INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

ALL VERTEBRATES exhibit cardiovascular and ventilatory responses to acute hypoxia that, together, help maintain an adequate oxygen supply to the metabolizing tissue. As an alternative strategy, some animals enter a hypometabolic state that reduces the actual rate of ATP turnover and lessens the demand on the cardiorespiratory systems during conditions of limited oxygen availability (16). The hypometabolic state is a regulated response that involves changes in membrane permeability, reductions in active membrane transport, and downregulation of various synthetic pathways, such as protein synthesis (2, 16, 24, 25). Oxygen shortage is suggested as the causative factor that induces the hypometabolic state (20, 31), and recently some of the cellular transduction pathways for this response have been described (16, 17, 25).

In ectothermic vertebrates, several studies document downregulation of cellular metabolism in response to anoxia; turtles, for example, reduce heat production by 85% during anoxic submergence (18, 26). Similar results are reported for amphibians (38). However, although the reduction in metabolism as a response to anoxia is well established for a variety of ectothermic vertebrates (18, 26, 38), the effects of milder oxygen lack (i.e., hypoxia) have received less attention. It therefore remains to be investigated whether ectothermic vertebrates exhibit metabolic depression during hypoxia.

Many reptiles and amphibians experience internal hypoxia caused by intermittent breathing patterns and the presence of large right-to-left (R-L) cardiac shunts (4, 22, 33, 35, 40). During diving and periods of apnea, the development of the R-L shunt can occur rapidly and reduces arterial oxygen levels, in some cases, to values approaching mixed venous oxygen levels (4, 40). In addition, the R-L shunt recirculates metabolically produced CO₂ to the systemic arterial blood as venous admixture, consequently increasing arterial PCO₂ (Pa_CO2) and reducing arterial pH (39). Is it possible, therefore, that the development of a R-L shunt and the subsequent changes in arterial blood gases and pH can trigger a hypometabolic state?

The present study was designed to determine if hypoxemia or a combination of hypoxemia and hypercapnia, in the range normally experienced by turtles during diving, reduces aerobic metabolism. We chose to study the red-eared slider (Trachemys scripta) because this species normally exhibits long-lasting breath holds that are associated with the development of large R-L shunts and reduced blood oxygen levels (4, 36, 40, 41).

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Experimental animals. This study was performed on freshwater red-eared turtles, T. scripta Gray (body mass, 1.1-1.3 kg; mean = 1.26 kg; n = 5) obtained from Lemberger (Oshkosh, WI) and air-freighted to Odense University. All animals were housed (4-6 wk before study) in a large (0.8 × 1.2 m) fiberglass tank containing fresh water and free access to dry platforms allowing for behavioral thermoregulation. Animals were fed on trout pellets several times a week, but food was withheld for at least 3 days before experimentation.

Animal preparation. On the day of experimentation, turtles were anesthetized with an intramuscular injection of pentobarbital sodium (Nembumal; 30 mg/kg). Animals were placed in the supine position, tracheostomized, and mechanically ventilated with humidified room air (2-3 breaths/min) at a tidal volume of 25-30 ml/kg (Harvard ventilator, model HI 665). Occlusive PE-50 polyethylene catheters were inserted into the left carotid artery and the jugular vein. These saline-filled catheters were advanced toward the heart ~2-4 cm. These catheters were used for blood sampling, and the venous catheter also acted as an infusion port for 2,4-dinitrophenol (DNP). To access the central vascular blood vessels, a 2- × 2-cm portion of the plastron was removed with a bone saw. The pectoral muscles were gently loosened from the excised piece, and bleeding from small superficial vessels was stopped by cauterization (ROBOZ RS-232). For measurements of systemic blood flows (Qsys), 1-cm sections of the left aortic arch (LAo) were freed from connective tissue for placements of 2S transit-time ultrasonic blood flow probes (Transonic System, Ithaca, NY). To enhance the signal, acoustical gel was infused around the blood flow probes. The small plastron piece was replaced and sealed with small strips of duct tape. Body temperature was continuously monitored using a thermistor inserted 3-4 cm into the rectum. After instrumentation the turtles were placed in the prone position for the remainder of the experiment.

Measurements. All measurements were made on anesthetized animals at an ambient temperature of 25 ± 2°C. Rate of oxygen consumption (O₂) and carbon dioxide production (CO₂) were computed by measuring the average inflow-outflow gas concentrations multiplied by the lung ventilation over a given time period. Inflow gas concentrations were sampled from the Wosthoff gas-mixing pump after humidification and outflow gases were sampled from a mixing chamber attached to the excurrent line of the Harvard ventilator. The expired O₂ and CO₂ concentrations were measured with a Normocap 200 (Datex Corp, Finland) that combined a paramagnetic O₂ sensor with an infrared detector for CO₂. Lung ventilation was continuously recorded with a pneumotrachograph connected to the expiratory line of the ventilator and attached to a differential pressure transducer (Validyne). The pneumotrachograph was calibrated at the beginning and end of each experiment. To adjust for small variations in body temperature between animals, O₂ and CO₂ were corrected to a body temperature of 25°C, assuming a Q₁₀ of 2, and are presented normalized by the weight of the animals in kilograms and corrected to STPD.

Blood samples were collected (1 ml/sample) at the end of each normoxic, hypoxic, or hypercapnic exposure and immediately analyzed for arterial PO₂ (Pa_O2) and plasma pH with a Radiometer (Copenhagen, Denmark) O₂ electrode (E5046-0) and a capillary pH electrode (PS-1 204), respectively. Both electrodes were maintained at 15°C in a BMS Mk3 electrode assembly. Total amount of O₂ and CO₂ content (ctO₂ and ctCO₂) in the blood and total CO₂ content in plasma (separated from the red cells after 2 min centrifugation at 12,000 rpm) were measured as described by Tucker (34) and Cameron (6) and corrected in accordance with Bridges et al. (1). Hemoglobin (Hb)-bound O₂ (HbO₂) was calculated as ctO₂  O₂ × Pa_O2, where O₂ is the solubility of O₂ in human blood at 25°C (7). The outputs from all electrodes were displayed on two PHM 73 meters (Radiometer). Pa_CO2 and plasma HCO₃ concentration ([HCO₃]) were calculated from the measurements of ctCO₂ and pH using the Henderson-Hasselbalch equation and the pK' and CO₂ solubility values provided by Nicol et al. (28).

Hb concentration was measured spectrophotometrically at 540 nm after conversion to cyanomethemoglobin and applying a millimolar extinction coefficient of 11.0 (42). The fractional Hb oxygen saturation was calculated as HbO₂ relative to the O₂ capacity of functional Hb (i.e., total Hb minus methemoglobin). Hematocrit was determined after 3 min centrifugation at 12,000 rpm in capillary tubes. Plasma lactate was measured enzymatically with Sigma kits (nos. 826-UV and 171-UV, respectively). Potassium was determined with a Perkin Elmer (2380) atomic absorption spectrophotometer.

Beat-to-beat heart rate (f_H) was calculated on the basis of the instantaneous blood flow profile in the LAo. Qsys was determined from the average blood flow in the LAo (QLAo). Several studies on anesthetized and nonanesthetized freshwater turtles show that Qsys can be accurately estimated from QLAo × 2.85 (8, 33, 35). Signals from the differential pressure transducer, blood flow meter, O₂ and CO₂ analyzer, oximeter, and thermistor were continuously collected onto a computer with a commercial data acquisition system (AcqKnowledge MP 100, Goleta, CA) sampling at 50 Hz.

Hypoxic and hypercapnic gas mixtures were prepared with a Wostoff gas mixing pump (Bochum, Germany). The DNP (Sigma) was prepared as a 1.5% solution in sodium bicarbonate at a pH of 7.5.

Experimental protocol. Experiments were begun after a 30-40 min period to ensure that expired gases were in steady-state conditions. The experiment was divided into three phases. In the first phase, the animals were exposed to normoxia [fractional inspired O₂ (FI_O2) = 0.21, balanced N₂] for 30-40 min. This was followed by a 30-min exposure to hypoxia (FI_O2 = 0.05, balanced N₂). The FI_O2 was then increased (FI_O2 = 0.10, balanced N₂) for an additional 30 min. Finally, the animals were returned to normoxic levels for an additional 30 min. In the second phase of the experiment, the animals were exposed to a normoxic-hypercapnic gas mixture [FI_O2 = 0.21, fractional inspired CO₂ (FI_CO2) = 0.05, balanced N₂] for up to 1 h. The order of administration of the hypercapnic hypoxic gases was the same as in the first phase. The final phase of the experiment was conducted at the end of the FI_O2 = 0.1, FI_CO2 = 0.05 exposure. A single dose of DNP (20 mg/kg) was injected intravenously during hypoxia (FI_O2 = 0.1, FI_CO2 = 0.05), and all measurements were made 20-30 min after injection. After the experimental protocol was complete, all turtles were killed by intravascular injections of KCl.

Data analysis and statistics. All recordings of blood flows were analyzed with AcqKnowledge data analysis software (version 3.2.3; Biopac). For each oxygen level mean values for QLAo, f_H, ventilation, and arterial saturation were taken for the last 5 min of the exposure period. A two-way ANOVA for repeated measures was employed to assess the effects of reductions in FI_O2 and increased FI_CO2 on the reported parameters. Mean values that were significantly different from the control condition (FI_O2 = 0.21 and FI_CO2 = 0.00) were identified by a subsequent post hoc Dunnett's test. For each animal, the relationship between O₂ and arterial and venous oxygen levels was determined by regression analysis. From this regression analysis, the effects of DNP injections on O₂ were compared with the O₂ predicted by the measured blood gases. Differences between predicted and measured values were tested by means of a one-tailed paired t-test. A fiducial limit for significance of P  0.05 was applied, and all data are presented as means ± SE.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Effects of hypoxia. Reductions in FI_O2 and the associated decrease in PO₂ and arterial and venous ctO₂ (Table 1) were accompanied by a significant decrease in oxygen uptake (P = 0.017). In normocapnia, O₂ was maintained at the normoxic level at a FI_O2 of 0.10, but a further reduction to 0.05 elicited a significant reduction to 73% of the normoxic value. In all animals, the reduction in O₂ at the lowest FI_O2 was reversed on exposure to normoxia (Fig. 1). Hypercapnia increased Pa_CO2 and venous PCO₂ approximately threefold and was associated with a 0.4% reduction of pH (Table 1). Because plasma [HCO₃] remained unaffected, this acidosis was a result of the increased PCO₂. Hypercapnia alone did not reduce O₂ during normoxia and did not enhance the effects of hypoxia (Fig. 2). During hypercapnia, a reduction of FI_O2 to 0.10 reduced O₂ to 76% of the normocapnic-normoxic value, and O₂ was further reduced to 62% at a FI_O2 of 0.05. The reductions in O₂ during normocapnic-hypoxia and hypercapnic-hypoxia were not associated with significant changes in Qsys, f_H, or systemic stroke volume (Table 2).

View larger version (28K): [in this window] [in a new window]   Fig. 1.   Rate of O2 consumption (O2) of Trachemys scripta, as a function of fractional inspired O2 (FIO2). Reduction in O2 at FIO2 = 0.05 was reversed on reexposing animals to FIO2 of 0.21. Data are means ± SE; n = 5 animals.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Effects of reducing arterial (A) and venous oxygen content (B) on O2 in T. scripta (). Reduction in O2 was not enhanced by addition of CO2 to inspired air (). Injection of 2,4-dinitrophenol (DNP) () during hypercapnia-hypoxia increased O2 to levels measured during normoxia. All values represent means ± SE (n = 5 animals). * Significantly different from normocapnic-normoxic control; ** significantly different from preceding hypercapnic-hypoxic condition.

Effects of DNP injections. DNP caused a significant fall in arterial and venous ctO₂ and reduced PO₂, approaching the levels measured at FI_O2 of 0.05 (Table 1 and Fig. 2). Therefore, it was necessary to compare the measured O₂ after DNP injections with the O₂ that would have prevailed at these blood oxygen levels (Fig. 2). By this comparison, the O₂ after DNP injection was significantly higher than predicted by blood oxygen levels (P = 0.035). DNP injection was also associated with a large reduction in arterial and venous pH and a fourfold increase in plasma lactate (Fig. 3). Thus the reduction in pH resulted from the combination of lactate formation and increase in blood PCO₂ (Table 1).

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Effects of reducing arterial PO2 (PaO2) on plasma lactate in Trachemys scripta (). Injection of DNP significantly increased plasma lactate ().

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

This study shows that hypoxia induces a decrease in oxygen uptake in anesthetized turtles. This hypometabolic response is correlated with the reduction in arterial and venous oxygen levels, and was not associated with significant increases in plasma lactate. In all animals, increasing FI_O2 back to normoxic levels reversed the reduction in O₂. In addition, injections of DNP increased O₂ during hypercapnic hypoxia in some animals to levels greater than those measured during normoxia. We conclude that the reduction in O₂ most likely reflects downregulation of ATP turnover rates.

Comparison with data on conscious turtles and the choice of experimental preparation. Conscious turtles exhibit cardiorespiratory responses to reduced FI_O2 to compensate for the reduced oxygen availability and may increase activity in connection with escape behaviors (37). These responses are associated with some metabolic cost (21, 39) that may mask a hypometabolic state and confound the analysis of the data. This is not the case in the present study, because the in situ anesthetized preparation does not exhibit cardiopulmonary responses to hypoxia or changes in muscle activity during activity. Consequently, changes in oxygen consumption most likely reflect altered tissue metabolism. Nevertheless, as a potential disadvantage of this preparation, anesthesia generally reduces Qsys (9), and it is therefore possible that the reduced systemic oxygen delivery renders anesthetized turtles more sensitive than conscious animals to hypoxia. However, because our data concord with values for conscious turtles, and because our experimental design allows the confounding effects of activity to be dismissed, it is likely that the hypometabolic state reported here applies to conscious animals.

In conscious turtles, the effects of hypoxia on O₂ are markedly augmented with increased temperature (12, 19, 23). Thus at 10°C, O₂ is maintained at the normoxic level until Pa_O2 falls to <10 mmHg, whereas O₂ falls progressively to values as low as 50% of the normoxic value whenever Pa_O2 is reduced to 12 mmHg at 30°C (12). Extrapolations of the gas exchange data obtained in these previous studies on conscious turtles (12, 19) to 25°C are similar to those reported in the present experiments.

Delivery vs. a controlled response: DNP. The reduced O₂ during hypoxia may be caused by either an inability of the cardiopulmonary system to supply sufficient oxygen to the tissues or by a downregulation of metabolism, i.e., reduction of ATP demand. To distinguish between these two possibilities, we injected the metabolic uncoupler DNP and found that the hypoxic hypometabolism was reversed even though arterial and venous blood O₂ levels remained reduced and cardiac output remained constant. In some of the experimental animals, the level of O₂ after DNP injection was greater than O₂ measured during normoxia. The pharmacological stimulation of O₂ measured in our study is qualitatively similar to the results reported for isolated cells and similar to results reported in other animals. In the aquatic turtle, Chrysemys picta, isolated hepatocytes undergo rapid metabolic depression during anoxia that is reversed by treatment with DNP (3). In adult rats, hypoxia (FI_O2 = 0.10)-induced hypometabolism is reversed by a single injection of DNP (17 mg/kg), and the pharmacologically stimulated O₂ during hypoxia can exceed the normoxic value (30).

We conclude that the reduced O₂ during hypoxia in the present study reflects a reduction in ATP turnover rate. This conclusion is based on the observations that 1) plasma lactate concentrations did not increase significantly in hypoxia, indicating that the net ATP production derived from anaerobic metabolism was not increased; and 2) that O₂ could be pharmacologically stimulated by injection of DNP during hypoxia, suggesting that oxygen delivery did not limit aerobic metabolism. Similar conclusions have been reached previously with comparable data in both adult and newborn mammals during hypoxia (27, 29). In mammals, inhibition of the pathways involved in thermogenesis may be the most important contributor to the reduced metabolism, although other mechanisms play a role (10, 11, 27). In ectotherms, thermogenic processes are absent, and the reduction in O₂ at the organismic level may therefore be caused by the same mechanisms that depress metabolism during anoxia. As originally proposed more than a century ago, low oxygen levels appear to be the causative factor that leads to the suppression energy turnover (16, 27).

Hypoxemia-induced hypometabolism: a possible physiological role for R-L cardiac shunts. In freshwater turtles, diving is often associated with significant bradycardia, increased pulmonary vascular resistance, reduction in pulmonary blood flow, and the development of a large R-L intracardiac shunt (32, 33, 35, 40). In turtles, the magnitude of the cardiac shunt is determined by factors that affect heart rate, myocardial contractility, and the vascular resistances in the pulmonary and systemic circulations, and is primarily under cholinergic and adrenergic control (13, 14). These cardiovascular changes can occur rapidly on diving (<30 s), and as a consequence the addition of venous admixture to the arterial circulation rapidly reduces arterial blood oxygen levels toward venous values, whereas lung PO₂ declines more slowly (4, 40). It has been suggested that the reduction of pulmonary blood flow associated with the R-L shunt increases aerobic dive times by prolonging the use of lung oxygen stores (5). However, a recent theoretical analysis indicates that as long as tissue metabolism remains at predive levels, the development of a R-L shunt does not necessarily improve aerobic dive times (15). In our experiment, the arterial blood oxygen levels that initiated metabolic depression are similar to the levels measured in conscious turtles after the development of a R-L shunt during diving (4, 40). Thus the results of our study suggest that development of a R-L shunt may trigger the rapid onset of a hypometabolic state in the tissues and therefore contribute to the prolongation of aerobic dive times in freshwater turtles.