Purines, that is, adenosine and ATP, are not only products of metabolism but are also neurotransmitters. Indeed, purinergic neurotransmission is involved in thermoregulatory processes that occur during normoxia. Exposure to severe hypoxia elicits a sharp decrease in body core temperature (Tco), and adenosinergic mechanisms have been suspected to be responsible for this hypothermia. Because ATP per se and its metabolite adenosine could have complex interactions in some neural networks, we hypothesize that both adenosine and ATP are involved in the central mechanism of hypoxia-induced hypothermia. Their role in the thermoregulatory process was therefore investigated in a 24-h hypobaric hypoxia (FiO2 = 10%), using CGS-15943, a nonselective antagonist of adenosine receptors, and suramin, an ATP receptor antagonist. Tco and spontaneous activity (AS) were monitored by telemetry in conscious rats, receiving CGS-15943 (10 mg/kg ip), suramin (7 nmol icv), or both. The same treatments were done in normoxia to evaluate the specificity of their thermoregulatory action observed in hypoxia. Suramin/CGS-15943 treatment blunted the profound hypothermia observed in control rats throughout the hypoxia exposure, whereas CGS-15943 treatment blunted hypothermia during only 3 h, and suramin treatment had no effect. These results suggest that suramin potentiates the CGS-15943 effects and consequently that adenosine and ATP signaling act in synergy. In normoxia, suramin/CGS-15943 induced an increase in Tco but to a far lesser extent than observed in hypoxia. Thus it might be suggested that the suramin/CGS-15943 blunting of hypoxia-induced hypothermia would be specific to hypoxia-induced mechanisms.
- body core temperature
- spontaneous activity
- hypobaric hypoxia
adenosine and atp are not only products of metabolism, but also neurotransmitters depending on the cell compartments where they act. Thus far, several types of purinergic receptors have been identified: four for adenosine (A1, A2a, A2b, A3) and five for ATP (P2X, P2Y, P2U, P2T, P2Z), which are not distributed in the same areas of the brain (16, 17).
In rats exposed to normoxic conditions, the purinergic neurotransmission is involved in thermoregulatory processes. On the one hand, conscious rats exhibited a decrease in rectal temperature after an intravenous injection of the adenosine agonists l-2-phenyl isopropyl-adenosine and 5′-N-ethylcarboxamido-adenosine, acting through A1 or A2 receptors (21). Moreover, 5′-N-ethylcarboxamido-adenosine and cyclopentyladenosine cause a decrease in brain temperature when injected in the preoptic area (27). Thus the pharmacological activation of adenosine receptors mediates an hypothermia through modification of central thermoregulatory processes. On the other hand, the intracerebroventricular injection of the stable ATP analog α,β-methylene-ATP results in hypothermia in the cold, no change at thermoneutrality, and an hyperthermia in the heat (10). Thus the activation of ATPergic neurotransmission might induce a poikilothermic-like responses, suggesting that it could limit the thermoregulatory zone.
In rats, exposure to severe hypoxia elicits a sharp decrease in body core temperature (2, 19, 20). Moreover, it is part of a global hypometabolic syndrome, along with locomotor activity reduction, food intake drop, and body weight loss (2, 13, 14, 20). In such a context, hypoxia-induced hypothermia has been related to a centrally regulated mechanism (5, 8) (for a review, see Ref. 25), which could involve several mediators: lactate, histamine, arginine-vasopressin, opioids, nitric oxide, and adenosine (25, 26). Adenosinergic mechanisms had been suspected to be responsible for the hypoxic hypothermia because cerebral adenosine concentration increases in rats after acute exposure to hypoxia (30, 32), and the intracerebroventricular injection of the nonspecific adenosine receptor antagonist aminophylline attenuates the extent of hypoxia-induced hypothermia in rats (1). In spite of a lack of specific information on the putative involvement of ATPergic neurotransmission in thermoregulation in conscious rats exposed to hypoxia, we hypothesize that adenosine and ATP are both involved in the central mechanisms of hypoxia-induced hypothermia. Accordingly, 1) a 5-min hypoxia exposure induces an ATP release from the ventrolateral medulla in anesthetized rats (9), and 2) the high activity of the brain ecto-nucleotidases observed in native tissues (18, 33) suggests that the ATP released is rapidly converted into adenosine. Its is thus likely that a separate analysis of both transmitters would be difficult, except through a pharmacologically selective inhibition. On a functional side, the analysis of the interaction between ATP and adenosine suggests that the overall response of ATP is double. In hippocampus, ATP and adenosine block the evoked neuronal activity synergistically, whereas in the nucleus of the solitary tract, ATP regulated the ratio of evoked vs. spontaneous depolarization. Indeed, ATP per se facilitated spontaneous depolarization, whereas adenosine reduces evoked depolarization (12). Because ATP is metabolized into adenosine, it can modulate physiological function such as thermoregulation as it is released in hypoxia; it was thus hypothesized that the blockade of purinergic neurotransmitters would have an effect on the hypoxia-induced hypothermia.
To analyze the role of ATP and adenosine in hypoxia-induced hypothermia, we carried out a study in hypobaric hypoxic conditions. We used CGS-15943, an antagonist of adenosine receptors (29), and/or suramin, a P2 receptor antagonist, which has been speculated to act more on P2X receptors (28). Because ATP and adenosine are also involved in thermoregulatory processes acting in normoxia, the investigation was replicated in a normoxic environment to evaluate the degree of specificity of their effects to the hypoxia-driven mechanisms.
MATERIALS AND METHODS
Sixty-four male Sprague-Dawley rats (OFA, Charles River Laboratories) weighing 280–300 g were used for the investigation. Each rat participated in only one experiment. The investigation was reviewed and approved by the institutional ethics committee for animal care.
After their arrival at the laboratory, the animals were housed for 2 wk in individual cages placed into a climatic chamber to allow adaptation to the new environmental conditions: ambient temperature (Ta), 22–24°C; relative humidity (rh), 50 ± 10%, and 12:12 h-light-dark (LD; lights on at 0700 and lights off at 1900). The rats were given standard food and water ad libitum. Throughout the investigation, body weight and food intake were measured daily at 0900.
Rats were anesthetized with pentobarbital sodium (60 mg/kg ip). A telemetric device (Physiotel TA10TAF40, Data Sciences, St. Paul, MN) was implanted into the abdominal cavity of each rat for continuous recording of body core temperature (Tco) and spontaneous activity (AS). Then, a 22-gauge guide injection cannula (C313G, Plastic One, Roanoke, VA) was implanted stereotaxically into the right lateral ventricle [stereotaxic coordinates: 0.8 mm caudal to bregma, 1.6 mm lateral to midline, and 4 mm ventral from the surface of the skull, according to the atlas of Paxinos and Watson (22)]. Three small stainless screws were placed into the skull to fix the cannula, which was secured with dental acrylic cement. The guide cannula was obstructed with a dummy cannula (C313DC) 0.2 mm longer than the guide cannula. After surgery, the animals went back to their individual cages. They received then a postsurgery antibiotic (Gentamicine, Panpharma, France, 1 mg/kg im) and an analgesic treatment (Finadyne, Schering-Plough, France, 1 mg/kg sc). The rats were allowed to recover for at least 2 wk thereafter.
The functionality of the intracerebroventricular probe was tested 1 wk before experiments with an intracerebroventricular injection of 5 μl of an ANG II solution (Sigma A9525; 50 ng/μl). The test was considered to be positive when the injection was followed by drinking behavior. The functionality of the intracerebroventricular probe was also tested before death. A 5-μl solution of Evans blue was intracerebroventricularly injected; then a lethal injection of pentobarbital sodium (300 mg/kg ip) was administered. The brain was removed and observed macroscopically. The probe was considered as permeable when the ventromedial part of the third ventricle was colored blue. Lastly, the location of the cannula was confirmed histologically. The animals were included into statistical analysis only when the probe was permeable and well located.
Microinjections into the right lateral ventricle were made in freely moving rats using an internal injection cannula (C323ICT) connected through a tubing (C313CS) to a 50-μl syringe (Hamilton, Reno, NV). The 5-μl injections were made over a period of 1–2 min, and the injection cannula was removed 2–3 min later.
Suramin sodium salt (Sigma-Aldrich, Saint Louis, MO) was dissolved in artificial cerebrospinal fluid (aCSF) at the amount of 7 nmol in 5 μl of aCSF. The aCSF was reconstituted with (in mM): 147.0 NaCl, 4.0 KCl, 1.2 CaCl2, 1.0 MgCl2 dissolved in sterile pyrogen-free water.
Suramin was injected intracerebroventricularly because it does not efficiently cross the blood-brain barrier (11, 23), whereas CGS-15943 was given intraperitoneally in order not to inject a large amount of DMSO directly into the brain.
A hypoxic environment was obtained through the use of a hypobaric chamber in which the ambient pressure was reduced from 938 mbar, a laboratory environment value considered to be normobaric, to 485 mbar. Under the hypobaric condition, oxygen partial pressure was considered to be equivalent to a 10% inspiratory oxygen fraction (FiO2) in normal barometric pressure. The other environmental characteristics of the hypobaric chamber were: Ta = 22–24°C, rh = 50 ± 10% and 12:12-h LD cycle (with lights on at 0700).
A normoxic environment was produced in a climatic chamber in which the following environmental characteristics were reproduced: Ta = 22–24°C, rh = 50 ± 10% and 12:12-h LD cycle (with lights on at 0700).
After the recovery period, the rats were randomly distributed into eight groups, according to the treatment and the environmental condition. The animals received one of four treatments: suramin and CGS-15943 (Suramin/CGS), suramin and DMSO (Suramin/DMSO), aCSF and CGS-15943 (aCSF/CGS), or both vehicles (aCSF/DMSO). The hour preceding the injections was considered to be the baseline period. Both injections were done between 0900 and 1000 h for the normoxic animals and between 1000 and 1100 for the hypoxic animals. After injections, the animals were replaced in their cages either in the climatic chamber (normoxic groups) or in the hypobaric chamber to be exposed to 485 mbar during 24 h (hypoxic groups). Another 24 h recording in normoxic condition was carried out in order to check delayed pharmacological and environmental effects.
The experiment was divided into four sessions of 16 rats. In each experimental session, two rats for each treatment group were recorded in each environmental condition.
Physiological measurements and data analysis.
Tco and AS were monitored with the implanted telemetry devices (Physiotel TA10TAF40, Data Sciences, St. Paul, MN). The Tco was considered as the measured value of the abdominal temperature, whereas AS was defined as changes in the transmitted power signal due to the transmitter position changes. Recordings were made using a computer-based data acquisition system (Dataquest IV, Data Sciences, St. Paul, MN) every 5 min during 10 s; thus, each 5-min epoch value was the average of the 10-s records.
Because AS is a variable that could be dependent on the telemetric device, the AS values (x) were normalized using the z-score calculation The parameters of z-score calculation, that is, standard error (σ) and mean (μ), were determined for each rat using the values acquired during the day preceding the injections. The Tco and AS data were averaged into 1-h samples.
The statistical analysis was done using the Statistica software (Statsoft-France, Maison-Alfort, France). The time course of Tco and AS was examined for statistical significance using a two-way ANOVA (suramin treatment and CGS-15943 treatment) for repeated measures (time), followed when necessary by a Fisher least significant difference post hoc test using the aCSF/DMSO group as reference. The ANOVAs were done separately in each 24-h recording and for each environmental condition. A difference was considered to be statistically significant when P < 0.05; tendencies were noted for P < 0.10. The values were expressed as means ± SE.
A correlation analysis between Tco and AS was carried out in normoxic conditions using the 5-min data sampled during the first 3 h after injections. Each point corresponds to the mean value of the treatment group for the 5-min epoch considered. The significance of the correlation coefficient (r) was tested and considered to be statistically significant when P < 0.05.
Baseline Tco measurements did not show any difference between rat groups (Table 1, Fig. 1A). After the injections, the aCSF/DMSO rats exhibited a higher Tco than during baseline (Table 1, Fig. 1A). The suramin treatment modified Tco in a time-dependent manner (suramin treatment, P < 0.001; time, P < 0.001; interaction, P < 0.001). For both suramin-treated groups, Tco dropped immediately after the injections to reach its minimal values during the first hour (Table 1, Fig. 1A). An increase in Tco was observed from the 6th to the 9th h after injections in Suramin/CGS rats compared with the aCSF/DMSO rats (Fig. 1A). No CGS-15943 treatment effect was observed on Tco. On the following day, suramin increased Tco independently of its circadian variation [suramin treatment, P < 0.001; time, P < 0.001; interaction, not significant (ns)] with a higher Tco in the Suramin/CGS group compared with the aCSF/DMSO group (Fig. 1A).
AS was similar in all groups during baseline (Fig. 2A). The injection of CGS-15943 is followed by a time-dependent increase in AS (CGS-15943 treatment, P < 0.05; time, P < 0.001; interaction, P < 0.001). The sharp increase in AS, observed in both CGS-15943-treated groups during the hours following the injections, reached its maximum during the first hour (Fig. 2A). In the lights-off active phase, only the Suramin/CGS rats exhibited several epochs of lower AS than in aCSF/DMSO rats. On the following day, a circadian variation without any remaining treatment effect was observed (CGS-15943 treatment, ns; suramin treatment, ns; time, P < 0.001).
Considering the first 3 h after injections, a significant positive correlation between Tco and AS (r = 0.87, P < 0.05, Fig. 3) was evidenced in the aCSF/DMSO rats. Such a correlation was not observed in the Suramin/DMSO rats, Tco being totally independent of AS. A weaker positive correlation existed between Tco and AS in the aCSF/CGS rats (r = 0.61, P < 0.05, Fig. 3), and a negative correlation was seen in Suramin/CGS rats (r = −0.89, P < 0.05, Fig. 3).
Baseline Tco measurements did not show any difference between rat groups (Table 1, Fig. 1B). During exposure to hypoxia, Tco decreased in aCSF/DMSO rats to reach minimal values 3 h after the beginning of the exposure. A slight increase in Tco occurred thereafter, that is, from the 3rd to 23rd h of hypoxia (Fig. 1B). During the hypobaric hypoxia exposure, CGS-15943 and suramin treatments modified Tco in a time-dependent manner without any interaction between them (CGS-15943 treatment, P < 0.001; suramin treatment, P < 0.001; interaction between drugs, ns; time, P < 0.001; interaction between time and suramin and CGS-15943 treatment, P < 0.01 and P < 0.001, respectively). During the first 3 h of hypobaric hypoxia exposure, the aCSF/CGS rats exhibited higher Tco values than the aCSF/DMSO rats. Thereafter, no difference was observed between the aCSF/CGS and aCSF/DMSO groups (Fig. 1B). During the first hour of hypoxia exposure, Tco dropped for both suramin-treated groups. Thereafter, Tco of the suramin/DMSO rats did not differ from that of the aCSF/DMSO rats, whereas the Tco of the suramin/CGS rats was higher than that of aCSF/DMSO rats. The latter difference remained throughout the following 20th h of hypoxia exposure (Fig. 1B). On the following day spent in normoxic condition, a daily variation in Tco was observed without any remaining treatment effects (CGS-15943 treatment, ns; suramin treatment, ns; time, P < 0.001).
During baseline, AS was similar in all groups (Fig. 2B). A time-dependent effect of CGS-15943 on AS was observed (CGS treatment, P = 0.07; time, P < 0.001; interaction, P < 0.001). Indeed, the aCSF/CGS rats exhibited higher AS values compared with aCSF/DMSO rats during the first 2 h of hypoxia but not after (Fig. 2B). On the following day spent in the normoxic condition, a daily variation in AS was observed without any remaining treatments effect (CGS-15943 treatment, ns; suramin treatment, ns; time, P < 0.001).
In our investigation, hypoxia induced a strong decreases in Tco and AS in the aCSF/DMSO-treated animals, both effects being achieved in 2–3 h and observed until the hypoxia was removed. Such a result is in accordance with other reports done in normobaric (2, 4, 5, 8, 20) or hypobaric conditions (20). This hypothermia is considered as a regulated Tco decrease since animals exposed to normobaric hypoxia selected lower ambient temperature than normoxic animals (5, 8). On the other hand, adenosine was a putative mediator of the hypoxia-induced hypothermia (25, 31) and was metabolically (33) and/or functionally linked to ATP (12). In this context, using a model of 24-h hypobaric hypoxia, we studied the role of purinergic, that is, adenosinergic and ATPergic, neurotransmission in hypoxia-induced hypothermia. Our results showed that 1) CGS-15943, an antagonist of adenosine receptors, reduced hypoxic hypothermia, and 2) suramin, a P2 receptor antagonist, reduced hypoxic hypothermia, especially in association with CGS-15943.
Role of adenosine in hypoxia-induced hypothermia.
The fact that blockade of the adenosine receptors due to CGS-15943 blunts the hypoxia-induced hypothermia suggests that adenosine is involved in hypoxia-induced hypothermia. This result confirms others from previous studies in which the use of other antagonists of adenosine receptors, like theophylline or aminophylline, attenuated the hypoxia-induced hypothermia (1, 6). The effect of the CGS-15943 lasted only 3 h, a limitation linked either to the half-life of the product (103 ± 30 min; see Ref. 7) or the occurrence of other regulation mechanisms. Nevertheless, our results correlate with those of the other studies, as in these studies, normobaric hypoxia exposure did not exceed 30 min (1, 6). During this phase, the effect of CGS-15943 is likely to be linked to hypoxia. Indeed, we did not observe any change in Tco in the normoxic environment, despite a strong increase in AS, which would likely increase it. This is in agreement with other reports showing that the blockade of adenosinergic neurotransmission induces an hyperthermia, depending to the dose used. Administration of adenosine antagonists, such as aminophylline, theophylline, or caffeine, at optimal doses, has been reported to have no effect on the temperature in different species (1, 3, 6, 24). In contrast, at higher doses, these antagonists increase rectal temperature in conscious rats independently from the ambient temperature or the method of administration (15, 21). However, it could also be suggested that the effect of blockade of adenosine transmission is dependent on its level of activation. Indeed, adenosine signaling is enhanced when adenosine concentration in the central nervous system is increased, especially in the case of hypoxia (30, 32) and injection of adenosine agonists (21, 27). Thus our result based on CGS-15943-induced adenosinergic inhibition suggests that adenosinergic neurotransmission is involved in hypoxia-induced regulated hypothermia at least in the early phase of hypoxia.
Role of ATP in hypoxia-induced hypothermia.
The second step of our study concerns the role of ATP in hypoxia-induced hypothermia. We studied it using suramin, a P2 receptors antagonist acting possibly more on P2X receptors (28). Suramin alone did not have any effect on hypoxia-induced hypothermia. However, when associated to CGS-15943, suramin was able to strongly reduce the hypoxic hypothermia. The interactive effect between suramin and CGS-15943 remained steady throughout the hypoxia exposure and disappeared when hypoxia was removed. These results merit discussion of the relationship between ATP and adenosine neurotransmission and the specificity of suramin-induced effects in hypoxic conditions. The time course of the suramin/CGS-15943 cotreatment and the environmental conditions could be divided in two phases: 1) the first 3 h during which both drugs are effective and 2) the last 12 h during which suramin is likely to act alone. During the first phase, it could be speculated that the physiological effects could be related to the interactions between both drugs and environment. During the second phase, suramin is likely to have remaining effects since its half-life (44–54 days) largely exceeds the duration of the investigation (see Ref 28). This feature is supported by the several hyperthermic episodes observed in suramin-treated rats exposed to normoxia more than 24 h after the injections.
In the first hour after the end of drug injections, suramin induced a hypothermia in the normoxic condition. A greater hypothermia was observed in hypoxic condition. This result supports a cumulative effect of suramin and hypoxia. In both environments, the suramin-induced hypothermia was not reduced by CGS-15943, suggesting that the mechanism of such hypothermia was only dependent on the acute interaction of suramin with the ATP receptors. This conclusion is supported by the fact that the correlation analysis between Tco and AS demonstrated that, during the first 3 h, suramin treatment acted only on Tco and CGS-15943 acted only on AS and that both effects were observed with the two treatments. However, the mechanism of the acute suramin-induced hypothermia remains to be further studied because it was not observed by others in a study in which the same amount of suramin was used (10).
After that, in both environmental conditions, we observed a hyperthermic effect of suramin principally when associated to CGS-15943. This effect was not linked to an enhanced locomotor activity but rather to a modification in thermoregulatory processes, as we did not observe an increase in AS at this time. Because the rats treated only with suramin exhibited the same hypothermia in hypoxic conditions as the control rats, it could be concluded that ATP blockade alone is unable to counteract the hypoxia-induced hypothermia. Thus this suggests that the hypothermic regulatory mechanisms initiated by hypoxia did not go through the sole ATP receptors. The extent of the CGS-15943-induced attenuation of hypothermia seemed to stabilize after suramin cotreatment beyond its half-life and throughout the hypoxic exposure. These results led us to surmise that the Tco decrease under hypoxia was driven by a first mechanism blocked by CGS-15943, whereas the steady state of the hypothermia was influenced by another mechanism affected by suramin. Thus the effects of these treatments might not be simply additive, but rather complementary. This suggests that an interaction between receptors might exist on the thermoregulatory processes activated during exposure to hypoxia.
During the past 12 h of hypoxia exposure, the difference in Tco observed in the early phase of hypoxia exposure remained steady: neither suramin alone nor CGS-15943 alone modified the hypoxia-induced hypothermia, whereas the association of both drugs blunted it. However, this effect could be nonspecific to hypoxia because the suramin/CGS-15943 rats exhibited several hyperthermic episodes in normoxia. Such an hypothesis seems unlikely because 1) there is a difference of magnitude in the cotreatment-induced hyperthermia depending on the environmental condition, the reduction of the hypoxia-induced hypothermia being far more important than the hyperthermia in normoxia, especially at the end of hypoxia exposure; and 2) the Tco difference observed in hypoxia between suramin/CGS-15943 and control rats disappeared as soon as hypoxia was removed. This suggests that the Tco difference was actually related to an interaction with hypoxia-activated thermoregulatory processes.
The first part of our results confirms that adenosine is involved in hypoxia-induced hypothermia. Moreover, the prolonged blunting of hypoxia-induced hypothermia by suramin and CGS-15943 cotreatment together with the lack of specific activity of suramin alone suggests that ATP acts probably in interaction with adenosine. Further studies are required to analyze the interplay between both neurotransmitters.
We thank Dr. Alain Buguet for his critical review of the manuscript, Mr. David Popieul for the animal care, and Mr. Denis Bernabé for the histological analyses.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 the American Physiological Society