytes. Figure1 A demonstrates the effect of moving the oscillating electrode away from the cell membrane, through the extracellular Ca2+ gradient. Ca2+ flux diminishes toward background over the first micron of movement away from the cell, and a stable background measurement is reached at a distance of 5 μm. Figure1 Bdemonstrates the effect of increasing excursion amplitude through the extracellular gradient, toward a maximum excursion of 10 μm. An increase in the Ca2+-specific ΔμV signal was observed with larger excursion amplitudes. Figure 1,C andD, demonstrates the change in the magnitude of the Ca2+ signal (ΔμV, representative of the Ca2+ gradient) and Ca2+ flux over a 100-fold change in [Ca2+]e. Although there is considerable scatter, Ca2+ flux rates demonstrate a parabolic pattern, with maximum rates occurring between [Ca2+]eof 10–60 μM. Background measurements remained independent of [Ca2+]e.

Fig. 1.

A: dissipation of measurable Ca2+ flux with distance from cell membrane (representative single experiment). Electrode was vibrated over a 10-μm excursion and moved in increments outward from cell membrane. B: effect of increasing excursion amplitude on Ca2+ signal obtained from a point of closest approach to cell membrane. ForA andB, symbols represent means ± SD from data collected from 10 cells. C andD: effect of increasing extracellular Ca2+ concentrations on Ca2+ signal and Ca2+ flux, respectively. Both panels demonstrate a 2nd-order regression with 95% confidence limits presented in D. ▵, Background measurements taken 80 μm from cell; •, measurements taken at point of closest approach to cell membrane (at cell). Each plotted point represents mean of 50 data points/cell (n = 44 cells from independent cultures).

Characterization of the Ca2+ Flux Signal

A directional value of Ca2+ flux can be obtained by applying the ΔμV signal into a modified Fick equation. Consistent with findings from a wide range of cultured cells (24), this technique resolves a steady-state Ca2+ efflux that was sustained over at least 1 wk in culture. Vanadate and lanthanum ions can render broad concentration-dependent effects on a wide variety of Ca2+ transport mechanisms; however, both inhibit Ca2+-ATPase activity over micromolar concentrations (7, 10). Table1 demonstrates the inhibitory effect of vanadate or lanthanum ions on the magnitude of the Ca2+ efflux. In each case, a maximal inhibition of Ca2+ efflux to 30% of control values was observed at 50 and 100 μM for vanadate and La3+, respectively. Substitution of Li+ for Na+ in the culture medium caused a moderate, but statistically insignificant, increase in the magnitude of the Ca2+ efflux signal over 50 min of culture.

View this table:
Table 1.

Cation substitution effects on Ca2+ efflux

Ca2+ Flux During Progressively Severe Hypoxia

Figure2 Apresents the pattern of apparent Ca2+ efflux during anoxia and normoxic recovery. Anoxia resulted in a 90% suppression of Ca2+ flux compared with normoxic controls, with the suppression occurring well before oxygen was completely diminished from the medium. Reinstatement of normoxic oxygen tensions resulted in a rapid recovery of Ca2+ flux to preanoxic control values. Compiled data given in Fig. 2 Bdemonstrate that the anoxic suppression in Ca2+ flux was statistically significant relative to normoxic controls. Over the hypoxic transition, the half-maximal flux rate (apparentK m) lay at 145 μM of oxygen (Fig.3), suggesting that pathways regulating the suppression of the apparent Ca2+efflux have the capacity to respond to an oxygen signal over high oxygen concentrations.

Fig. 2.

Ca2+ flux during progressive hypoxia and normoxic recovery. A: representative single experiment. Data points are means with negative SD shown over 50 data points collected. Long horizontal dashed line represents mean flux at background position.B: compiled results. Values are means with SD (n = 5). Bg, background; Nor, normoxia; Anx, anoxia; Re-Oxy, reoxygenation. * P < 0.05.

Fig. 3.

Oxygen conformity of Ca2+ flux. Plot presents mean of every 50 data points from compiled results given in Fig. 2 B. Dashed lines: 95% confidence limit (n = 5).K m, Michaelis constant.

Oxygen Availability and PKC Activation and Inhibition

Figure4 Apresents a representative recording of Ca2+ flux from a single cell, as affected by the PKC activator TPA and the PKC inhibitor sphingosine. Grouped data are represented in Fig.4 B. Addition of 50 nM TPA caused a statistically significant reduction in Ca2+ efflux toward baseline, which is sustained for several minutes. Sequential addition of TPA to 100 and 150 nM lead to pronounced reductions in signal amplitude. Addition of 50 μM sphingosine (reduces the availability of free diacylglycerol) returns Ca2+efflux to normoxic control values. Under anoxia, sphingosine administration led to a rapid and pronounced Ca2+ influx to +4.3 ± 1.6 10−18 mol Ca2+ ⋅ μm−2 ⋅ s−1, (n = 6) that was shortly followed by cell swelling and rupture (Fig. 5).

Fig. 4.

Effect of protein kinase C activation by phorbol 12-myristate 13-acetate (TPA), and inhibition by sphingosine (Sphg), on Ca2+ flux rates.A: representative single experiment. Data points are means with negative SDs over 50 data points collected. Horizontal dashed line represents mean flux at background position.B: compiled results. Note that 150 nM TPA treatment was repeated in only 2 experiments and, therefore, is not represented in grouped data. Values are means with a SD (n = 7). * P < 0.05.

Fig. 5.

Representative single experiment in real time showing rapid reversal of Ca2+ flux during anoxic administration of 50 μM sphingosine. Each plot is a single data point. Experiment was taken as representative fromn = 6 measurements.

Oxygen Availability on Adenosineand 8-PT-Mediated Ca2+Flux

Adenosine has been widely implicated as a signaling metabolite in the control of both systemic and cellular adjustments during oxygen-mediated metabolic suppression. In hepatocytes, adenosine mediates increases in free [Ca2+]i(20, 26); therefore, it was of interest to examine the effects of oxygen availability on adenosine-mediated control over Ca2+ flux. In normoxia, administration of up to 100 μM adenosine resulted in a near-twofold increase in the magnitude of Ca2+ efflux, which was effectively abrogated toward control by addition of the A1 purinoreceptor antagonist 8-PT (Fig. 6). In anoxia, neither adenosine nor 8-PT administration significantly altered the oxygen-mediated suppression of Ca2+ flux.

Fig. 6.

Effect of aerobic and anaerobic administration of adenosine (AD; 100 μM) and A1-purinoceptor antagonist 8-phenyltheophylline (8-PT; 10 μM) on Ca2+ efflux. Values are means ± SD (n = 6). * P < 0.05.

Effect of Metabolic Blockade on Ca2+ Flux

The energetic requirements of sustained Ca2+ flux were assessed by inhibiting oxidative phosphorylation through addition of KCN, followed by complete metabolic blockade with addition of the glycolytic inhibitor IAA (Fig. 7). Aerobic addition of KCN did not significantly alter rates of Ca2+ flux from the control “at cell” condition. Addition of 10 mM IAA, together with the 2 mM KCN already present in the medium, led to a general reduction in Ca2+ flux toward baseline and, in three out of five cases, resulted in an apparent overall reversal in the direction of Ca2+ flux before cell rupture. In each experiment, KCN and IAA together ultimately led to blebbing, followed by cell death.

Fig. 7.

Representative single experiment demonstrating effect of metabolic blockade on aerobic Ca2+ flux after additions of 2 mM KCN followed by 10 mM iodoacetate (IAA). Long horizontal dashed line denotes mean background flux. Data points are means with negative SDs over each 50 data points recorded. Experiment was taken as representative of n = 5 measurements.


The noninvasive Ca2+-selective self-referencing electrode resolves a steady-state apparent Ca2+ efflux that can be modulated by vanadate, La3+, hypoxia, PKC activation and inhibition, adenosine, and the A1 purinoreceptor antagonist 8-PT. The efflux did not stem from a chronic loss of [Ca2+]i, as we were able to resolve it at a constant rate over 1 wk of culture, a time frame well beyond that supported by uncompensated Ca2+ efflux, at the rates measured in ths study. Furthermore, [Ca2+]eremained well above the expected level of [Ca2+]i, ensuring an inward diffusional force. This technique resolves a similar pattern of Ca2+ efflux among a wide range of biological systems (24). To examine this phenomenon further, we are pursuing two possibilities; first, that the efflux compensates Ca2+ influx from a point distant to the site of measurement; and second, that the time required for data processing results in a preferential resolution of slow Ca2+-transport events such as pump activity. In support of the latter, we observed a 70% inhibition of Ca2+ efflux by vanadate and La3+ over concentration ranges that characteristically inhibit Ca2+-ATPase activity.

This study demonstrates that oxygen availability, and its interaction with second-messenger systems, plays a central role in regulating pathways of Ca2+ homeostasis during the transition toward hypoxia. Progressive hypoxia was associated with a reversible oxygen-dependent suppression of Ca2+ efflux with an apparentK m for oxygen of 145 μM. This suggests that the suppression of Ca2+ efflux is achieved through a signaling mechanism capable of detecting changes in oxygen availability over concentration ranges that are two orders of magnitude above theK m for oxygen of the electron-transfer system in liver [0.7 μM oxygen for tightly coupled isolated liver mitochondria (27)]. In support of this was the failure of KCN, administered over concentration ranges that achieve rapid inhibition of oxygen consumption in turtle hepatocytes (17), to diminish Ca2+flux when oxygen was present in the medium.

The oxygen signal was transduced by a pathway that involved the activation of PKC, as anaerobic inhibition of PKC with sphingosine eliminated the controlled suppression of Ca2+ efflux, causing a marked Ca2+-influx followed by cell swelling and rupture. According to the literature, the role PKC might play in coordinating the suppression of Ca2+ efflux during hypoxia is vague. Aerobic experiments have detailed a role for PKC in the opening of capacitatively coupled Ca2+channels in the plasma membrane after release of internal Ca2+ pools by thapsigargin (1, 2). This fits with the case in rat hepatocytes, where PKC activation by TPA has been observed to prolong the time course of cytosolic Ca2+ resequestration into cellular compartments during glucagon-induced Ca2+ cycling (23). In glioma cells, PKC activation by TPA results in an increase in [Ca2+]ithat stems, in part, from an activation of plasma membrane Ca2+ channels (2). In normoxia, there is clearly a precedent to suggest that PKC activation opens capacitatively coupled Ca2+channels.

In the present study, we identified both an aerobic TPA-induced reduction in Ca2+ efflux and an absolute requirement for PKC activation in the controlled suppression of Ca2+ efflux toward a new steady state during the hypoxic transition. Given the clear role of PKC activation in the opening of Ca2+channels, we cannot exclude the possibility that the observed suppression of Ca2+ efflux was associated with an elevated, inwardly directed Ca2+ flux component. When compounded together, the absolute requirement for PKC activation, the ensuing sustained pattern of hypoxic Ca2+ flux, and the prolonged survivability of these cells during anoxia serve to underscore this as a regulatory event rather than as a precursor to the uncontrolled capacitative increase in [Ca2+]i, the central feature of the cell-death cascade in anoxia-intolerant tissues (8, 10, 19).

Receptor-mediated Ca2+ entry in hepatocytes is under the control of two distinct, agonist-activated pathways that exhibit different cation selectivities and modes of control (20). The first operates through an agonist-sensitive (vasopressin or thapsigargin) inositold-myo-inositol (1,4,5)-trisphosphate [Ins(1,4,5)P3] -generating pathway to mobilize the intracellular Ca2+-pool followed by a Ca2+-specific influx over the plasma membrane by capacitative coupling. This pathway is distinct, as it will not conduct Mn2+. The second mechanism is less selective for Ca2+ over Mn2+ and requires continuous hormone-receptor binding to a G-protein complex with the subsequent activation of Ca2+ channel conductance (20, 25). In the present study, it was important to assess the capacity for the oxygen signal to interact with existing receptor-mediated pathways of Ca2+mobilization. We examined the adenosine-purinoceptor-mediated pathway of Ca2+ mobilization, because the production of extracellular adenosine during periods of metabolic stress has the general effect of reducing energy demand and increasing energy supply, making this an important regulator of pathways associated with the survival of anoxia (22). In hepatocytes, specific aerobic effects of adenosine include a suppression of protein synthesis (26) and gluconeogenesis (21), activation of glycogen phosphorylase, and enhanced rates of glycogenolysis (12) and urea synthesis (9). All events are associated with increases in [Ca2+]i, with activation of glycogenolysis and urea synthesis demonstrating an absolute dependence for the Ca2+signal. Adenosine binding to the purinoceptor signals increases in [Ca2+]iby activating the Ins(1,4,5)P3-mobilized Ca2+ pool with subsequent capacitative entry of Ca2+ over the plasma membrane.

Adenosine administration activated Ca2+ efflux almost twofold over normoxic controls. The response was abrogated in the presence of 10 μM of the specific A1 subclass purinoceptor antagonist 8-PT. As adenosine is well characterized to transiently increase [Ca2+]iin hepatocytes, we interpret the increase in apparent Ca2+ efflux as an elevation in outwardly directed Ca2+-ATPase activity to compensate capacitative Ca2+ entry. When repeated under anoxic conditions, neither adenosine nor 8-PT administration altered the characteristic suppression of Ca2+ efflux, suggesting a capacity for the oxygen-signaling pathway to behave as a modulator of existing regulatory pathways of Ca2+ efflux during severe oxygen lack.

The oxygen-dependent suppression of Ca2+ flux in turtle hepatocytes indicates that processes associated with Ca2+ homeostasis are under tight control during the onset of hypoxia. Taken together with previous studies, we now have an emergent picture of anoxic survival in turtle hepatocytes that involves the selective shutdown of individual cell processes involved with both ATP supply and demand, an oxygen-regulated change in protein expression profiles, preservation of the cellular membrane potential, and the avoidance of a lethal accumulation of cytosolic Ca2+ by oxygen-dependent modulation of Ca2+ fluxes. Perhaps most importantly, this study demonstrates that critical cell processes, such as protein turnover (17) and Ca2+ flux, can be regulated by oxygen-sensitive signal transduction mechanisms that enable an adaptive response to be mounted before oxygen availability is critically limiting to aerobic metabolism. Moreover, this form of control appears to possess the capacity to override or modulate other mechanisms involved in the regulation of the same cell process.

In conclusion, tight control over patterns of Ca2+ flux appears to be an important component of the evolved mechanism of anoxia tolerance in turtle hepatocytes. This pattern of control is acutely sensitive to oxygen availability with the mechanism for the detection of the oxygen signal being independent of the role of oxygen in electron transfer. The oxygen signal juxtaposes second-messenger signaling systems displaying an absolute requirement for PKC activation and a capability to modulate adenosine A1-receptor-mediated changes in Ca2+ efflux.


Comments from an anonymous reviewer significantly improved the presentation of this work. We also thank Jane Maclaughlin for her helpful suggestions.


  • Address for reprint requests: S. C. Land, Biocurrents Research Center, Marine Biological Laboratory, Woods Hole, MA 02543 (E-mail:sland{at}

  • This work was supported by a Council for Tobacco Research USA, Inc. operating grant to S. C. Land and by a National Institutes of Health, National Center for Research Resources operating grant to P. J. S. Smith.

  • A preliminary version of this work appeared as a short report after presentation at the General Scientific Meetings of the Marine Biological Laboratory, Woods Hole, MA (see Ref. 18).


View Abstract