INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

SURVIVAL OF NEURONS in the brain of the extremely anoxia-tolerant Western painted turtle Chrysemys picta is associated with depression in the activity of key ion channels that preserve critical ion gradients (2, 5, 17). These changes are part of a suite of adaptations that are thought to coordinately decrease ion pumping and energy expenditure in hypoxia-tolerant cells (10, 11). An example of an important ion channel undergoing inactivation during anoxia in the turtle brain during anoxia is the N-methyl-D-aspartate (NMDA) receptor, which decreases in activity by 30-40% (4, 5). The NMDA receptor mediates fast excitatory neurotransmission and, during hypoxic/ischemic insults in the mammalian brain, contributes to uncontrolled Ca²⁺ influx and Ca²⁺-mediated cell injury (7). In turtles, alterations in NMDA receptor activity during the weeks to months of anoxia experienced during submergence dormancy could be important for survival, but long-term changes in receptor properties have not been studied.

One of the key factors thought to be important for the downregulation of ion channel function during anoxia in turtles is the accumulation of adenosine in the brain extracellular space (14, 16, 17). Adenosine has been shown to reduce the activity of potassium channels (17) and to partially explain the reduction in NMDA receptor ion channel activity as well (4, 5). However, adenosine levels increase over a 2-h period after the onset of anoxia and thereafter decrease toward baseline (16), leaving the role of adenosine in long-term receptor inactivation, such as that presumed to occur during prolonged dormancy, unclear.

Dramatic changes in the ionic composition of cerebrospinal fluid (CSF) of turtles during anoxic dormancy have important implications for the regulation of ion channels. The NMDA receptor, which is permeable to Ca²⁺ and Na⁺, is modulated by several extracellular ions whose concentrations change profoundly during prolonged anoxia. The NMDA receptor is inhibited by elevated magnesium and acidity (1, 22), whereas increased CSF Ca²⁺ levels will increase Ca²⁺ influx via the receptor channel (1). During 10 days of anoxia at 10°C, CSF Ca²⁺ concentration increases almost sixfold and Mg²⁺ concentration increases fivefold. pH decreases from 7.6 to 7.2, which should decrease NMDA receptor activity by 50% (22). We hypothesized that the increase in Mg²⁺ and acidity that occurs in CSF during anoxia depresses NMDA receptor activity sufficiently to compensate for the increased CSF Ca²⁺ concentration. To test this hypothesis, we measured NMDA receptor activity in turtle cerebrocortex during 6 wk of anoxia.

	METHODS

Top Abstract Introduction Methods Results Discussion References

With the approval of the University of California, San Francisco, Committee on Animal Research, 1-1.5 kg adult Western painted turtles (Chrysemys picta belli) were obtained from commercial suppliers and housed under 12:12-h light-dark photoperiods at 25°C. Studies were conducted during the fall and winter months when turtles are normally dormant. Turtles were placed in 5- or 10-gallon aquariums in a cold room at 2-4°C for several days to allow thermal acclimatization. Nitrogen gas was then continuously bubbled through the water. The aquariums were then sealed for periods between 2 h and 6 wk. A small gas head space at the top of the aquarium was allowed to exist for 1-2 days. Turtles became inactive and minimally responsive during this time. Samples of water showed PO₂ was <5 mmHg. Two turtles not allowed a period of acclimation with a gas head space above the aquarium water died, but 18 other animals survived up to 6 wk without apparent injury.

To study the activity of cortical NMDA receptors, cortical sheets, containing a superficial layer of cortical neurons with abundant NMDA receptors (23), were incubated in turtle artificial CSF (aCSF) containing the Ca²⁺ indicator fura 2. The methods for dissecting, loading slices with fura 2, and measuring intracellular Ca²⁺ changes are exactly as described in Buck and Bickler (4). Cortical sheets from anoxic turtles were dissected and fura 2-loaded in a nitrogen atmosphere (Atmos bag, Fisher Scientific). Four to six sheets were obtained from each animal. During fura loading, slices were continuously bubbled with 95% N₂ and 3% CO₂. NMDA receptor-mediated changes in intracellular Ca²⁺ were measured during application of NMDA (final concentration 100 µM) to cortical sheets in a fluorometer cuvette. To prevent the exogenous NMDA from triggering action potentials and releasing neurotransmitters, sheets were incubated in 1 µM tetrodotoxin, 0.1 µM -conotoxin, and 0.5 µM agatoxin GIVA. With these sodium and Ca²⁺ channel blockers present, we assumed that application of NMDA triggered Ca²⁺ changes almost exclusively due to Ca²⁺ influx through the activated NMDA receptor complex (9).

The activity of the NMDA receptor during normoxia and 6 wk of anoxia was compared using two benchmarks: the response of the receptor in normal Ca²⁺ and in aCSF with the higher Ca²⁺ found in anoxic animals. This was done to follow the changing activity of the NMDA receptor in reference to a normoxic control and to discover if the response of the receptor to high Ca²⁺ changes during long-term anoxia. Basal intracellular Ca²⁺ concentrations ([Ca²⁺]_i) and Ca²⁺ changes after NMDA application were measured in one of two types of aCSF. The first was standard aCSF containing the following ions present in the CSF of normoxic turtles (in mM): 129 Na⁺, 91 Cl, 1.9 K⁺, 1.3 Ca²⁺, 1.2 Mg²⁺, 2.9 lactate, 10 glucose, and 41 HCO₃, bubbled with either 95% O₂ and 3% CO₂ or 95% N₂ and 3% CO₂, PCO₂ 25 mmHg, pH 7.63 at 25°C. The second was aCSF containing ions found in the CSF of turtles after 10 days of anoxic dormancy (8): 131 Na⁺, 62 Cl, 4.0 K⁺, 8.4 Ca²⁺, 5.1 Mg²⁺, 69 lactate, 10 glucose, and 31 HCO₃, bubbled with 91% N₂-9% CO₂, PCO₂ 64 mmHg to achieve pH 7.25 at 25°C. Cortical sheets were randomly assigned for treatment with one of these media and were studied only once.

Experimental design. Basal [Ca²⁺]_i and [Ca²⁺]_i changes produced by application of NMDA (final concentration in cuvette 100 µM) were measured in cortical sheets from normoxic turtles kept at 25°C, from normoxic turtles kept at 2-4°C for 5 wk, and from turtles kept in anoxic aquariums for between 2 h and 6 wk. Measurements were also made in cortical sheets from turtles after 24 h recovery in room air after 4 wk of anoxia. Cortical sheets obtained from anoxic turtles were kept in N₂-CO₂ bubbled aCSF during study. Measurements of [Ca²⁺]_i were done in a sealed fluorometer cuvette temperature controlled to 25°C.

Statistics. ANOVA with Dunnett's post hoc test was used to compare values to control. Data are expressed as means ± SD.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Eighteen of twenty turtles survived up to 6 wk of anoxia without apparent injury. Although the blood of these turtles was black in color at the time of brain dissection, the heart of all animals was beating at the time of death. The brain tissue of turtles after 1 or more weeks of anoxia appeared softer and more edematous than normoxic controls. The brain tissues of turtles recovered in air (at room temperature) for 1 day after 6 wk of anoxia appeared completely normal.

[Ca²⁺]_i during normoxia, anoxia, and recovery. [Ca²⁺]_i in cortical sheet neurons from room air-breathing turtles studied in oxygenated aCSF at 25°C was 135 ± 12 nM (n = 27). After 2 h of anoxia in vivo before brain dissection, [Ca²⁺]_i increased to 164 ± 23 nM (n = 13, P < 0.05) and by 3 wk of anoxia had risen to 182 ± 12 nM (n = 5, P < 0.01 compared with control). However, at the end of 40 days of anoxia at 3°C, [Ca²⁺]_i returned to approximately the same level as seen after 2 h of anoxia (Fig. 1). In turtles returned to room air after 4 wk of anoxia, [Ca²⁺]_i returned to preanoxic levels within 24 h.

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Intracellular Ca2+ levels in turtle cerebrocortex. Cortical tissue was obtained from turtles during 2 h to 6 wk of anoxia in vivo. Intracellular Ca2+ levels ([Ca2+]i) were measured in tissue bathed in either artificial cerebrospinal fluid (aCSF) containing ions present in normoxic turtles (1.3 mM Ca2+) or in turtle aCSF containing 8.4 mM Ca2+ (found in CSF after 6-8 wk of dormancy; see text for other ions present). * Significant difference compared with 2-h anoxia group for 1.3 mM Ca2+ conditions.  Significant difference for 8.4 mM Ca2+ conditions.

NMDA receptor activity. NMDA receptor-mediated Ca²⁺ changes (NMDA-Ca²⁺) were measured as an index of receptor activity. In cortical sheets, NMDA-Ca²⁺ decreased to 60 ± 14% of control after 2 h of anoxia (examples in Fig. 2). During 6 subsequent weeks of anoxia, NMDA-Ca²⁺ was measured in slices bathed either in anoxic normal Ca²⁺ aCSF (ionic composition and pH identical to normoxic turtles) or in anoxic high- Ca²⁺ aCSF that was more acidic and contained greater amounts of Ca²⁺ and magnesium (see METHODS). In anoxic aCSF containing 1.3 mM Ca²⁺, NMDA-Ca²⁺ remained stably reduced at ~60% of preanoxic controls during 5 wk of anoxia (Fig. 3). In anoxic high-Ca²⁺ aCSF, NMDA-Ca²⁺ responses were approximately twice controls during the first several weeks of anoxia (Figs. 2 and 3). However, after 3-6 wk, responses measured in this high-Ca²⁺, high-Mg²⁺, and low-pH medium were approximately equal to counterparts studied in standard aCSF. These changes represented a dramatic decrease in Ca²⁺ influx via the NMDA receptor in the presence of high extracellular Ca²⁺. This is illustrated by a decrease in the ratio of NMDA receptor activity in high and low Ca²⁺ (Fig. 4) during the 6 wk of anoxia studied.

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Examples of N-methyl-D-aspartate (NMDA)-mediated increases in intracellular Ca2+ concentration ([Ca2+]i) (expressed as change relative to normoxic controls). Arrows show time of injection of NMDA into study cuvette; compound remained in cuvette thereafter. Shown are representative traces for cerebrocortical tissue bathed in either 1.3 mM Ca2+ (A) or 8.4 mM Ca2+ (B) after 2 h or almost 6 wk of anoxia in vivo. Mean NMDA receptor activity during these periods of anoxia is shown in Fig. 3.

View larger version (19K): [in this window] [in a new window]   Fig. 3.   NMDA receptor activity (NMDA-Ca2+, in nM) in cortical sheets from normoxic turtles and from turtles after 2 h to 5 wk of anoxia. Receptor activity was assessed in anoxic aCSF containing ion present in CSF of normoxic turtles (1.3 mM Ca2+, see text) or in aCSF of turtles after 10 days of anoxia (8.4 mM Ca2+, see METHODS for complete ionic composition). * Significantly different from control in 1.3 mM Ca2+.

View larger version (13K): [in this window] [in a new window]   Fig. 4.   Decreasing activity of NMDA receptor in high Ca2+/high Mg2+/high acidity relative to that in normoxic Ca2+ levels during 6 wk of anoxia. Receptor activity is expressed as ratio of NMDA-Ca2+ responses in anoxic dormancy CSF (8.4 mM Ca2+) to those in standard aCSF (1.3 mM Ca2+).

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

I found that NMDA receptor activity in turtle cerebrocortex decreases during 6 wk of anoxia and that functional changes result in depressed Ca²⁺ permeability of the NMDA receptor ion channel despite a greater than four- to fivefold increase in CSF [Ca²⁺] that probably occurs over this time. These changes may play important roles in decreasing neuronal excitability and neuronal energy expenditure and preventing excessive Ca²⁺ influx during anoxic dormancy. As far as I am aware, this is the first demonstration of such a long-term adaptive change in an ion channel in a dormant animal.

Reduction of NMDA receptor activity during anoxia occurred during a 2-h period of anoxia, and when measured under pH and ionic conditions like that of CSF in air-breathing turtles, this same degree of inactivation was present for 6 wk. NMDA receptor inactivation during the first several hours of anoxia is mediated by both adenosine receptors (4, 5) and phosphorylation of the NMDA receptor complex by neuronal phosphatases 1a or 2b (2). Therefore, other mechanisms (e.g., receptor internalization or removal from the plasma membrane) are not necessarily needed to explain the long-term receptor suppression during 5 additional weeks of anoxia. However, the results do not rule out long- or short-term changes in receptor number, nor do they imply that phosphorylation or actions of adenosine are necessarily the only mechanism by which receptor inactivation occurs.

Changes in NMDA receptor activity in the presence of the low-pH, high-Mg²⁺, and high-Ca²⁺ conditions found in CSF during prolonged anoxic dormancy suggest a more complicated picture than simple single-step inactivation processes occurring at the beginning or end of anoxia. Apparently, the Ca²⁺ sensitivity of the NMDA receptor undergoes a continuing process of alteration during long-term dormancy, as suggested by the progressive decrease in the ratio of NMDA receptor activity in high and low Ca²⁺ during the 6-wk period of study (Fig. 4). Mammalian NMDA receptors are inhibited by low pH (22) and elevated Mg²⁺ (1, 22), whereas increased extracellular Ca²⁺ will increase Ca²⁺ influx via the receptor ion channel. I hypothesized that the low pH and high Mg²⁺ present in CSF during long-term anoxia would decrease the activity of the NMDA receptor to a significant degree compared with responses in 1.8 mM Ca²⁺. Clearly, this hypothesis is incorrect, since NMDA-Ca²⁺ in this medium is about twice as great as in standard aCSF, even though [Ca²⁺] is five times as great. NMDA-Ca²⁺ in high-Ca²⁺ aCSF decreased progressively during anoxia, suggesting that alterations in the Ca²⁺ permeability of the receptor occur during prolonged dormancy. The following processes could be involved in producing such long-term alterations in receptor function: 1) phosphorylation/dephosphorylation of regulatory sites on the cytoplasmic domains of receptor subunits; 2) allosteric modulation by regulatory factors including glycine, polyamines, protons, magnesium, and arachadonic acid; 3) changes in the number of NMDA receptors present in the cell membrane; and 4) expression of other NMDA receptor subtypes composed of different subunits. Hume et al. (12) found that single amino acid substitutions in glutamate receptor subunits markedly influence Ca²⁺ permeability. Further study will reveal which of these occur. It should be noted that numbers of NMDA receptors do not change in turtle forebrain during 2-6 h of anoxia (21).

[Ca²⁺]_i increased ~22% during 2 h of anoxia, and continued to increase for 3-4 wk, before returning to control levels. [Ca²⁺]_i could thus be an important messenger that initiates and sustains key regulatory processes that regulate ion channels such as the NMDA receptor. This hypothesis is supported both by the temporal relationship between the increase in [Ca²⁺] and NMDA receptor activity and data showing that elevated [Ca²⁺]_i inactivates the NMDA receptor. A Ca²⁺-dependent inactivation of NMDA receptors has been documented to occur by 1) Ca²⁺-dependent actin-NMDA receptor depolymerization (20) and 2) binding of Ca²⁺-calmodulin to the cytoplasmic domains of NMDA receptor NR₁ subunits (24). It is also possible that elevated Ca²⁺ activates protein kinases or phosphatases via Ca²⁺-calmodulin or other effector molecules. The actual role of Ca²⁺ in activating these processes in turtle neurons during anoxia will be defined in future studies.

Several methodological issues deserve comment. Estimates of [Ca²⁺]_i with fura 2 are subject to a number of possible problems, as reviewed by Moore et al. (15). It is possible that chemical or physical changes in the turtle cortical neurons during anoxia alter the properties of fura 2, such as dye binding affinity. However, such changes in fura 2 Ca²⁺ binding properties are unlikely to significantly alter our conclusions, since very large changes in viscosity or ionic strength are required to alter fura 2 Ca²⁺ binding by >10% (15).

Another potential problem is the way in which we have estimated NMDA receptor activity by assuming that activity is estimated by Ca²⁺ changes (Ca²⁺ influx) stimulated by NMDA. Even though we have shown that activity measured this way correlates well with receptor activity changes measured with cell-attached patch-clamp measurements of receptor open probability (5), conditions occurring during prolonged anoxia could rupture this correlation. Future studies with direct measurements of ion channel activity (channel open probability measurements with patch clamping) will be needed to settle this issue.

Perspectives

Significance of NMDA receptor inactivation during prolonged hypoxia and dormancy. The significance of the downregulation of excitatory receptors during anoxia relates to three key events that contribute to the long-term survival of the turtle brain during anoxia: 1) reduction in the consumption of ATP required for maintenance of electrical excitability and neurotransmission; 2) a reversible reduction in neuron excitability, which is consistent both with behavioral inactivity during dormancy and with the capability of increases in neuronal activity at the end of a dive or at the termination of dormancy; and 3) decreased possibility of lethal [Ca²⁺]_i accumulation.