Johnson, Stephen M., Rebecca A. Johnson, and Gordon S. Mitchell. Hypoxia, temperature, and pH/CO2 effects on respiratory discharge from a turtle brain stem preparation. J. Appl. Physiol. 84(2): 649–660, 1998.—An in vitro brain stem preparation from adult turtles (Chrysemys picta) was used to examine the effects of anoxia and increased temperature and pH/CO2on respiration-related motor output. At pH ∼7.45, hypoglossal (XII) nerve roots produced patterns of rhythmic bursts (peaks) of discharge (0.74 ± 0.07 peaks/min, 10.0 ± 0.6 s duration) that were quantitatively similar to literature reports of respiratory activity in conscious, vagotomized turtles. Respiratory discharge was stable for 6 h at 22°C; at 32°C, peak amplitude and frequency progressively and reversibly decreased with time. Two hours of hypoxia had no effect on respiratory discharge. Acutely increasing bath temperature from 22 to 32°C decreased episode and peak duration and increased peak frequency. Changes in pH/CO2increased peak frequency from zero at pH 8.00–8.10 to maxima of 0.81 ± 0.01 and 1.44 ± 0.02 peaks/min at 22°C (pH 7.32) and 32°C (pH 7.46), respectively; pH/CO2 sensitivity was similar at both temperatures. We conclude that1) insensitivity to hypoxia indicates that rhythmic discharge does not reflect gasping behavior,2) increased temperature alters respiratory discharge, and 3) central pH/CO2 sensitivity is unaffected by temperature in this preparation (i.e., Q10 ∼1.0).
in chelonia, respiratory motor output is modulated by sensory feedback from chemoreceptors located in the central nervous system (CNS) (15), the peripheral vasculature (1), and the lungs (25). With increased temperature, turtles (Pseudemys scripta and Chrysemys picta) generally increase their ventilation and have an increased sensitivity to changes in arterial or central pH/CO2 (9-12, 18, 19). Although studies with intact animals provide important information as to how overall ventilation and chemoresponsiveness are changed by temperature, it is difficult to determine how individual components of the respiratory control system are altered.
In this study we used an isolated in vitro brain stem preparation from adult turtles to examine the effects of anoxia and increased temperature on central respiratory neural discharge and central pH/CO2 chemosensitivity. In vitro brain stem preparations are well suited for this purpose, because peripheral chemo- and mechanosensory inputs as well as inputs from more rostral structures of the CNS are eliminated. In addition, the temperature and composition of the fluid bathing the tissue are easily controlled. Turtles are especially suited for in vitro studies, because the turtle brain is remarkably resistant to hypoxia (seediscussion). The in vitro turtle brain stem preparation has been used previously to show that an episodic pattern of respiratory motor output (i.e., periods of bursting activity interrupted by apnea) is intrinsic to brain stem structures; however, preliminary experiments indicated that this preparation exhibited inconsistent responses to alterations in pH/CO2 (5).
The specific objectives of the present study were1) to characterize respiratory motor discharge of the in vitro turtle preparation in greater detail under control conditions (22°C), 2) to examine the stability of respiratory discharge for 6 h at 22 and 32°C to determine the viability and usefulness of the preparation,3) to investigate the effects of bathing the brain stem preparation with hypoxic superfusate to test whether hypoxia modifies respiratory discharge by establishing a pathological respiratory pattern such as gasping,4) to determine how respiratory discharge is altered by acute increases in temperature from 22 to 32°C, 5) to determine whether the preparation exhibits pH/CO2responsiveness, and 6) to determine whether increased temperature changes pH/CO2 responsiveness of this preparation.
Adult turtles (C. picta) were obtained from local suppliers (Kons Scientific, Germantown, WI, and Lemberger, Oshkosh, WI) and kept in a large open tank, where they had access to water for swimming and heat lamps and dry areas for basking. The turtles had carapace lengths of 10–18 cm and weighed 0.45 ± 0.03 kg (range 0.17–0.84 kg).
Turtle Brain Stem Preparation
All surgical procedures and experimental protocols were approved by the Institutional Animal Care and Use Committee. Before the induction of anesthesia, pancuronium bromide (0.3–0.5 mg) was injected subcutaneously to paralyze the turtles, allowing an endotracheal tube to be placed without struggling. Shortly after an endotracheal tube was inserted into the trachea, the lungs were ventilated (10–15/min) with a mixture of halothane (4%) in O2 for 5–10 min. In nonparalyzed turtles used in other experiments (n = 15), this procedure eliminated corneal reflexes and the head withdrawal reflex within 1–2 min and reflexive limb withdrawal to foot pinch within 3–4 min. The plastron was rapidly removed with an autopsy saw, and the ascending aorta was perfused with ice-cold oxygenated (95% O2-5% CO2) superfusate (see below) for 2 min. After the animals were decapitated, the bone and muscle covering the brain stem dorsally were rapidly removed and the tissue was submerged in ice-cold oxygenated superfusate solution. After the brain stem was isolated, the tissue was trimmed with scissors, leaving intact the portion of the brain stem caudal to the optic lobes (corresponds roughly to the level 9 section in Fig. 16 of Ref. 3) and rostral to the spinal C1 roots (Fig.1 A). The tissue was pinned down on Sylgard (ventral surface upward) in an in vitro recording chamber (10-ml volume), where it was bathed with oxygenated superfusate (2.0–3.0 ml/min) at room temperature (hereafter referred to as 22°C). The pH of the solution in the chamber was constantly monitored with a calomel glass pH electrode (Digi-Sense, Cole-Parmer Instrument, Vernon Hills, IL). The tissue was allowed to recover and equilibrate for 60–90 min before initiation of a protocol.
To increase the temperature within the chamber, a Teflon-coated nichrome wire was wrapped inside the recording chamber and attached to a temperature controller (model TC-344A, Warner Instruments, Hamden, CT). The reservoir containing the superfusate was also heated to 32°C using a water bath. The composition of the superfusate was as follows (in mM): 100 NaCl, 23 NaHCO3, 10 glucose, 5N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES) sodium salt, 5 HEPES-free acid, 3 KCl, 2.5 CaCl2, and 2.5 MgCl2 (all chemicals were obtained from Sigma Chemical, St. Louis, MO). The concentration was based on an analysis of turtle cerebrospinal fluid (CSF) (14). HEPES was used in conjunction with to avoid precipitation of divalent cations in alkaline superfusates containing high concentrations. The pH of the superfusate when equilibrated with CO2 was temperature dependent (Table 1).
Nerve Recording and Analysis
To record respiratory activity, glass suction electrodes were attached to hypoglossal (XII) nerve roots (5). The signals were amplified (×10,000) and band-pass filtered (10 Hz–10 kHz) using a differential alternating-current amplifier (model 1700, A-M Systems, Everett, WA) before being recorded onto magnetic tape using a pulse-code modulation analog-to-digital converter (Vetter Instruments, Redersburg, PA). Ten to 12 episodes were analyzed for each condition. For preparations with a very low episode frequency, a minimum of three peaks per 30 min were required before being included in data analysis. After an experiment, signals from the tapes were rectified and integrated (time constant = 200 ms) using a moving averager (model MA-821/RSP, CWE, Ardmore, PA) before being digitized and analyzed using pClamp software (Axon Instruments, Foster City, CA).
Episode and peak variables were measured as shown in Fig. 1. Two or more peaks separated by less than the average duration of a peak within that recording were defined as an episode (26). Episode frequency was defined as 1/T epi, whereT epi is the time between the onset time of two sequential episodes. In cases where only single peaks were observed, the peaks were considered individual episodes with one peak per episode, and episode frequency was therefore equal to peak frequency (i.e., the time between the onset time of 2 sequential peaks; Fig. 1 B). Neural minute activity was defined as the integrated discharge (in arbitrary units) per minute. In cases where the discharge pattern was not clearly separated into distinct peaks (e.g., episode discharge in Fig.1 B), a set of criteria was established as follows: 1) a burst of neural activity in the XII nerve root was defined as a peak if its amplitude and area were >50 and >35%, respectively, of that for a typical peak in the same recording;2) within a prolonged set of bursts with multiple potential peaks (i.e., an episode), a burst had to meet the criteria stated above, and the onset and termination of the burst must be within 15% of the baseline before the burst was considered an individual peak; and 3) bursts not meeting the criteria for a peak were considered as extrarespiratory discharge.
To test preparation viability and stability over time, respiratory activity was continuously recorded for 6 h immediately after equilibration at 22 or 32°C. The gas bubbling the superfusate was maintained at 5% CO2 for experiments at 22°C and at 7–9% CO2 for experiments at 32°C to roughly match pH (∼7.45) at the two temperatures. Different preparations were tested at 22 or 32°C.
Preparations were equilibrated with 5% CO2-95% O2 at 22°C as described above, and control data were recorded. The superfusate bathing the preparation was rapidly switched to a reservoir equilibrated with 5% CO2-95% N2 at the same flow rate. Data were recorded for up to 2 h after the switch to hypoxia superfusate.
Rapid temperature change experiments.
Because it was found that the preparation did not consistently produce stable respiratory discharge for prolonged periods at 32°C, a more rapid temperature change protocol was used to study the effects of increased temperature within the same preparation. For the control period, the preparations were first bathed with superfusate equilibrated with 5% CO2 at 22°C, then the temperature of the solution in the chamber and the reservoir was increased to 32°C within 20–30 min. The preparation was allowed to reach steady state for 12–15 min before data were collected. Because the pH of the superfusate increased at 32°C (Table 1), the superfusate was then equilibrated with 8–9% CO2 at 32°C to match the pH to that during the control period (i.e., 7.39). The preparations were again allowed to reach steady state for 12–15 min, and data were collected. Finally, control conditions were restored (5% CO2, 22°C), and data were collected 12–15 min later.
pH/CO2 sensitivity experiments.
After equilibration of the preparations at 22 or 32°C (5% CO2) as described above, the superfusate was rapidly switched to a reservoir equilibrated with gas of a different percent CO2. Care was taken to ensure that flow rates of the solutions into the chamber remained constant. The percent CO2in the O2-CO2gas mixture used to aerate the reservoirs was determined using a gas-mixing pump (Wösthoff, Calibrated Instruments, Hawthorne, NY) or rotameters analyzed with an infrared CO2 analyzer (model LB-2, Beckman Instruments, Schiller Park, IL). The pH in the chamber was monitored to determine when a new steady state had been reached (usually after 12–15 min). The preparation was allowed to reach a steady state for 12–15 min before data were collected. The sequence of CO2 changes was 1-3-5-7-9% CO2. A few preparations were tested with a reverse sequence, but there were no differences in the results, and thus data were pooled.
By use of separate brain stem preparations, a set of two-point pH/CO2 sensitivity experiments were performed to obtain data before the potential “rundown” of the preparation at 32°C (Figs.2 B,3, and 4). After equilibration at 22°C, the preparations were heated to 32°C with the superfusate equilibrated with 2.2% CO2. Once a new steady-state pH was established, data were collected (20–30 min), and CO2 levels in the superfusate were switched to 5.3% (CO2 levels of 2.2 and 5.3% were chosen, because the pH values fell within the steep portion of the pH/CO2 response curve). After the attainment of a new steady-state pH (15–20 min) and equilibration, data were collected (10–20 min). The total time of the experiment at 32°C was 96 ± 5 min (range 80–114 min).
Values are means ± SE. Statistical inferences were made using one- or two-way analysis of variance with a repeated measures design; Student’s t-tests with the Bonferroni correction for multiple comparisons were used for pairwise comparisons (Sigma Stat, Jandel Scientific Software, San Rafael, CA). Measurements in arbitrary units (e.g., episode integral, neural minute activity, and peak amplitude) were normalized to initial control values for the viability, rapid temperature change, and hypoxia experiments. For the CO2 sensitivity experiments, arbitrary unit measurements were normalized to values at 5% CO2 at 22°C and 7% CO2 at 32°C, because the pH was identical (∼7.45) at these two CO2 values.
Isolated in vitro turtle brain stems produced a rhythmic motor output that consisted of large-amplitude bursts (i.e., peaks) of neural discharge. The peaks of discharge occurred as singlets (single peaks only), doublets (2 peaks together), or episodes (several peaks grouped together; Fig. 1 B). In superfusate equilibrated with 5% CO2 at 22°C, 64% of the brain stem preparations (n = 25) produced singlets with occasional doublets; the other 36% produced doublets or episodic discharge. Peak frequency was 0.74 ± 0.07/min (range 0.24–1.80/min) with an average of 1.7 ± 0.2 peaks/episode (range 1.0–4.5 peaks/episode); peak duration was 10.0 ± 0.6 s (range 5.4–18.1 s). Episode frequency was 0.53 ± 0.05/min (range 0.06–1.25/min), and episode duration was 30.2 ± 3.8 s (range 12.0–94.0 s). In episodic discharge that contained more than one peak, the interval between peaks was 17.0 ± 1.4 s (range 6.8–29.6 s, n = 15). The pattern of discharge within a burst was highly variable, with the most consistent pattern being triangular shaped, with the peak rise time (time from onset of peak to peak maximum 4.6 ± 0.3 s, range 1.7–7.9 s) roughly equal to one-half the peak duration (Fig.1 C).
Viability Experiments: Preparations Stable at 22°C but Not at 32°C
At 22°C, the brain stem preparations (n = 6) produced neural discharge that was relatively stable (Fig. 2 A), with nearly all the episode and peak measurements unchanged for 6 h (Figs. 3 and 4). Episode frequency, episode duration, and episode integral did not change with time (Fig. 3,A–C), whereas the neural minute activity decreased significantly (by 27%), but only after 6 h (Fig.3 D). Likewise, the peak frequency, peaks per episode, rise time for peak amplitude, peak duration, and peak amplitude were not significantly altered for 6 h (Fig. 4). At 32°C, however, the brain stem preparations (n = 6) produced two general types of discharge patterns: the amplitude was maintained with a moderate decrease in episode and peak frequencies (Fig.2 B, n= 3 of 6) or amplitude and frequency slowly decreased, sometimes being abolished by 4 h (Fig. 2 C,n = 3 of 6). The decrease in XII nerve motor output did not mean that the preparation had died, because a return to 22°C always restored respiratory discharge (Fig.2 C, recovery trace). When data from all 32°C experiments were pooled, all variable values declined to varying degrees. At 6 h, episode duration and episode integral decreased gradually with time to 39 and 43% of control values (Fig. 3, B andC), respectively. Episode frequency and neural minute activity decreased relatively rapidly, reaching values significantly below control after only 2–4 h and being nearly abolished at 6 h (Fig. 3, A andD). The number of peaks per episode, rise time for peak amplitude, and peak duration decreased slightly over time to 80–90% of control values (Fig. 4,B–D). Peak frequency decreased relatively rapidly in a manner similar to the decrease in episode frequency, with the decrease reaching statistical significance after 4 h (Fig. 4 A). Peak amplitude decreased by 40% at 6 h, but the large variability prevented this value from reaching statistical significance (Fig.4 E). Two-way analysis of variance of the control curves for 22 and 32°C showed significant time and temperature interactions for the episode frequency, neural minute activity, peak frequency, and peak amplitude data (allP < 0.05). The episode frequency, episode duration, episode integral, and peak amplitude data had significant temperature and time effects (Figs. 3 and 4).
Hypoxia Resistance of Respiratory Discharge
Turtle brain stem preparations (n = 6) continued to produce a robust respiratory motor discharge pattern, even after 2 h of superfusion with superfusate equilibrated with 5% CO2-95% N2 at 22°C (Fig.5). Baseline peak frequency was 0.73 ± 0.22 and 0.71 ± 0.15 peak/min after 2 h of hypoxia. Similarly, neural minute activity and peak amplitude were 1.25 ± 0.33 and 0.97 ± 0.13, respectively, relative to baseline after 2 h of hypoxia. On several occasions, brain stem preparations that were stored in 50 ml of superfusate for 7–14 days at 1–4°C still produced respiratory discharge after being placed in the recording chamber and rewarmed to 22°C (data not shown).
Rapid Temperature Change Experiments
For these experiments (n = 6), increasing the temperature of the superfusate (5% CO2) by 10°C (from 22.2 ± 0.3 to 31.9 ± 0.2°C) caused the pH of the superfusate to increase from 7.39 ± 0.02 to 7.59 ± 0.01 (uncompensated pH). By increasing the percent CO2 in the superfusate to 8–9%, the pH was returned to its original level of 7.39 ± 0.01 (compensated pH). On return to the lower temperature (22.6 ± 0.3°C) and 5% CO2, the pH of the superfusate was 7.35 ± 0.01. During these temperature and pH/CO2 maneuvers, two general types of changes in the respiratory motor discharge pattern were observed that could be attributed to increased temperature at constant pH (Fig. 6). The first change was an increase in episode frequency (53%, althoughP > 0.05) and peak frequency (77%) at high temperature and compensated pH, but not at the uncompensated pH (Fig. 7,A andE). The second change was a 30–40% decrease in episode duration (Fig.7 B) and episode integral (Fig.7 C) that was paralleled by a 30–37% decrease in peak duration and rise time (Table2). The net result was a 37% decrease in neural minute activity at the uncompensated pH followed by a return to baseline levels at the compensated pH (Fig.7 D). Increased temperature did not change the number of peaks per episode (Fig.7 F) or peak amplitude (Table 2).
pH/CO2 Sensitivity Experiments: Does Higher Temperature Increase Sensitivity?
When CO2 levels (and pH; Table 1) of the superfusate were altered sequentially, respiratory discharge was altered in a similar manner at 22°C (n = 7) and 32°C (n = 9). In general, the frequency of episodes and peaks increased with increasing CO2/H+concentration from zero to maximum values at pH 8.00–7.46. For example, at 1.0% CO2 (pH 8.00–8.10), there was no discharge in any of the preparations at 22 or 32°C (Fig. 8). At 3% CO2 and 32°C (pH 7.76 ± 0.01), only two of nine preparations were active; at 22°C (pH 7.60 ± 0.03), all preparations produced a motor output. Within 3–5% CO2 (pH 7.46–7.60), episode and peak frequency increased with increasing CO2 levels. No further increases in episode and peak frequency occurred at CO2 >5% (pH 7.46–7.23).
Of the measured variables, episode frequency, episode integral, neural minute activity, and peak frequency were sensitive to changing pH/CO2. The curves for episode frequency at 22 and 32°C were comparable in shape, with a rapid rise from zero (pH >8.00) to maximal levels of 0.47 ± 0.08 and 0.64 ± 0.12 episodes/min at 22°C (5% CO2) and 32°C (7% CO2), respectively (pH 7.46 for both); episode frequency decreased by up to 25% at pH <7.46 (Figs.9 A and10 A). At 22°C, episode integral increased during hypercapnia to a maximum at 7% CO2, whereas the relationship was relatively flat or decreased with increased CO2 at 32°C, resulting in a significant percent CO2-temperature interaction (Fig.9 C). Neural minute activity increased from zero (pH >8.00) to a plateau at pH <7.46 at both temperatures (Fig. 9 D). Likewise, peak frequency increased from 0.00 (pH >8.00) to 0.81 ± 0.15 and 1.43 ± 0.33 peaks/min at 22°C (7% CO2) and 32°C (5% CO2), respectively (Figs.9 A and10 B). Episode duration, peak duration, and peak rise time were relatively unaffected by changes in pH/CO2 at 22 or 32°C but were decreased at 32°C as previously shown in the rapid temperature change experiments (Fig. 9 B, Table 2). Peak amplitude was not altered by changes in pH/CO2 at either temperature (Table 2).
From the two-point pH/CO2experiments at 32°C (n = 7), only the episode and peak frequency data were analyzed (Fig.10). Episode and peak frequency increased from 0.35 ± 0.09 and 0.47 ± 0.14/min at 2.2% CO2 (pH 7.92 ± 0.02) to 0.88 ± 0.24 and 1.00 ± 0.23/min at 5.3% CO2 (pH 7.61 ± 0.02), respectively. All seven preparations at 32°C produced motor discharge at both CO2 levels. The slopes of the responses from the two-point pH/CO2 experiments at 32°C were calculated and compared with the slopes of the responses at 22°C with 3 and 5% CO2 (pH 7.60 ± 0.03 and 7.46 ± 0.02). The episode frequency slope at 22°C (−1.07 ± 0.53 episodes ⋅ min−1 ⋅ pH unit−1) was not significantly different from that at 32°C (−1.76 ± 0.60 episodes ⋅ min−1 ⋅ pH unit−1; Fig.10 A). Also, the peak frequency slopes were nearly identical at both temperatures, with slopes of −1.81 ± 0.94 and −1.77 ± 0.78 peaks ⋅ min−1 ⋅ pH unit−1 at 22 and 32°C, respectively (Fig. 10 B). Thus temperature had no significant effects on the pH/CO2 sensitivity of respiratory motor output in this preparation.
The goal of this study was to characterize the pattern of respiratory motor output produced by an in vitro turtle brain stem preparation and determine how hypoxia, increased temperature, or changes in pH/CO2 altered that pattern. Turtle brain stem preparations produced a stable respiratory motor pattern for 6 h at 22°C under control (possibly hyperoxic) conditions and for ≥2 h under anoxic conditions, demonstrating the hardiness and viability of the preparation. Increasing the bath temperature to 32°C at constant pH reversibly increased peak frequency and decreased episode duration, peak rise time, and peak duration, demonstrating that the respiratory rhythm and pattern were sensitive to temperature changes. Central pH/CO2 chemosensitivity was intact in turtle brain stem preparations as peak frequency, episode frequency, and episode integral rate increased with decreasing pH within the pH range of 8.10–7.40. Central pH/CO2 chemosensitivity, however, was not sensitive to temperature changes, because the slopes of the peak frequency responses were similar.
Advantages and Limitations of the Turtle Brain Stem Preparation
The advantages and limitations of using reduced in vitro preparations for the study of respiratory neural control are well known (27). Turtle brain stem preparations, however, have the distinct advantages of1) being derived from a fully mature (as opposed to embryonic or neonatal) vertebrate that aspirates air (like mammals), 2) having low metabolic rates at the animal’s normal temperature, thereby minimizing tissue , , and pH gradients, and3) being studied within a physiological temperature range (cooling the tissue below the normal body temperature as is commonly done with mammalian tissue is not necessary to preserve viability).
Does the pattern of respiratory discharge produced by turtle brain stem preparations at 22°C resemble that produced in intact turtles at room temperature? In intact turtles, phasic expiratory and inspiratory spinal motoneuron activity can be recorded from the appropriate expiratory and inspiratory motor nerves during each breath (33). Respiratory cranial motor nerve root activity, however, consists of only a single burst of neural discharge for each breath (5); thus one can compare the breaths from intact vagotomized turtles with the “peaks” from turtle brain stem preparations (Table 3) (26). These comparison data indicate that turtle brain stem preparations had episode and peak values that were close to episode and breath frequency values in vivo. The variances of measurements from in vitro and in vivo turtles were also similar. On the other hand, episode and peak durations in vitro were 50–80% larger than episode and breath durations in vivo. The reason for differences in episode and breath (peak) durations may be a lack of descending inputs from higher centers in the CNS or from a complete loss of peripheral sensory afferent inputs. On the whole, there is a reasonable similarity of breathing patterns between the in vitro brain stem preparation and awake resting turtles (26).
XII nerve activity: an index of respiratory motor output?
Although cranial motor outflow and, specifically, XII nerve activity often exhibit respiratory-related, rhythmic activity that is tightly linked with breathing in mammals and turtles (5), the precise function of the XII nerve in turtles is not known. It is possible that XII nerve activity during inspiration prevents the tongue from blocking airflow while turtles are breathing through the nares. Accordingly, peak and episode frequencies and the number of peaks per episode are the only faithfully reflected variables of respiratory activity that can be drawn from XII nerve root activity.
Episodic breathing of more than one breath is not consistently observed in vitro. The isolated in vitro turtle brain stem preparation was previously shown in some cases to produce episodic breathing in a manner similar to that in the intact turtle (5). This earlier study confirmed that episodic breathing in turtles is an intrinsic property of the central respiratory control system and does not require peripheral chemoafferent feedback. In the present study we showed that roughly one-third of the isolated turtle brain stem preparations produced episodic discharge, whereas most preparations produce single bursts. It is hypothesized that all vertebrate respiratory control systems are capable of expressing episodic discharge, depending on the species and physiological conditions (24). The factor(s) that determines whether the respiratory discharge pattern is episodic is not known, although a relationship between respiratory drive and the number of breaths in an episode has been postulated (24). Factors that may play a role in this preparation include vagotomy (which reduces ventilatory responses to CO2), the specific site of the rostral transection, and other uncontrolled factors. Nevertheless, the similar mean values and variances characterizing ventilatory pattern in vivo vs. in vitro suggest that the real pattern differences between these preparations are minor.
Respiratory Neural Discharge In Vitro: Transformed Eupneic Pattern or Gasping?
The present controversy as to whether neonatal rodent in vitro brain stem (spinal cord) preparations are producing eupneic breathing or gasping will remain unresolved until the cellular and network mechanisms underlying eupnea and gasping are fully understood. In mammals, gasping is a last-chance attempt at autorescusitation during severe hypoxia: breathing frequency decreases, phrenic burst amplitude increases, and phrenic discharge patterns within bursts shift from a progressively augmenting-rapidly decrementing pattern to a rapid onset-slowly decrementing pattern (reviewed in Ref. 32). Some investigators hypothesize that neonatal mammalian in vitro brain stem preparations (and other in vitro preparations) are gasping because of the low frequency of respiratory-related neural discharge, the rapid onset-slowly decrementing pattern of inspiratory discharge, and the lack of expiratory phase-related discharge (32). Consistent with this hypothesis, the center of the neonatal rat brain stem preparation is hypoxic with high extracellular concentrations of H+ and K+ (2, 29). Others argue that the isolated in vitro respiratory neural circuit undergoes a transformation due to vagotomy and reduced temperature, such that the pattern of discharge is “gasplike” in some respects, but that the output is still produced by the neural circuitry that produces eupnea under normoxic conditions (2, 8). Consistent with this view, thin medullary slices from neonatal rats have rhythmic discharge patterns similar to those produced by intact brain stems in vitro (31), yet the slices probably are neither hypoxic nor acidic (21).
The turtle brain stem preparation makes an interesting contribution to this debate, because turtles can survive for several hours or days without O2, depending on the temperature [e.g., C. pictasurvives a forced submersion for 2 days at 26°C (28)]. Calculations based on the relatively low metabolic and O2 consumption rates of aquatic ectothermic vertebrates suggest that isolated brain tissue from these animals will not be hypoxic under in vitro conditions. In support of this hypothesis, the center of an in vitro tadpole brain stem preparation was recently found to have a of ∼240 Torr at a depth of 850–900 mm by using superfusate equilibrated with 90% O2 at 23°C (29). The center of an in vitro turtle brain stem is also likely to be hyperoxic, because the diffusion distance to the center of the brain stem is similar (turtle brain stems have a U-shaped geometry that puts the center at ∼1.0 mm from the outer surface). Thus it is unlikely that there is a hypoxic center in the turtle brain stem preparation analogous to that observed in neonatal rat brain stem preparations.
Furthermore, respiratory discharge in the turtle brain stem preparation was not sensitive to severely hypoxic conditions, in which the of the hypoxic superfusate was probably <50 Torr and the tissue may have been anoxic. In similar in vitro experiments using neonatal rat preparations, the of hypoxic superfusate (bubbled with 5% CO2-95% N2) was 5–30 Torr (2, 35). For the in vitro tadpole brain stem preparation, decreases by 60 Torr/100 μm, starting at 200 μm above the surface until 200 μm below the surface, at which point further decreases in are dependent on the position of the electrode tract (29). If one assumes that the in vitro turtle brain stem has a metabolic rate similar to that in the tadpole, it is highly likely that most of the turtle brain stem was anoxic during our hypoxia experiments, yet there were no alterations in respiratory discharge. In contrast, neonatal rat brain stem-spinal cord preparations undergo transient excitation followed by prolonged depression of respiratory discharge within minutes of exposure to hypoxic superfusate (2, 35). The turtle’s remarkable resistance to hypoxia is further demonstrated by the fact that intracellular recordings of turtle cortical neurons in vitro show no change in membrane potential, input resistance, or excitability after 180 min of anoxia (4). Taken together, these findings support the idea that the rhythmic motor output produced by the turtle brain stem preparation is respiratory related and not due to the establishment of some other pathological rhythm.
Changes in Breathing With Increased Temperature in Turtles
Early reports concerning the effects of increased temperature on turtle ventilation are controversial perhaps because of the use of different species and experimental approaches (reviewed in Refs. 10 and 12). More recent studies on unanesthetized, freely movingChrysemys, however, are in agreement that ventilation increases with increased temperature. The increased ventilation between 10 and 20°C is due primarily to increased frequency, whereas the increased ventilation between 20 and 30°C is due primarily to increased tidal volume (10, 12). To our knowledge, the present study is the first to examine temperature effects on respiratory motor discharge from the deafferented turtle nervous system, a procedure that eliminates many peripheral afferent inputs that are relevant to breathing. The three major effects of increased temperature on respiratory discharge from the turtle brain stem preparation are discussed below.
Stability of respiratory discharge at 22 and 32°C.
For the turtle brain stem preparation at 22°C, all measured respiratory discharge variables were stable for ≥6 h. After >2 h at 32°C, however, decreases in neural minute activity were observed (due to decreased peak frequency and amplitude). The decreased neural minute activity at 32°C could be due to increased activity of inhibitory neurons or a buildup of inhibitory neurochemicals within the brain stem. It is unlikely that the tissue was deteriorating, because respiratory discharge was rapidly restored on return to 22°C in all cases. Other investigators have studied ventilation in awake, intactP. scripta at 33–35°C without reporting any deleterious effects (9, 18), although the relevance of these findings to in vitro brain stem preparations is uncertain.
Increased peak frequency with increased temperature.
In our rapid temperature alteration and pH/CO2 sensitivity experiments, peak frequency increased by 72–75% (Figs.7 E and9 E). Thus acute exposures to increased temperature increased peak frequency in vitro. In contrast, intact turtles have no change (or a slight decrease) in breath frequency with temperature increases from 20 to 30°C (10,12), as in vagotomized snakes (13). In spontaneously ventilating vagotomized alligators (Alligator mississippiensis), breath frequency increases slightly, but not significantly, from 1.9 breaths/min at 20°C to 2.6 breaths/min at 30°C (6). The reason for the discrepancy with intact turtles is likely due to the absence of peripheral sensory inputs and rostral CNS structures in the turtle brain stem preparation. The lack of an increased breath frequency in vagotomized snakes and alligators, however, may be due to species differences, intact rostral structures, or differences in brain stem pH/CO2 levels (e.g., brain stem CO2 levels may have been higher in the turtle brain stem preparation).
Decreased peak duration with increased temperature.
In this study, peak (and episode) duration decreased with increasing temperature. To our knowledge, only two studies report similar changes in breath duration in reptiles with increased temperature. In restrained, unanesthetized P. scripta, breath duration decreased from 9.1 ± 0.7 to 7.7 ± 0.7 s after a temperature increase from 28 to 33–34°C (9). Similarly, in vagotomized water snakes tested at 15 and 30°C, the inspiratory time decreased from 6.3 to 3.5 s, whereas there was no change in inspiratory time in snakes with intact vagi (13). The decreased peak duration in the turtle brain stem preparation may be due to temperature-induced changes only at the level of XII motoneurons, or changes may additionally be occurring presynaptic to XII motoneurons (e.g., within premotoneurons, respiratory rhythm-generating neurons, or modulatory neurons).
Central pH/CO2 Sensitivity in Turtles
Physiological range of arterial and CSF pH in turtles.
In this study, turtle brain stems were tested in superfusates with pH of 7.21–8.10, although the center of the brain stems probably had a lower pH. For example, the center of the tadpole brain stem was as much as 0.5 pH unit lower than the outer bath (29). Even if the center of turtle brain stems is similarly acidic, most of the neuronal structures within the brain stem would be exposed to pH levels that fall within the range of normal arterial and CSF pH values reported forChrysemys andPseudemys at different temperatures or under different conditions. Arterial pH in these species is 7.85–7.96 at 10°C, 7.74–7.78 at 20°C, and 7.58–7.63 at 30°C (11, 15, 20). The pH of CSF inPseudemys is reported to be relatively more acidic than arterial pH: 7.79 at 10°C, 7.64 at 20°C, and 7.49 at 30°C (15). In addition, arterial pH was measured at 6.809 in Pseudemys after a 2- to 4-h dive (20), demonstrating the enormous range of arterial pH tolerated by these semiaquatic turtles.
Temperature effects on central pH/CO2sensitivity in reptiles.
The increase in ventilation with hypercapnia in turtles and tortoises varies from mild to robust (1, 9, 10, 16, 17, 19, 25, 26, 30). The increased ventilation is due to an increase in breath frequency and tidal volume along with a decrease in the periods of breath holding. Interestingly, the increase in breath frequency with hypercapnia is severely attenuated with vagotomy in turtles (Chrysemys) (26), tortoises (Testudo) (1), and alligators (6). In the present study the turtle brain stem preparation mildly increased peak frequency with increasing acidification at a rate comparable to the increases obtained in vagotomized turtles and tortoises at room temperature (1, 26). Indeed, it is likely that larger increases in peak frequency at a lower pH would be obtained with stimulation of vagal afferent inputs to the turtle brain stem preparation. This hypothesis is supported by the finding that stimulation of vagal afferents in an in vitro bullfrog preparation augments the increase in respiratory burst frequency to CO2/H+(23).
In the turtle brain stem preparation at 22 and 32°C, episode and peak frequency as well as neural minute activity were pH/CO2 sensitive; all other variables were remarkably insensitive. Data from the pH/CO2 experiments conducted at 32°C must be interpreted cautiously, even though there was no evidence for a consistent time-dependent rundown in peak amplitude and frequency after 3–4 h (as was observed in the viability experiments). It is possible that none of the preparations used in the pH/CO2 experiments at 32°C underwent rundown. Alternatively, the protocol for the pH/CO2 experiments at 32°C may have masked rundown, because the sequential decrease in pH during the pH/CO2 experiments may have increased respiratory “drive.” To avoid the problem, rapid two-point pH/CO2 experiments were performed within 90–120 min at 32°C to minimize the effects of rundown. Because the episode and peak frequency vs. pH slopes from the two-point experiments were not larger than those from the five-point experiments, it is unlikely that increased respiratory drive was masked by a decline in respiratory discharge due to time-dependent factors.
Collectively, these data suggest that central chemoreceptors in the turtle brain stem preparation are functional at both temperatures studied and act to regulate the number of peaks (breaths) per minute. The Q10 for pH/CO2 sensitivity in the turtle brain stem preparation was ∼1.0, because neither the episode nor the peak frequency vs. pH slopes were significantly affected by temperature. Although intact turtles and alligators have a Q10 for CO2 sensitivity of ∼2.0 (6, 10), it is striking to note that the Q10 for CO2 sensitivity in vagotomized alligators is also ∼1.0 (6). It has been hypothesized that the Q10 for central CO2 sensitivity (or the central integrator of chemosensory inputs) in reptiles is <1.0 to compensate for the large temperature dependence of vagal CO2-sensitive receptors, which may have Q10 values of ∼3.0 or more (7). How the brain stem (or CNS) modifies the chemosensory input at different temperatures in ectothermic vertebrates is an intriguing problem about which little is known. The development of in vitro preparations such as the turtle brain stem preparation may facilitate a resolution of these issues.
The authors thank Dr. R. Kinkead for helpful comments in the preparation of the manuscript and Dr. K. B. Bach for the drawing in Fig. 1 A.
Address for reprint requests: S. M. Johnson, Dept. of Comparative Biosciences, School of Veterinary Medicine, University of Wisconsin, 2015 Linden Dr. West, Madison, WI 53706.
This work was supported by National Heart, Lung, and Blood Institute Grants HL-53319 and HL-36780. S. M. Johnson is a Parker B. Francis Fellow in Pulmonary Research.
Preliminary reports of this work have appeared in abstract form (22).
- Copyright © 1998 the American Physiological Society