Main

The breathing pattern of newborn and adult humans and other animal species is markedly modified by changes in environmental and body temperature. In anesthetized adult humans and several other animal species, a significant reduction in respiratory rate, tidal volume, and alveolar ventilation has been described during HT (1–3). In unanesthetized adult rats, the ventilatory response to Hy and CO2 was markedly reduced when their body temperatures were decreased from 38 to 35°C (4). A significant decrease in the initial as well as the late ventilatory response to Hy was also observed in newborn infants exposed to a cold environment when compared with their ventilatory response in normothermic conditions (5). However, the mechanism responsible for this depressed ventilatory response has not been made clear. A possible explanation for this decrease in ventilation is the reduction in metabolic demands that has been shown in newborn kittens exposed to a lower environmental temperature when compared with those in a thermoneutral environment (6).

In recent years, there has been more emphasis on the effects of AA neurotransmitters and/or modulators such as glutamate and GABA on the ventilatory response to Hy. Experimental evidence suggests that the biphasic ventilatory response to Hy results from the interaction between the effects of inhibitory and excitatory neurotransmitters and/or modulators on the central respiratory neurons (7–9). Therefore, the Hy ventilatory depression observed during HT may be explained in part by either a decrease in the release of excitatory AA neurotransmitters, glutamate or aspartate, and/or an increase in the release of inhibitory AA neurotransmitters, GABA or glycine, in the CNS.

Recently, McCormick et al. (10) demonstrated that the initial increase in the ventilatory response to Hy was absent in hypothermic newborn piglets, and this was followed by a decrease in ventilation below baseline values after 15 min of Hy. This change in the ventilatory response to Hy was accompanied by a decrease in the extracellular concentration of glutamate in the nucleus tractus solitarius. However, this study has not ruled out the possibility that the depressed ventilatory response to Hy observed in hypothermic newborn piglets is in part due to an increase in the release of endogenous inhibitory AA neurotransmitters such as GABA.

GABA is an important central modulator of cardiorespiratory functions (9, 11, 12). It has been demonstrated that the administration of GABA or its agonists in the ventriculocisternal area of adult dogs results in a significant reduction in the basal ventilation (12). The administration of bicuculline, a GABAA receptor blocker, reversed the hypoxia (Hy) ventilatory depression observed in adult cats (13). Soto-Arape et al. (9) demonstrated that the application of bicuculline into the ventrolateral medullary surface of adult rats results in a significant increase in the phrenic nerve activity during normoxia and Hy and eliminates the “roll off” observed during Hy in the control group. In newborn piglets, the intravenous administration of GABAA or GABAB receptor blockers also reversed the hypoxic ventilatory depression observed before the drug administration (8).

Several studies have reported an increase in the release of GABA in the striatum during cerebral ischemia in adult animals (14–16). A greater increase in the brain GABA levels was observed during cerebral ischemia in adult rats when they were exposed to a cold environment (16). These studies suggest that HT may enhance the increase in the CNS GABA concentra-tion observed during Hy, and this increase in GABA levels may mediate the decreased ventilatory response to Hy during HT.

The present study was performed to test the hypothesis that the decrease in the ventilatory response to Hy observed in hypothermic newborn piglets is in part due to an increased CNS concentration of GABA. In this case, the hypoxic ventilatory depression observed in hypothermic conditions should be reversed by GABA receptor blockade. The objective of the study was to evaluate the effects of GABAA receptor blockade (BM) on the cardiorespiratory response to Hy in normothermic and hypothermic newborn piglets.

METHODS

Thirty-one newborn piglets (≤7 d) were anesthetized with an intraperitoneal injection of urethane (250 mg/kg) and chloralose (40 mg/kg).

The femoral artery and vein were cannulated for ABP and HR measurements, blood sampling, and the infusion of maintenance fluids and drugs. ABP was measured with a pressure transducer (model P-23 ID, Gould Instruments, Cleveland, OH) and recorded on a multichannel recorder (model 2800, Gould Instruments).

The central respiratory output was expressed as MPO. The phrenic nerve was dissected from the C5 root down to the thorax, desheathed, and cut distally. The nerve was placed on a platinum bipolar electrode and preserved with mineral oil-vaseline paste. The phrenic nerve activity signal was amplified (model S75–05, Coulborne Inst., Lehigh Valley, PA) and filtered through an adjustable band pass filter (48 dB/octave, model S75–34, Coulborne Inst.) at a bandwidth of 30 to 2500 Hz. The signal was digitized at 5000 Hz (AT-CODAS, Dataq Instruments) and recorded into a personal computer. MPO was analyzed by moving time average (area under the curve) multiplied by the nerve burst frequency over a 1-min collection period or a minimum of 10 breaths.

To avoid interference between the mechanical breaths and phrenic nerve output, a computerized system was developed to control the timing of the time-cycled pressure-limited ventilator (model IV-100B, Sechrist, Anaheim, CA). The phrenic nerve signal was band pass filtered (30–2500 Hz), rectified, passed through a moving time averager with a time constant of 200 ms, and amplified. This processed signal was digitized at 100 Hz by the computer system. Every burst of the phrenic nerve was identified and validated by the program that then triggered the ventilator after a preset time had elapsed at the end of each phrenic nerve burst (10).

VO2 was measured by sampling the exhaled gas from the side port located in the expiratory line of the ventilator (17). The difference between inspiratory and expiratory oxygen concentration was measured continuously using an oxygen analyzer (model 570-A, Servomex, Crowborough, Sussex, UK). VO2 was calculated by the following formula: VO2 = VS · (Fio2 − Feo2), where VS is the flow rate through the system, Fio2 is the fraction of inspired oxygen concentration, and Feo2 is the oxygen concentration in mixed expired gas.

The brain temperature of the normothermic and hypothermic groups was maintained at 38.5 ± 0.5°C and 34.0 ± 0.5°C, respectively. HT was accomplished by decreasing the temperature on the radiant warmer and water-heated blanket. To monitor the changes in brain temperature during the study period, a temperature probe (model 6510, Mallinckrodt Medical, St. Louis, MO) was inserted into the right side of the skull below the meningeal layer. The rectal temperature was continuously monitored with a thermistor probe (Mallinckrodt Medical).

Each piglet was placed in the prone position with its head fixed in a stereotaxic holder (model 1530, David Kopf, Tujunga, CA). An incision was made along the midline of the scalp and extended to the occipital region. A 2-mm-diameter hole was drilled through the occipital bone in the midline 1.0 cm posterior to the lambda. A 20-gauge guide catheter 3.2 cm long was inserted into the cisterna magna via this hole and cemented with dental acrylic. Then, a 22-gauge stainless steel pipette was inserted through the guide into the cisterna magna and used for administration of the drug (BM) or PL (Ringer's solution).

The animals were randomly assigned to either the NT or HT group and then further subdivided into two groups: PL or treated (BM). The groups were as follows: NT-PL, NT-BM, HT-PL, and HT-BM.

BM (Sigma Chemical Co., St. Louis, MO) was freshly dissolved in Ringer's solution (NaCl 140 mM, KCl 2.5 mM, CaCl2 1.3 mM, MgCl2 0.9 mM, Na2HPO4 4.8 mM, NaH2PO4 1.3 mM, distilled water 100 mL) and titrated with 0.2 M Na2HPO4 to a final pH of 7.34. Ringer's solution with pH of 7.34 was administered to the animals in the PL group.

The dose of BM for ICI was selected on the basis of neurophysiologic studies (18) that demonstrated that the administration of 5 to 25 μg of BM to the medullary ventral surface reversed the GABA effects in cats. In this study, the administration of a total dose of 25 μg of BM increased minute ventilation from 361 to 455 mL/min, blood pressure from 164 to 177 mm Hg, and HR from 191 to 204 beats/min. In our pilot study, different doses (10, 12.5, 15, 20, 25 μg) of BM were injected into the piglets by ICI during Hy under hypothermic conditions in an attempt to determine the dose that would elicit changes in MPO. A dose of 10 μg of BM produced an increase of 11.2% in MPO without significant change in systemic arterial pressure. This BM dose range (10–25 μg) did not trigger convulsions, which were monitored by EEG. EEG electrodes consisted of a pair of insulated stainless steel wires (Cooner Wire Co., Chatsworth, CA) connected to bifrontal stainless steel screws that were anchored into the skull and insulated with dental acrylic. The signal was processed by an EEG preamplifier (Gould) and acquired by a computer program by means of an analog-to-digital converter (AT-Codas, Dataq Instruments, Akron, OH).

After the surgery, the animals were paralyzed with pancuronium bromide (0.2 mg/kg as bolus, followed by 0.4 mg·kg−1·h−1), and the ventilator was adjusted to maintain a constant Paco2 within 2 mm Hg of the value obtained during spontaneous breathing. One hour after the surgery was completed, the brain temperature of the animal was adjusted to the temperature value of the assigned group. After 30 min of stable temperature, baseline measurements (MPO, ABP, HR, arterial blood gas, and VO2) were obtained during normoxic conditions. MPO, ABP, HR, and end-tidal Pco2 were continuously monitored. The animals were then exposed to an Fio2 of 0.10 immediately after the baseline measurements were obtained, and measurements were performed at 1, 5, and 10 min of Hy. An infusion of 20 μL of BM (10 μg) or Ringer's solution was administered over a 1-min period into the cisterna magna while the animal remained hypoxic. All measurements were repeated at 11, 15, and 20 min of Hy.

At the end of the experiment, a craniotomy was performed to confirm the location of the site of infusion.

Handling and care of the animals were in accordance with guidelines of the National Institutes of Health, and the study protocol was approved by the Animal Care Committee of the University of Miami.

Data are expressed as mean ± SEM. Data were analyzed by repeated measures ANOVA for comparison of the pattern of the cardiorespiratory response to Hy before and after the ICI of PL or BM within the group and between groups. ANOVA was also used to compare the baseline cardiorespiratory measurements between groups. A p < 0.05 was considered statistically significant.

RESULTS

No significant differences were observed in weight and ages among the following groups: NT-PL (n = 8, wt 1.8 ± 0.1 kg, age 4.6 ± 0.4 d); NT-BM (n = 7, wt 1.9 ± 0.2 kg, age 4.6 ± 0.6 d); HT-PL (n = 8, wt 1.9 ± 0.1 kg, age 4.5 ± 0.5 d); HT-BM (n = 8, wt 1.9 ± 0.1 kg, age 4.5 ± 0.5 d).

Figure 1 shows the percentage changes in MPO from room air to Hy before and after the ICI of PL or BM in NT and HT groups. When the NT piglets were exposed to Hy, an initial increase in MPO (50 ± 6%) was observed, followed by a decrease to values that remained above the baseline at 10 min (24 ± 8%). In the HT group, the initial increase in MPO with Hy was abolished, and this was followed by a gradual decline in MPO to values (35 ± 4%) below the baseline. The changes in MPO with Hy were significantly different between NT and HT groups (NT versus HT, p < 0.03). After administration of BM, there was a significant increase in MPO during Hy in both groups when compared with their PL groups (p < 0.002 in NT-BM group, p < 0.0001 in HT-BM group). This increase in MPO in the NT-BM group was mainly due to an increase in area under the curve (p < 0.01), whereas the burst frequency did not change. However, the magnitude of the increase in MPO was significantly greater in the HT-BM group. In this group, MPO increased by 76 ± 5% at 11 min, 70 ± 5% at 15 min, and 55 ± 4% at 20 min after BM infusion, whereas an increase of only 17 ± 4%, 16 ± 5%, and 9 ± 4% was observed at 11, 15, and 20 min in the NT-BM group (NT versus HT, p < 0.0001), respectively, when compared with 10 min of Hy. The marked increase in MPO during Hy after BM in the HT group was due to an increase in burst frequency (p < 0.001) and moving time average (area under the curve) of phrenic nerve activity (p < 0.002). Administration of Ringer's solution did not modify the changes in MPO during the additional 10 min of Hy in NT-PL and HT-PL groups.

Figure 1
figure 1

Percentage changes in MPO from room air (RA) to Hy before and after the ICI of PL or BM in NT and HT groups. A significant decrease in MPO was observed during Hy in the HT group before BM or PL infusion. After BM infusion, a greater increase in MPO occurred in the HT group compared with the NT group.

Full size image

Figure 2 illustrates the mean values for VO2 during normoxia and Hy in NT and HT groups before and after PL or BM infusion. Although there was a significant decrease in basal VO2 in the HT groups compared with the NT groups (p < 0.0003, HT-PL versus NT-PL;p < 0.003, HT-BM versus NT-BM), the decrease in VO2 with Hy was similar in both groups before and after PL or BM infusion.

Figure 2
figure 2

Changes in VO2 with Hy before and after the ICI of PL or BM. A significant decrease in basal VO2 was observed in the HT groups compared with the NT groups (p < 0.0003, HT-PL vs NT-PL;p < 0.003, HT-BM vs NT-BM), but the decrease in VO2 with Hy was similar in NT and HT groups before and after PL or BM infusion.

Full size image

Table 1 shows the values for arterial blood gases and acid-base status during normoxia and Hy in the NT and HT treated piglets. As expected, the Paco2 remained constant during normoxia and Hy in NT and HT groups. Changes in pH, Pao2, and base excess with Hy before and after the ICI of PL and BM were not different in NT and HT groups.

Table 1 Arterial blood gases, acid-base values, ABP, and HR in RA after 10 min of Hy and an additional 10 min of Hy after the ICI of BM or PL in normothermic and hypothermic conditions Values are mean ± SEM. Changes in arterial blood gases and acid-base, ABP, and HR values with Hy were not different between both groups before and after PL or BM infusion. RA indicates room air; BE, base excess.
Full size table

The mean values for ABP and HR during normoxia and Hy in NT and HT treated piglets are also presented in Table 1. There was no significant difference in basal ABP between the NT and HT groups. Although ABP increased in the NT group and decreased in the HT group during Hy before administration of BM, the increase in ABP after BM infusion was similar in NT and HT groups. Changes in ABP with Hy were similar in NT and HT groups after infusion of Ringer's solution. Basal HR decreased significantly in the HT group compared with the NT group (p < 0.0001), but changes in HR with Hy were not significantly different between groups before and after the ICI of PL or BM.

DISCUSSION

The present study demonstrates that the depressed ventilatory response to Hy observed in hypothermic newborn piglets is reversed by the ICI of BM, a GABAA receptor blocker. This suggests that the endogenous inhibitory AA neurotransmitter GABA may mediate the depressed ventilatory response to Hy in hypothermic newborn piglets through GABAA receptors in the brain stem area. Because bicuculline was infused into the cisterna magna, it is impossible to localize precisely the site of BM action on the CNS, and it is likely that the drug diffused beyond the area of the cisterna.

The biphasic ventilatory response to Hy observed in normothermic newborn piglets has been widely described (8, 10, 17, 19). The mechanisms underlying the late depression of ventilatory response to Hy have not been very clear, although some hypotheses have been put forward. One possibility is that the late decrease in ventilation during Hy is the result of an imbalance of the excitatory and inhibitory AA neurotransmitters in the area of central respiratory neurons (20). Glutamate is the most important excitatory AA neurotransmitter, whereas GABA has inhibitory effects on the central cardiorespiratory control mechanisms. When glutamate and GABA are released into synaptic clefts adjacent to respiratory neurons, they can bind to postsynaptic or other receptor sites, open ion channels, and change the polarization state or discharge rate of these neurons. Thus, respiratory output can be modulated by the interaction kinetics of glutamate and GABA with their specific receptors (21).

Yamada et al. (18) have shown that GABA and its agonist muscimol administered either into the cisterna magna or the fourth ventricle to chloralose-anesthetized cats caused respiratory depression, hypotension, and bradycardia. They also demonstrated that the application of GABA to the brain stem surface (Schlaefke's area) produced similar effects on respiratory activity and ABP. This effect of GABA was reversed by the application of BM to the same site (22). In addition, the application of bicuculline into the ventrolateral medullar area of adult rats eliminated the “roll off” of the phrenic nerve activity observed during Hy in untreated animals (9). We have previously reported that the intravenous administration of bicuculline to sedated newborn piglets increased minute ventilation by 13% at 10 min of Hy (8). In the present study, MPO increased by 9% at 10 min of Hy after an intracisternal administration of BM. This small discrepancy may be due to the different dose of bicuculline and the route of administration. These studies are in agreement with the findings of the present study that showed a significant increase in MPO with Hy in NT and HT piglets after administration of BM. However, the increase in MPO was significantly greater in the HT animals, suggesting that the lower temperature enhanced the increase in CNS GABA concentration and, thereby, the depression in ventilation observed during HT and Hy.

A possible explanation for the increase of CNS GABA during HT is a reduced rate of removal of GABA from the synaptic clefts adjacent to respiratory neurons. Na+-dependent binding to membranes and transport into intracellular sites may be the major mode of removal of GABA from synaptic sites in mammalian brain (23). This mode of removal of GABA by the Na+ pump is an ATP-dependent system. During Hy, energy metabolism becomes anaerobic and, consequently, the production of ATP decreases. As the Na+ pump fails, the removal of GABA from synaptic sites is reduced. Iversen and Neal (24) demonstrated that the uptake of GABA was temperature sensitive in rat cerebral cortex slices incubated in a medium containing [3H]-GABA. When incubation was performed at 0 and 15°C, the rate of uptake of [3H]-GABA was lower than that at 25°C.

A significant increase in the release of GABA during brain ischemia has been described in hypothermic adult rats (16). This suggests that the increase in GABA concentration in the central respiratory neurons during Hy is enhanced by HT. Another possible explanation for the increased inhibitory effects of GABA on the CNS respiratory neurons during HT is a lower extracellular concentration of excitatory AA, such as glutamate in the nucleus tractus solitarius. This finding was observed in our laboratory in hypothermic newborn piglets exposed to Hy (10). This may enhance the inhibitory effect of GABA on the central respiratory system. Because the activity of glutamic acid decarboxylase (GAD) was not assessed during Hy in hypothermic newborn piglets, we cannot rule out the possibility that the CNS glutamate and GABA concentrations were determined by the changes in GAD activity during Hy and HT. However, GAD is an anaerobic enzyme and, therefore, continues to convert glutamate to GABA during Hy (20). Additionally, in vitro, GAD remains stable for 1 h when the temperature is kept between 30 and 40°C (25). In the present study, hypothermic piglets were maintained for approximately 1 h at 34 ± 0.5°C.

Because the GABAA receptor binding was not assessed in the present study, we cannot rule out that Hy had a direct effect on the affinity and number of GABA receptors. This could explain the decrease in ventilation observed during Hy in NT and HT animals. An increase in the density of the GABA receptors (Bmax) has been described during anoxia or prolonged Hy in different animal species (26–28), but the changes in affinity of these receptors (KD) have not been consistent (26, 27). In the present study, the animals were exposed to Hy for only 10 min, and it is, therefore, unlikely that this possibility is the sole mechanism for the inhibitory effect of GABA on the central respiratory control.

Because metabolic rate is reduced during HT and Hy, the hypoxic ventilatory depression observed in hypothermic newborn piglets may in part be explained by this mechanism. Furthermore, a significant decrease in VO2 and CO2 production has been reported in dogs after ventriculocisternal perfusion with GABA agonists (29). Therefore, it is possible that the greater increase in MPO observed after BM in the HT group compared with NT animals was due to an increase in the metabolic rate. However, although VO2 was significantly decreased in HT animals compared with NT animals before the ICI of BM or PL, the changes in magnitude in VO2 with Hy were similar between NT and HT groups after BM, ruling out this possibility.

Previous papers have reported that anesthetics also may act at the GABAA receptor sites in the brain, enhancing GABAA receptor-mediated chloride currents to modulate either the release or uptake of GABA (21, 30, 31). Thus, it is possible that anesthesia may potentiate GABA effects and the administration of BM may reverse anesthesia-induced respiratory depression by a GABA-mediated mechanism. However, the anesthetics used in our study were urethane and chloralose, which have been reported to have no influence on the release or uptake of GABA in adult rats (32, 33). There is also a possibility that BM effects may be explained by changes in the sleep state of the animals. However, changes in EEG were not observed in either the NT or HT piglets after the drug administration, which makes it unlikely that the effect of BM on ventilation was due to behavioral changes.

Another possible explanation for the marked increase in MPO after BM infusion is a change in the cardiovascular response to Hy. It has been reported that intraventricular GABA administration to adult dogs resulted in a decrease in mean ABP and HR (12). Furthermore, BM applied to the surface of the ventrolateral medulla reverses the cardiovascular effect of GABA (18). In our study, however, changes in ABP and HR with Hy were not significantly different between NT-BM and HT-BM groups after BM infusion. These disparities of BM effects may be the result of different animal models, drug dosages, and age of the animals.

An increase in the production of lactic acid in the striatum area of adult rats has been reported during cerebral ischemia (34). It has been demonstrated that moderate lactic acidosis on the chemosensitive neurons on the surface of the ventrolateral medulla resulted in a slight stimulation of respiration during progressive brain Hy (35). In our study, however, no statistically significant difference in arterial pH or base excess during Hy was observed between NT and HT groups after BM infusion, suggesting that changes in acid-base status were not responsible for the increase in MPO after the ICI of BM in hypothermic newborn piglets. Although the changes of acid-base status in arterial blood may not accurately reflect those in brain stem tissues during HT, the shifts in calculated medullary cerebrospinal fluid pH were similar in direction but smaller in magnitude than those in arterial blood (36, 37). Therefore, the changes of acid-base status in arterial blood may at least exhibit the tendency of changes of acid-base status in medullary cerebrospinal fluid.

In summary, these data suggest that the depression in the hypoxic ventilatory response produced by HT in newborn piglets is in part due to an increased inhibitory effect of GABA in CNS and this effect is mediated through GABAA receptors.