After occurrence of spinal cord injury, it is not known whether the respiratory rhythm generator undergoes plasticity to compensate for respiratory insufficiency. To test this hypothesis, respiratory variables were measured in adult semiaquatic turtles using a pneumotachograph attached to a breathing chamber on a water-filled tank. Turtles breathed room air (2 h) before being challenged with two consecutive 2-h bouts of hypercapnia (2 and 6% CO2 or 4 and 8% CO2). Turtles were spinalized at dorsal segments D8–D10 so that only pectoral girdle movement was used for breathing. Measurements were repeated at 4 and 8 wk postinjury. For turtles breathing room air, breathing frequency, tidal volume, and ventilation were not altered by spinalization; single-breath (singlet) frequency increased sevenfold. Spinalized turtles breathing 6–8% CO2 had lower ventilation due to decreased frequency and tidal volume, episodic breathing (breaths/episode) was reduced, and singlet breathing was increased sevenfold. Respiratory variables in sham-operated turtles were unaltered by surgery. Isolated brain stems from control, spinalized, and sham turtles produced similar respiratory motor output and responded the same to increased bath pH. Thus spinalized turtles compensated for pelvic girdle loss while breathing room air but were unable to compensate during hypercapnic challenges. Because isolated brain stems from control and spinalized turtles had similar respiratory motor output and chemosensitivity, breathing changes in spinalized turtles in vivo were probably not due to plasticity within the respiratory rhythm generator. Instead, caudal spinal cord damage probably disrupts spinobulbar pathways that are necessary for normal breathing.
- control of breathing
- episodic breathing
in response to injury, the central nervous system undergoes morphological and physiological changes to compensate for loss of function, such as altered gene expression, neuronal reorganization, and altered synaptic plasticity. Likewise, after incomplete spinal cord injury, remaining descending spinal pathways and spinal motor networks undergo reorganization and neuroplasticity to compensate for loss of function (6, 20, 51). Because respiratory dysfunction is the leading cause of death in spinal cord-injured patients (9, 59), understanding the mechanisms by which the respiratory control system undergoes compensatory plasticity may lead to novel therapeutic strategies for reducing respiratory complications and subsequent morbidity.
After spinal cord injury, breathing pattern is altered to maintain adequate ventilation (V̇e). For example, high cervical spinal cord injury compromises synaptic drive to phrenic and intercostal motoneurons in mammals, resulting in decreased tidal volume (Vt). To compensate, breathing frequency increases in humans to maintain blood-gas homeostasis (40, 41). Likewise, rats with cervical or thoracic spinal cord injuries have decreased Vt and increase the frequency of normal and augmented breaths (19, 21, 56). Injury-induced changes in breathing pattern may be due to alterations in lung and chest wall mechanical properties since intact vagal nerves are required (19). Another compensatory mechanism is augmentation of excitatory drive to accessory muscles (11) and remaining functional spinal motoneurons (8, 17, 20). However, the mechanisms underlying these compensatory changes in respiratory control are not well understood.
To address this question, a model of spinal cord injury in an aspirating adult ectothermic vertebrate was developed to test whether compensatory breathing changes are due to long-lasting changes within the respiratory rhythm generator. Turtles were chosen because breathing is easily measured in awake, freely swimming adult animals, and their breathing can be disrupted by spinalization. Chelonians, such as semiaquatic turtles (Pseudemys scripta), do not have a diaphragm; instead, they breathe by moving their soft pectoral girdle (neck and forelimbs) and pelvic girdle (tail and hindlimbs) inward and outward like a bellows to generate positive (expiratory) and negative (inspiratory) lung pressures (15, 16, 43). A complete spinal transection at segments D8–D10 (just rostral to the lumbar spinal cord) eliminates use of the pelvic girdle [the main respiratory pump (15, 16, 43)] and forces the turtles to use only their pectoral girdle for breathing. In addition, brain stem (and brain stem-spinal cord) preparations from adult turtles can be maintained in vitro for several hours, thereby permitting studies on spontaneous respiratory motor output under more controlled conditions (28, 57). In spinalized turtles, we tested the hypotheses that 1) decreased Vt is compensated by increased breathing frequency, 2) hypercapnic responses are blunted due to a compromised ability to increase Vt, 3) spontaneous recovery occurs within 2 mo, and 4) breathing pattern changes observed in vivo persist in isolated in vitro brain stems, thereby representing a form of supraspinal neuroplasticity (18). This is the first study to address such hypotheses in an aspirating ectothermic vertebrate that breathes episodically. A preliminary report of this work was published in abstract form (4).
MATERIALS AND METHODS
All procedures were approved by the Animal Care and Use Committee at the University of Wisconsin-Madison School of Veterinary Medicine. Adult red-eared slider turtles (Pseudemys, n = 51) were obtained from commercial suppliers and kept in a large open tank where they had access to water for swimming and heat lamps and dry areas for basking. Room temperature was set at 27–28°C [close to their preferred temperature range (3)], with 14 h/day of light. Turtles were fed ReptoMin floating food sticks (Tetra, Blacksburg, VA) three to four times per week.
Turtles (n = 7; 822 ± 19 g) were preoperatively given antibiotic (5 mg/kg sc enrofloxacin), anti-inflammatory (2–4 mg/kg im ketoprofen), and analgesic drugs (0.05–1.0 mg/kg im buprenorphine). Turtles were then intubated and ventilated with 5% isoflurane (balance = 100% O2) until the forelimb withdrawal reflex to noxious foot pinch was eliminated. Anesthetized turtles were placed in ice water for 20 min to reduce blood pressure and minimize bleeding during surgery. A drill was used to remove a 1.5 × 1.0-cm area of carapace dorsal to spinal segments D8–D10. A laminectomy was performed to expose and remove a 0.5- to 1.0-cm length of spinal cord. A 1.0 × 1.0-cm square of Surgicel was placed in the spinal column to control bleeding and protect the spinal cord from rapid-setting epoxy glue, which was used to form a watertight seal over the carapace opening. Sham turtles (n = 3; 849 ± 25 g) underwent the same procedure without spinal cord removal; Surgicel was placed dorsal to, but not touching, the dorsal surface of the spinal cord before the site was sealed with epoxy glue. All operated turtles were placed in shallow warm water (30–35°C) to stimulate breathing and allow them to eliminate isoflurane and restore blood-gas homeostasis. Postoperative care consisted of administering antibiotic (5 mg/kg sc enrofloxacin daily for 4 days), anti-inflammatory (2–4 mg/kg im ketoprofen daily for 4 days), and analgesic drugs (0.05–1.0 mg/kg im buprenorphine given once 2 days postsurgery). Spinalized and sham turtles were allowed to recover in their normal holding tanks for 8 wk. During the recovery period, spinalized and sham turtles gained 43 ± 10 and 13 ± 2 g (P = 0.103), respectively.
Turtle V̇e Measurements
V̇e in awake, freely swimming turtles was measured using established methods (13, 14). Turtles were placed in a black opaque container (20 × 30 × 40 cm) that was filled to the top with water at room temperature (Fig. 1). A circle (diameter = 11.5 cm) was cut in the top, and a translucent plastic container (volume = 250 ml) was inverted and placed on the cut hole to make an air-tight seal. This inverted plastic “breathing” chamber provided the only location within the tank for the turtles to breathe. Flowmeters were used to maintain gas flow into the chamber at ∼500 ml/min. A pneumotachograph (Godart, Gould Electronics, Eastlake, OH) attached to the breathing chamber exit hole measured air flow. The pneumotachograph was calibrated by connecting a 25-ml glass syringe to the breathing chamber with plastic tubing. An electric motor was attached to the syringe plunger so that the plunger could be moved rhythmically back and forth at cycle periods of 1.5, 2.5, and 3.5 s, which is similar to the duration range for one turtle expiratory-inspiratory cycle. For each of these cycle periods, the syringe was set at several different volumes (range = 3.5–26.3 ml) and the motor was used to move the plunger for 10–15 cycles to generate simulated breaths. The amplitude and area of recorded signals were analyzed and plotted versus syringe volume. Because the three groups of calibration data were linear and had similar slopes, all data for the simulated breaths were pooled to calculate one calibration line that was used to convert signal amplitude and area from real turtle breaths to volume.
In Vivo Experimental Protocols
Two protocols were applied to the same turtles during the week before surgery and at 4 and 8 wk after surgery. The first protocol sequentially exposed turtles to room air (2 h), room air plus 2% CO2 (2 h), room air plus 6% CO2 (2 h), and room air (2 h), whereas the second protocol exposed turtles to room air (2 h), room air plus 4% CO2 (2 h), room air plus 8% CO2 (2 h), and room air (2 h). Protocols were applied at least 24–48 h apart to allow complete recovery. Electrical signals from the pneumotachograph were saved to computer using a data acquisition system (LabPro, Vernier Software and Technology, Beaverton, OR) and analyzed offline using Clampfit software (Axon Instruments, Union City, CA).
In Vitro Turtle Brain Stem Preparations
Brain stems were isolated from adult turtles (n = 53, 754 ± 23 g) as described previously (27). Approximately 60 days after surgery, control, sham, and spinalized turtles were intubated and anesthetized with 5% isoflurane (balance O2) until limb withdrawal to noxious foot pinch was eliminated. Turtles were rapidly decapitated and decerebrated. Brain stems were removed and pinned down in a recording chamber (13 ml volume) with the ventral surface facing upward (see Fig. 7A). Brain stems were superfused (4–6 ml/min) with solution (23–24°C) containing HEPES buffer as follows (in mM): 100 NaCl, 23 NaHCO3, 10 glucose, 5 HEPES (sodium salt), 5 HEPES (free acid), 2.5 CaCl2, 2.5 MgCl2, 1.0 K2PO4, and 1.0 KCl (bubbled with 5% CO2-95% O2). Reservoir pH was 7.38 ± 0.01 as measured periodically with a calomel glass pH electrode (Digi-Sense; Cole-Parmer Instruments, Vernon Hills, IL). To record respiratory motor bursts, glass suction electrodes were attached to hypoglossal (XII) nerve rootlets (see Fig. 7A). Signals were amplified (10,000×) and band-pass filtered (10–10,000 Hz) using a differential alternating current amplifier (model 1700; A-M Systems, Everett, WA) before being rectified and integrated (time constant = 200 ms) using a moving averager (MA-821/RSP; CWE, Ardmore, PA).
Expiratory signals (upward traces) and inspiratory signals (downward traces) primarily occurred in pairs with expiration preceding inspiration (Fig. 2). Expiratory or inspiratory signals occurred by themselves very rarely and were therefore ignored. Expiratory signals were analyzed to measure timing variables such as breath frequency (breaths/min), breaths/episode, intraepisode breath interval (time between expiratory peaks within an episode), and singlet breath frequency. A singlet was defined as one breath (i.e., one expiration followed immediately by one inspiration) separated from other breaths by >7 s (i.e., greater than twice the duration of a single breath). Both expiratory and inspiratory signals were analyzed separately to measure the average Vt per breath (ml/kg). V̇e (ml·min−1·kg−1) was calculated by multiplying frequency times Vt per breath. All measurements were averaged into 60-min bins and reported as means ± SE. For in vitro data, burst duration and rise time (time from burst onset to burst peak) and bursts/episode were measured. In vivo and in vitro data were analyzed with Clampfit and Axoscope software, respectively (Axon Instruments, Foster City, CA). A two-way ANOVA with repeated measures design (Sigma Stat; Jandel Scientific Software, San Rafael, CA) was used to compare data before and after surgery. If normality or equal variance assumptions were not satisfied, data were ranked and the two-way ANOVA with repeated measures design was performed. Post hoc comparisons were made using the Student-Newman-Keuls test. P values < 0.05 were considered significant.
Before surgery, turtles breathed by moving their pectoral and pelvic girdles inward for expiration and outward for inspiration. After surgery, spinalized turtles were unable to move their pelvic girdle or their hindlimbs voluntarily; hindlimb withdrawal reflexes were still intact. During breathing, the pelvic girdle of spinalized turtles moved paradoxically compared with intact or sham turtles. When the pectoral girdle was pulled inward during expiration, the pelvic girdle was forced outward presumably by the increased intracoelomic pressure inside the shell. When the pectoral girdle was pulled outward during expiration, the pelvic girdle was pulled inward by the decreased intracoelomic pressure. Spinalized turtles were able to swim and move on dry surfaces using their functional forelimbs. Sham-operated turtles appeared to swim and walk similarly to intact turtles.
Spinalization Increased Singlet Breathing During Room Air Breathing
Turtles breathed room air for 2 h before CO2 exposures. The first hour allowed the turtles to equilibrate to the breathing chamber. Data from the second hour of this 2-h period for both CO2 protocols were pooled together and used to establish baseline breathing variables before surgery and at 4- and 8-wk postsurgery (Table 1). Spinalization did not alter baseline frequency, V̇e, and Vt. There was a strong trend toward decreased number of breaths/episode (P = 0.092), which was due to a larger singlet breath frequency (Fig. 2B). The rate of singlets per hour increased from 3.9 ± 1.7 (preinjury) to 26.7 ± 9.1 (4 wk) and 27.7 ± 8.8 (8 wk) (P = 0.05, no individual group was different from preinjury levels).
Spinalization Decreased Hypercapnic Responses
Presurgical hypercapnic responses.
Intact turtles had robust responses to hypercapnia, but spinalized turtles had significantly diminished responses (Fig. 2). Intact turtles increased their breathing frequency from baseline (1.5–1.7 breaths/min) to maximal values of 7.7 ± 1.4 breaths/min (6% CO2) and 7.8 ± 0.8 breaths/min (8% CO2) (all values P < 0.05 vs. baseline; Fig. 3, A1 and A2). Vt decreased from baseline (22 ± 3 ml/kg) to 15–18 ml/kg (P < 0.05) in response to 2% CO2 but increased to a maximum of 45 ± 5 ml/kg at 6% CO2 (P < 0.05; Fig. 3B1). In response to 4–8% CO2, Vt increased from baseline (13 ± 2 ml/kg) to maximal values of 22 ± 3 and 58 ± 6 ml/kg, respectively (P < 0.05; Fig. 3B2). V̇e increased from 28 ± 1 ml·kg−1·min−1 (baseline) to maximal values of 61 ± 9 and 360 ± 66 ml·kg−1·min−1 at 2–6% CO2, respectively. There were significant surgery- and CO2-dependent effects but no significant surgery-CO2 interactions at individual time points (Fig. 3C1). V̇e increased from 16 ± 6 ml·kg−1·min−1 (baseline) to maximal values of 88 ± 34 and 450 ± 68 ml·kg−1·min−1 at 4–8% CO2, respectively (P < 0.05; Fig. 3C2).
Postspinalization hypercapnic responses.
Spinalized turtles responded similarly to hypercapnia at 4 and 8 wk postspinalization (Fig. 2B). Breathing frequency did not change during hypercapnic challenges; breathing frequency was 1.6 ± 0.4 breaths/min (baseline) and only 2.4 ± 0.6 breaths/min (6% CO2) and 3.6 ± 1.0 breaths/min (8% CO2) at 4 and 8 wk (P > 0.05; Fig. 3, A1 and A2). Thus breathing frequency was significantly lower compared with preinjury levels during hypercapnic challenges at 2, 6, and 8% CO2 (Fig. 3, A1 and A2). During hypercapnic challenges of 2–6% CO2, Vt increased from 16 ± 2 ml/kg (baseline) to maximal values of 24 ± 3 (P < 0.05) and 22 ± 4 ml/kg (P > 0.05) at 4 and 8 wk, respectively, at 6% CO2 (P < 0.05); 2% CO2 did not change Vt (Fig. 3B1). During hypercapnic challenges of 4–8% CO2, Vt increased from 14 ± 2 ml/kg (baseline) to maximal values of 21 ± 2 (4% CO2) and 31 ± 4 ml/kg (8% CO2) at 4 wk and from 12 ± 1 ml/kg (baseline) to 19 ± 1 (4% CO2) and 27 ± 3 ml/kg (8% CO2) at 8 wk (all values P < 0.05 vs. baseline). Compared with intact turtles, spinalized turtles had a decreased Vt during 6 and 8% CO2 exposures (Fig. 3, B1 and B2). Despite a modest increase in Vt, 2–6% CO2 did not alter V̇e (Fig. 3C1). In contrast, V̇e increased from 23 ± 8 ml·kg−1·min−1 (baseline) to 79 ± 29 (4% CO2) and 130 ± 60 ml·kg−1·min−1 (8% CO2) at 4 wk and from 18 ± 3 ml·kg−1·min−1 (baseline) to 69 ± 23 (4% CO2) and 110 ± 37 ml·kg−1·min−1 (8% CO2) at 8 wk (all values P < 0.05 vs. baseline). Compared with intact turtles, spinalized turtles had decreased V̇e only at 6 and 8% CO2 (Fig. 3, C1 and C2).
Spinalization Reduced Episodic Breathing and Increased Singlet Breathing
Semiaquatic turtles breathe episodically with a series of expiratory-inspiratory cycles followed by prolonged apneic periods (i.e., breath holding); singlet breaths are rarely observed. The number of breaths per episode was larger in intact turtles compared with spinalized turtles at 4 and 8 wk postsurgery (Figs. 2 and 4, A1 and A2). During 2–6% CO2, the mean number of breaths/episode ranged between 4.7 and 8.3 breaths/min and did not change during CO2 exposures in intact turtles. In contrast, spinalized turtles had mean values ranging from 1.7–3.7 breaths/episode (except for the 8-h point in the 4-wk data), which was not altered by CO2 (P < 0.05 for surgery effect; Fig. 4A1). For intact turtles breathing 4–8% CO2, episodic breathing increased from 3.1 ± 0.4 breaths/episode (baseline) to 6.2 ± 0.8 and 8.9 ± 2.7 breaths/episode, respectively (P < 0.05; Fig. 4A2). In spinalized turtles, however, episodic breathing had mean values ranging from 2.0 to 4.5 breaths/episode at 4 and 8 wk postsurgery with no CO2-dependent alterations. For intact and spinalized turtles, the interval between breaths within episodes was not altered in intact vs. spinalized turtles (Fig. 4, B1 and B2). There were, however, significant increases for all turtles during exposures to 4–8% CO2 due to increased breath duration at larger Vt values. Finally, intact turtle singlet frequency was 3.3 ± 0.8 singlets/h while breathing room air or 2–8% CO2 (Fig. 4, C1 and C2). At 4 and 8 wk postsurgery, singlet frequency was increased by similar amounts during the 2–6% CO2 and 4–8% CO2 protocols (P < 0.05; Fig. 4, C1 and C2). There were no hypercapnia-induced changes in singlet frequency at 2–6% CO2 (Fig. 4C1). However, singlet frequency increased from 22 ± 8 (baseline) to 55 ± 17 singlets/h (8% CO2) after 4 wk and from 21 ± 7 (baseline) to 51 ± 21 singlets/h (8% CO2) after 8 wk (Fig. 4C2; P < 0.05 but no individual points reached significance).
Sham Operation Did Not Alter Hypercapnic Response or Breathing Pattern
In sham-operated turtles (n = 3), respiratory variables pre- and postsurgery were similar to each other and to data from preinjured turtles. CO2-dependent increases in frequency, Vt and V̇e were observed during 2–6% CO2 (Fig. 5, A1, B1, and C1) and 4–8% CO2 (Fig. 5, A2, B2, C2), but no significant differences were found pre- and postinjury. For episodic breathing, there was a CO2-dependent increase in the number of breaths/episode during the 6% CO2 exposure but no differences pre- and postinjury (Figs. 6, A1 and A2). Likewise, there were no CO2-dependent or surgery-dependent changes in intraepisode breath interval in sham-operated turtles (Fig. 6, B1 and B2). Singlet frequency in sham-operated turtles ranged between 0 and 8 singlets/h throughout both protocols and was not altered by breathing CO2 (data not shown).
Spinalization Did Not Alter Respiratory Motor Output and Central Chemosensitivity in Isolated Brain Stems In Vitro
To test whether spinalization induced similar long-lasting changes in the respiratory rhythm generation, brain stems were isolated from sham (n = 11) and spinalized (n = 17) turtles at 73 ± 2 and 71 ± 2 days postsurgery, respectively. Respiratory motor bursts were recorded from XII nerve rootlets for 30 min (Fig. 7A), and the results compared with data taken from control brain stems (n = 171; baseline data from Refs. 30, 31, 57). Although there were trends toward higher burst frequency and shorter burst duration in brain stems from spinalized turtles, there were no significant differences (Table 2). In addition, there were no differences in burst rise time or bursts per episode (Table 2). With respect to discharge pattern, the fraction of brain stems producing singlet discharge (i.e., bursts/episode ≤ 1.25) or episodic discharge (i.e., bursts/episode ≥ 2.00) were similar for intact turtles (singlets = 56%, episodic = 25%; n = 171) and spinalized turtles (singlets = 47%, episodic = 35%; n = 17). To test whether central chemosensitivity was altered by spinalization, brain stems from controls (n = 21), sham (n = 10), and spinalized (n = 10) turtles were switched from control solution (pH 7.34 ± 0.01; 5% CO2) to high pH-low CO2 solution (pH 7.76 ± 0.01; 1–2% CO2) for 90 min (Fig. 7B). This protocol is similar to the protocol that we used previously to measure pH/CO2 sensitivity in turtle brain stems in vitro (27) and represents a hypercapnic challenge because normal arterial pH in this species ranges from 7.74 to 7.78 at room temperature (reviewed in Ref. 27). The frequency change in all three groups was similar with slopes of −0.58 ± 0.09, −0.40 ± 0.11, and −0.61 ± 0.16 bursts·min−1·pH unit−1 for control, sham, and spinalized turtles, respectively (P > 0.05; Fig. 7C). Thus central chemosensitivity in isolated brain stems was not altered by spinalization.
This is the first study to examine changes in breathing following spinal cord injury in a freely behaving, aspirating ectothermic vertebrate. We found that spinalized turtles did not have compromised V̇e while breathing room air, suggesting that compensation occurred despite the loss of the main respiratory pump muscle. Spinalized turtles, however, did have an altered breathing pattern with a significantly increased singlet breathing rate. In response to hypercapnia, spinalized turtles had significantly decreased ventilatory responses due to a reduction in frequency and Vt. Also, episodic breathing was decreased and singlet breathing was increased. There was little evidence for spontaneous recovery of respiratory function over a 2-mo period. Isolated brain stems from spinalized and sham turtles had similar patterns of respiratory motor output, and central chemosensitivity was not changed, suggesting that reduced hypercapnic responses and breathing pattern changes in spinalized turtles were not due to changes within the respiratory rhythm generator in the brain stem.
Breathing in Intact and Spinalized Turtles: Comparisons and Caveats
Respiratory variables in this study were similar to previously published results in red-eared slider turtles breathing room air with a similar apparatus (13, 50). Frequency in this study was slightly lower at 1.29 ± 0.18 breaths/min (compared with 1.8–3.0 breaths/min), Vt was slightly larger at 17.2 ± 1.1 ml/kg (compared with 9.0–13.5 ml/kg), and V̇e was similar at 22.1 ± 3.3 ml·kg−1·min−1 (compared with 16.7–24.8 ml·kg−1·min−1). For hypercapnic challenges from 2 to 8% CO2, frequency increased from 1.3 to 7.8 breaths/min, similar to the increase of 3.0–8.8 breaths/min reported by Funk and Milsom (14), although Vt and V̇e were larger in this study. For example, Vt and V̇e increased in this study from 17 to 58 ml/kg and from 22 to 450 ml·kg−1·min−1, respectively, compared with 9–15 ml/kg and 20–140 ml·kg−1·min−1 (14). The larger Vt in this study could be due to differences in experimental protocols. In this study, turtles breathed either 2 or 4% CO2 for 2 h before challenges at 6 and 8% CO2, whereas turtles were exposed to the various CO2 levels for 2 h in a random manner in the study by Funk and Milsom. Thus arterial Pco2 may have started at much higher levels in this study before 6 and 8% CO2 exposures. Subtle differences in calibration methods may have also contributed to differences between the two studies.
Spinalized Turtles Compensate for Injury While Breathing Room Air
Spinal cord injury alters breathing pattern because of impaired ability to pump air and maintain airway patency as well as altered breathing mechanics. Spinalization in the turtle caudal spinal cord abolished use of the pelvic girdle, the main respiratory pump muscle (15, 16, 43), and forced the turtles to breathe only with their pectoral girdle. Surprisingly, most breathing variables in intact and spinalized turtles were similar while breathing room air (changes in breaths/episode and singlet breathing are discussed below). One explanation is that pectoral girdle movement alone in spinalized turtles was sufficient for V̇e and maintenance of normal blood-gas levels. This could be because these spinalized turtles recovered in a tank and were relatively sedentary. If spinalized turtles were released into the wild where they would have to swim and walk to survive (i.e., feed and escape predators), an increase in metabolic demand may have revealed respiratory insufficiency while breathing room air. Measurement of arterial Po2, Pco2, and pH pre- and postinjury would have helped resolve this question. However, maintaining chronic arterial lines for more than 2 mo in semiaquatic turtles without causing infection is technically difficult. In intact turtles, arterial Pco2 values would be expected to be 27–29, 34–36, 40–45, and 49 Torr while breathing 2, 4, 6, and 8% CO2, respectively (14, 25), which is within the normal limits since arterial Pco2 reaches 130 Torr following a 2- to 4-h dive (25). Alternatively, turtles may have also altered their behavior and voluntarily moved their forelimbs back and forth to move the pectoral girdle and help ventilate their lungs.
A second compensatory mechanism might be decreased pulmonary stretch receptor activity because of lower lung volumes. Decreased pulmonary stretch receptor activity would result in increased Vt and breath frequency and a decrease in the apneic period between episodes (46). Third, synaptic drive may have been augmented to remaining respiratory muscles rostral to the spinal transection (due to changes in the brain stem or cervical spinal cord), i.e., to pectoralis and serratus muscles, which control the pectoral girdle (15). For example, there is electrophysiological evidence that bulbospinal (or propriospinal) premotor axons sprout onto thoracic motoneurons that are rostral to a spinal transection (8). Consistent with this hypothesis, preliminary evidence suggests that activity-dependent synaptic plasticity in descending inputs to respiratory cervical spinal motoneurons is shifted toward increased short-term potentiation and decreased long-term depression in spinalized turtles (29). Fourth, alterations at the level of motoneurons, neuromuscular junctions, and muscle fibers may have contributed to normal V̇e while breathing room air (42).
Hypercapnic Responses Are Attenuated in Spinalized Turtles
In response to hypercapnia, turtles increase their V̇e by increasing Vt, frequency, and breaths/episode (1, 10, 14, 22, 24, 50) with little change in expiratory and inspiratory duration (50). In this study, Vt and frequency decreased in spinalized turtles while breathing 6–8% CO2, whereas only Vt decreased with 2% CO2 (Fig. 3). It is not clear why Vt and frequency in spinalized turtles were not different from intact controls at 4% CO2. The injury-induced decrease in Vt was expected because use of the pectoral girdle was abolished; only the pectoral girdle was available for pumping air. CO2-dependent depression of pulmonary stretch receptors (32, 48, 49) may have contributed to increased bulbospinal respiratory drive to remaining muscles controlling the pectoral girdle, but this was insufficient to restore Vt to preinjury levels during CO2 breathing.
The inability of spinalized turtles to substantially increase breathing frequency during hypercapnia was surprising because many spinally injured animals, including humans, increase breathing frequency to compensate for decreased Vt (7, 19, 40). Likewise, because turtle central chemoreceptors regulate respiratory frequency (22), one would have expected respiratory frequency to increase during hypercapnic challenges. Thus it is possible that spinalized turtles may have developed decreased chemosensitivity and therefore did not respond to hypercapnic challenges to the same degree as intact turtles. However, spinalized and intact turtles had similar respiratory variables while breathing room air, suggesting that spinalized turtles were not hypoventilating. Also, isolated brain stems from control and spinalized turtles had similar hypercapnic responses, suggesting that central chemosensitivity was not altered.
Other mechanisms that may have contributed to a decreased hypercapnic response include muscle fatigue, altered breathing mechanics, and loss of spinobulbar inputs to the respiratory rhythm generator. For example, pectoralis and serratus muscles controlling the pectoral girdle maintained adequate Vt during 2–4% CO2 challenges but may have fatigued at 6–8% CO2. This would explain the decrease in Vt but not the decrease in frequency and episodic breathing (see below). Alternatively, breathing mechanics may have been altered by spinalization, but this is unlikely because the turtle lung is attached to the inside of the rigid carapace and plastron, which are unaltered following spinal transection. Finally, loss of propriospinal inputs from the lumbar region may have decreased cervical spinal motoneuron excitability. Turtles have propriospinal projections that connect neurons in the lumbar and cervical enlargements and help coordinate hindlimb and forelimb movements (39). Lumbar projections may also provide tonic excitatory synaptic input to cervical spinal respiratory motoneurons, which is lost with spinalization.
Reduced Episodic Breathing in Spinalized Turtles: Potential Mechanisms
All vertebrate respiratory control systems are hypothesized to be capable of producing episodic breathing patterns, depending on species, physiological condition and developmental stage (45). Factors that modulate episodic breathing in ectothermic vertebrates include peripheral chemo- and mechanosensory afferent inputs (34, 36, 37, 53, 54), descending inputs from rostral brain structures (34, 47, 52) and mechanisms intrinsic to the brain stem (55, 58). In this study, transection of the caudal turtle spinal cord significantly increased singlet breathing and attenuated the number of breaths/episode when challenged with hypercapnia.
One interpretation is that neurons or passing fibers in the caudal spinal cord modulate breathing by altering episodic breathing itself or factors that regulate episodic breathing (e.g., respiratory drive). Although evidence showing that neuronal activation in caudal thoracic and lumbar spinal cord alters respiratory timing is sparse and controversial (26), electrical activation of ventral cervical spinal cord alters respiratory timing in neonatal rats in vitro (5). In addition, activation of skeletal muscle sensory afferents increases V̇e via spinobulbar pathways that appear to terminate in the ventrolateral medulla (33). Because the turtle spinal cord has many features in common with other vertebrate spinal cords (2), it is possible that sensory afferents from oblique and transverse muscles (which control pelvic girdle movement during breathing) synapse onto spinobulbar pathways that project to the turtle respiratory rhythm generator in the brain stem (38). The absence of these inputs in spinalized turtles may limit the extent to which the rhythm generator can generate episodic breathing.
Alternatively, functional loss of the pelvic girdle may have established conditions within the central nervous system that led to an altered breathing pattern during hypercapnia. For example, spinalized turtles may have had chronically elevated arterial Pco2 and low pH levels such that, during hypercapnic challenges, further increases in respiratory drive (due to central chemoreceptor activation), and coupled with increased Pco2 (and low pH) in the brain stem, may have been disruptive and caused the respiratory rhythm generator to produce a more ataxic breathing pattern with increased singlet breathing. Also, low blood pH during hypercapnia may have caused ion shifts (e.g., hyperkalemia) that altered function within the respiratory rhythm generator. Further experiments are required to test these hypotheses.
Advantages of Spinalized Turtles for Studying Spinal Cord Injury
The spinalized turtle model is advantageous because respiration and chemoreflexes can be easily measured in awake, freely behaving spinally injured animals that require little postinjury care (e.g., bladder expression, feeding). Also, genetic differences due to inbreeding that may confound data interpretation in spinally injured animals (44) are not a problem because turtles are captured from the wild by commercial suppliers. Third, the lack of improvement in hypercapnic responsiveness over a 2-mo period means that specific therapeutic interventions can be tested and their effectiveness quantified without significant ongoing spontaneous recovery. In contrast, other species such as rats undergo significant spontaneous recovery following spinal cord injury that may depend on strain (44). Potential therapeutic interventions that could be tested in spinalized turtles include chronic intermittent hypoxia (12), chronic alterations in lung volume via experimentally induced buoyancy changes (46), and increased activity and exercise (23). Fourth, if a specific intervention successfully improved respiratory function, the brain stem (Fig. 7) or brain stem-spinal cord (28) could be removed and studied under in vitro conditions to test specific hypotheses under more controlled conditions.
This work was funded by the University of Wisconsin Graduate School, Center for Neuroscience and School of Veterinary Medicine.
We thank Frank J. Golder for critically reviewing early drafts of this manuscript and Julia E. R. Wilkerson for excellent technical assistance.
- Copyright © 2005 the American Physiological Society