Respiration Physiology. 84 (1991) 61-76 Elsevier
Respiratory patterns in anesthetised rats before and after anemic decerebration F. Hayashi and J.D. Sinclair Department of Physiology. School of Medicine, University of Auckland. Auckland. New Zealand (Accepted 12 January 1991) Abstract. Experiments were undertaken to test the comparability of changes in respiratory frequency and tidal volume during hypoxia and hypercapnia in rats with and without intact peripheral chemoreceptors and with intact vagi. Neural organisation of respiratory control was perturbed by anemic decerebration, achieved by ligation of the common carotid and basilar arteries. Ischemia of the brain was produced as far candal as the rostral pontine nuclei involved in respiratory control but left the medulla well perfused. The dominant respiratory effect in animals breathing air or oxygen was polypnea with hypocapnia (mean Pace, when breathing air 24.7 mmHg, when breathing oxygen 29.6 mmHg). After decerebration the increase of ventilation produced by breathing 10~o 0 , in N2 was reduced compared with responses in the intact state but levels of ventilation (VI) in hypoxia were similar to those before decerebration. After decerebration, the increase of ventilation produced by breathing 5 ?~oCO2 was greatly reduced and the level of ~'l in animals breathing CO2 was significantly less than in the intact state. Intermediate changes were seen in animals breathing 2-3 ?/o CO2 which converted the hypocapnia (Pace, 30.9 mmHg) to eucapnia (Pace, 46.4 mmHg). In the intact state, hypoxia dominantly caused increased frequency (f) and hypercapnia caused increased tidal volume (VT); after decerebration, hypoxia produced reduction of VT while hypercapnia produced reduction of f. Bilateral carotid sinus nerve section in decerebrate animals eliminated the ventilatory response to hypoxia but left the responses to hypercapnia unaltered. The results point to differences in the mechanisms by which hypoxia and hypercapnia influence respiration in both intact and decerebrate animals with carotid sinus and vagus nerves functional. The differences can now be interpreted in terms of specific neural features of respiratory control.
Control of breathing, hypoxia and hypercapnia; Decerebration, by anemia; Respiratory pattern, after decerebration
Widely accepted models of respiratory control proposed by Clark and von Euler (1972) predict that increased chemical drive will accelerate inspiration, producing an increased tidal volume in a briefer respiratory cycle. Hypoxia and hypercapnia are thus expected to produce similar patterns of increase of both tidal volume and frequency. However, longstanding observations on humans suggest that hypoxia specifically drives respira-
Correspondence to: J.D. Sinclair, Dept. of Physiology, School of Medicine, University of Auckland, Private Bag, Auckland, New Zealand. 0034-5687/91/$03.50 © 1991 Elsevier Science Publishers B.V,
F. HAYASHI AND J.D. SINCLAIR
tory frequency while experiments on decerebrate cats suggest that the direct output variable of the CO2 controller is tidal volume (Florez and Borison, 1967). Gautier (1976) showed differences between responses to hypoxia and hypercapnia in awake and anesthetised cats; and much neurophysiological evidence has accrued to support a difference between the actions of these two chemical stimuli on respiratory rhythm generation. St John (1979), from changes in apneustic breathing patterns ofvagc~tomised decerebrate cats, argued for a discrete effect of peripheral chemoreceptor input and a diffuse effect of central chemoreceptor input on hindbrain respiratory neurons. Sears et al. (1982) proposed a differential effect, with hypoxia initially affecting inspiratory, and hypercapnia initially affecting expiratory, motor output. Carotid body denervation produces a dramatic loss of the increase in frequency normally occurring in awake rats exposed to hypoxia (Martin-Body et al., 1985) while separation of frequency and tidal volume components is revealed by small variations in levels of decerebration (Martin-Body, 1988), and by application' of focal cold block to different parts of the exposed ventral medullary surface (Homma et al., 1988). Gautier (1976)concluded that the explanation of chemical influences required a more complex model of the control of rhythm generation than that of Clark and von Euler (1972). More recent models such as that of Richter et al. (1986) allow chemical drive to be asserted at more than one level and thus provide a better framework within which experimental observations may now be tested. In some earlier experiments concerned with the chemical regulation of breathing, we considered the effects of ligation of internal carotid and vertebral arteries in the rat, a procedure similar to that described for the cat as 'anemic decerebration' (Pollock and Davis, 1923). Residual responses to chemical drive included a slowing of respiration during hypercapnic stimulation but a contrasting acceleration during hypoxic stimulation. A precise interpretation of these results was limited by the variability of the levels of ischemic decerebration and by the lack of blood gas data. However, a slowing of frequency during hypercapnia has been demonstrated in anesthetised vagotomised rabbits tested at arterial CO2 levels as high as 250 mmHg (Kobayasi and Murata, 1979) e.nd in deeply anesthetised vagotomised rats breathing up to 10~o CO2 (Martin-Body and Sinclair, 1987). A further series of experiments was undertaken to study frequency and tidal volume changes in more detail, examining specifically the effects of hypoxia and hypercapnia before and after the destruction of forebrain and midbrain function. A uniform decerebration was achieved by ligation of common carotid arteries and of the basilar artery rostral to the posterior inferior cerebellar arteries. Arterial blood gas levels were measured. Special interest was taken in assessing the responses of frequency and tidal volume during hypercapnia and hypoxia, and when the hypocapnic state seen after decerebration was converted to a state of approximate eucapnia. The aim of the studies was to demonstrate that, in the presence of normal sensory inputs from chemoreceptors and lung inflation receptors, hypoxia and hypercapnia acted differently on the respiratory patterns established in the hindbrain.
CHEMICAL INFLUENCES ON DECEREBRATE BREATHING
Studies were completed in a total of 21 male white rats of the Wistar strain, weighing 290-350g anesthetised with a mixture of urethane (700mg/kg) and 0c-chloralose (75 mg/kg) (i.p.) supplemented as necessary with additional chloralose (i.p.). Cannulae were inserted into the femoral artery and vein and the trachea. Body temperature was monitored with a rectal probe and maintained at 37 + 0.4 °C with a heating lamp. Arterial pressure was monitored continuously. Inspired gases were supplied from calibrated cylinders via rotameters which provided a constant flow rate of 1 L/rain in a line connected to the tracheal tube by a T-piece without the incorporation of valves. Flow was detected by a pneumotachograph (Fleisch, pressure fall 10mmH20 for 22.67 ml/sec) placed on the expiratory flow line. The background flow provided linear characteristics in the pneumotachograph and prevented significant rebreathing of expired gas. Airway pressure was less than 0.4 mm H20. The pneumotachograph signal was detected by a differential pressure transducer (Validyne MP 45-1, + 2 cmH20) and signal conditioner (Validyne CD 15 carrier demodulator) and integrated in a summing device and processor to give inspiratory tidal volume, inspiratory minute volume, and frequency. Respiratory responses were measured over periods of I rain following 3 min exposure to 5% CO2 in 02 and after 2 min exposure to 10% 02 in N 2. The electrocardiograph was recorded continuously and the heart rate (HR) derived from the ECG signal. Arterial blood samples were taken in microcapillary tubes and analysed at 37 °C. Associated blood loss was replaced with a heparinised saline solution (Na 154 mM; CI 130 mM; HCO3 24 mM). Denervation of the carotid bodies was achieved by bilateral carotid sinus nerve section and the effectiveness confirmed by the immediate loss of the ventilatory response to hypoxia or to intravenous injection of NaCN (0.5 mg/kg). Cervical vagotomy was performed in 3 animals but none survived for long enough to allow valid measurements, presumably because of the pulmonary oedema which characterises the response of the lightly anesthetised rat to acute bilateral vagotomy (Martin-Body and Sinclair, 1987).
Anemic decerebration. This was achieved by ligation of both common carotid arteries and the basilar artery. Each common carotid artery was carefully separated from the vagal, sympathetic, and aortic nerves and tied. The ventral surface of the hindbrain was exposed surgically and the basilar artery ligated rostral to its posterior inferior cerebellar branch, using fine surgical thread (Ethicon 10-0 Ethilon). The hindbrain surface was thereafter kept warm and moist. In some animals temporal craniotomy was performed to prevent effects on respiration of increased intracranial pressure.
Experimentalprocedure. The common carotid arteries and basilar artery were exposed but not ligated and control records of respiration made in sequence while the animals breathed air, 100% 02, 5% CO2 in 02 and 10% 02 in N2. The three arteries were then ligated. To ensure the immediate survival ofthe animals it was found necessary to supply 100% 02 for the period immediately following ligation. After I h, the respiratory
F. HAYASHI AND J.D. SINCLAIR
measurements were repeated. In 8 animals, the carotid sinus nerves were exposed during preliminary surgery, and were sectioned after completion of the post decerebration studies. The ventilatory tests were repeated within the following 15 rain. The brain was then prepared for histological study.
Preparation of brain tissues. At the conclusion of the experiment, the anesthetised animal's chest was opened, the descending aorta was clamped, and 200 ml of physiological saline containing 5000 IU of heparin was perfused into the upper body by intracardiac injection. One ml of a filtered black waterproof ink (Rotting) was then injected at a pressure of approximately 120 mmHg and this was followed with 200 ml of fixative (4% paraformaldehyde in 0.1 M phosphate buffer) to precipitate the ink. Since unperfused brain areas were not fixed by this procedure, the brain and upper cervical cord were then gently removed and stored in the fixative at 5 °C for 7 days and in 12.5% sucrose in 0.1 M phosphate buffer for 2 days. The tissue was then sectioned (100 #m) in a cryostat in the sagittal plane and stained in 0.5% Neutral Red. Records of regions with and without ink-filled capillaries were made on appropriate plans of the rat brain (Paxinos et al., 1982).
Anatomical effects. Anatomical study of the brain demonstrated that the arterial ligations had produced a clear delineation between regions with and without residual perfusion. Cortex, diencephalon, mesencephalon and cerebellum showed no ink staining. In the hindbrain, ischemic decerebration was at pontine level, illustrated in a typical case by the photograph in fig. 1. Microscopic examination also showed abrupt and clearly defined limits to ink-filled capillaries (fig. 2). Medially the border between perfused and non-perfused tissues lay dorsally at the caudal edge of the tegmental and pontine raphe nuclei and ventrally, between the pontine nucleus and the trapezoid body. Laterally, the border was, dorsally, between the locus coeruleus (unperfused) and the medial vestibular nuclei (perfused) and ventrally, the pontine nuclei were not perfused while the nucleus ofthe trapezoid body and the periolivary and parvocellular nuclei were perfused. Among nuclei significant for respiratory control, the lateral and medial parabrachial nuclei of the pons were unperfused as was the motor trigeminal nucleus. Nuclei retaining perfusion included the nucleus of the tractus solitarius, the retrofacial nucleus (BOtzinger complex), nucleus ambiguus and associated nuclei, the retrotrapezoid nucleus, nuclei of 7th, 10th and 12th cranial nerves, and the paragigantocellular and lateral reticular nuclei. The results were consistent in 8 animals in which total histological examination was successfully completed. Cmdiorespiratory effects. The immediate effect of arterial ligation was an increase of arterial blood pressure and of heart rate, the 'cerebral ischemic response' (fig. 3). In 16 rats, the mean arterial BP increased from 105 + 6 (SEM) mmHg to a maximum of
CHEMICAL INFLUENCES ON DECEREBRATE BREATHING
Fig. I. Photograph of brain perfused with black ink following ligation of both common carotid arteries and of basilar artery rostral to posterior inferior cerebellar branch. Demarcation between regions with and witbout residual perfusion is visible at approximately midpontine level.
146 + 6 mmHg (P < 0.01); the mean heart rate (HR) increased from 417 + 8 to 465 + 8 beats per minute; (P < 0.01). The changes began within 10-20 sec of completion of vascular ligation, reached maximum in 1-2 min, and subsided over the following 1-2 min. After 60 rain the mean BP was 98 + 6 mmHg in air, 105 + 6 mmHg in 02; and mean HR was 494 + 9 in air, 479 + 8 in 02. Immediately after ligation of all arteries, the respiratory pattern was characterized by hyperventilation and frequent augmented breaths or 'sighs'. These now occurred at intervals of 20-30 sec, compared with intervals of 120-180 sec before decerebration. Measurements of minute ventilation (VI) tidal volume (VT) and respiratory frequency (f) in rats breathing air and 100~o 02 before and I h after decerebration are shown in table 1. In 15 animals breathing air, decerebration led to significant increases in ('~I) from a mean level of 138.6 + 4.1 (SEM) to 206.2 + 8.8 ml/min; the 49?/0 increase of ('~l) was the result of mean increases of 177/o in f and 29~o in VT. The arterial P¢o2 decreased significantly, from a mean value of 41.3 mmHg to 24.7 mmHg. When breathing 100 % 02, the animals still showed significant increases of (VI) after decerebration, by 43% ofthe level in the intact state (from 132.2 to 188.4 ml/min), again the result of increases of both f (by 7 ~o) and VT (by 41 ~o). Before decerebration, in these animals breathing 100~o 02, mean arterial P¢o: was 46.0 mmHg, significantly higher than during air breathing, and after decerebration, it decreased to 29.6 mmHg (P < 0.01).
F. HAYASHI AND J.D. SINCLAIR
Fig. 2. Drawings ofsagittai sections of rat brain (Paxinos and Watson, 1982) indicate regions made ischemic by anemic decerebration. Labelled nuclei include those with major significance in respiratory control (see text). 7 = facial nucleus; 12 = hypoglossal nucleus; Amb -- nucleus ambiguus; CVL = caudoventrolateral reticular nt;cleus; DPGi - dorsal paragigantocellular nucleus; IO = inferior olive; LC = locus coeruleus; LPB = lateral parabrachial nucleus; LPGi = lateral paragigantocellular nucleus; L r t - l a t e r a l reticular nucleus; LSO = lateral superior olive; LVe - lateral vestibular nucleus; MnR - median raphe nucleus; Me5 = motor trigeminal nucleus; MPB = medial parabrachial nucleus; MVe = medial vestibular nucleus; MVeV = medial vestibular nucleus, ventral part; MVPO = medioventral periolivary nucleus;
CHEMICAL INFLUENCES ON DECEREBRATE BREATHING
, , '!/ ,1 ,1 ,t /
ABP 100 a
I 1 min
Fig. 3. Recording showing the immediate increases of heart rate (HR) and arterial pressure mmHg (ABP) following ligation (at arrow) of common carotid and basilar arteries (RR respiratory frequency; V, expired volume, ml).
Responses to hypercapnia. When intact animals breathed 5 % CO2 in 02 after breathing 100% 02, arterial BP and HR were unchanged. In the decerebrate animals, BP did not change but HR decreased from 479 + 8 to 441 + 11 per min (P < 0.01). In the intact animals, (~'I) increased from 132.2 ml/min in 02 to 247.5 ml/min in 5% CO2, i.e. by 87%, while arterial P¢o2 rose from 46.0 to 60.0 mmHg. In the decerebrate animals, ('¢I) increased from a mean level of 188.4 ml/min to 203.9 ml/min, an increase of 8 % while arterial P¢o, rose from 29.6 to 54.8 mmHg. In 5 of 13 rats, (VI) diminished in the hypercapnic state. PCRtA = parvoceUular reticular nucleus, alpha part; Pn = pontine nuclei, py = pyramidal tract; pyx = pyramidal decussation; RtTg - reticulotegmentai nucleus of the pons; RVL - rostroventrolateral reticular nucleus; s c p - superior cerebellar peduncle; Sol - nucleus of the solitary tract; SolC = nucleus of the solitary tract, commissural part; S PO = superior paraolivary nucleus; SpVe - spinal vestibular nucleus; SubCA = subcoeruleus nucleus, alpha part; tz = trapezoid body.
F. HAYASHI AND J.D. SINCLAIR TABLE 1
Effects of anemic decerebration on arterial gas levels and on respiratory frequency, (f), tidal volume, (VT), and minute volume, (~h), in rats breathing air, 100% 02, 5% CO2 in 02 and 10% 02. Values are mean _+ SEM. Significance of differences between values in intact and decerebrate state are shown (*P < 0.05, **P < 0.01). In parentheses, number of animals. Inhaled gas
Air Pao2(mmHg) Paco2(mmHg) f(min) VT (ml) "~)l(ml/min)
(9) (10) (15) (I 5) (15)
92.5 ,-_2.1 41.3 +- 1.0 91.5 --.3.5 1.53 +- 0.05 138.6 +4.1
109.7 __ 24.7 +106.8 _ 1.97 __ 206.2 _
2.2 1.5 4.1 0. l I 8.8
** ** ** ** **
100% 02 Pao:, (mmHg) Paco2 (mmHg) f(min) VT (ml) Vl(ml/min)
(7) (7) (147 (14) (147
>450 46.0 _+ 1.3 80.1 +2.7 1.66 + 0.04 132.2 _+4.2
> 450 29.6 86.1 2.34 188.4
_+ + _+ _+
1.7 4.4 0.12 7.2
5 % CO2 in 02 Pao2 (mmHg) Paco2 (mmHg) ftmin) VT (ml) Vl (ml/min)
(7) (7) (13) (13) (13)
>450 60.0 I02.8 2.47 247.5
_+ 1.4 _+5.4 _+0.13 _+ 8.3
>450 54.8 75.7 2.80 203.9
_+ 2.0 _+ 7.0 _+ 0.15 _+ 13.8
10% 02 Pao:(mmHg) Paco2 (mmHg) f(min) VT (ml) ~'l (ml/min)
(7) (7) (13) (137 (137
53.4 30.2 128.8 !.66 206.1
+ 1.6 + 1.4 :t:7.3 + 0.12 + 8.2
61.6 23.0 !15.8 i.93 212.3
+ _+ + + +
2.5 1.4 7.2 0.15 6.6
In the control (intact) state the change was effected by an increase of 2 8 ~ in f and 4 9 ~ in VT (table 1). After decerebration, the corresponding changes were a mean decrease of 12~o in f and a mean increase of 2 0 ~ in VT. The difference between the frequency responses in the two states was highly significant (P < 0.01) as was the resulting difference in ('¢I) response. An example of depression of f in hypercapnia is shown in fig. 4. Because of the possibility that the hypercapnic response reflected changes specific to hypocapnia, further studies were conducted in which respiration was measured after decerebration in 6 rats breathing 100~ 02, 5~o CO2 in 02, and 2 - 3 ~ CO2 in 0 2 to produce measurements in approximately eucapnic conditions. The respiratory charac= teristics at arterial CO2 levels close to those found in intact rats breathing 100~ 02
CHEMICAL INFLUENCES ON DECEREBRATE BREATHING
'ool 5O 40"
Fig. 4. Reco~'d of the ventilatory response to CO2 after anemic decerebration. When the inspired gas is switched from 100% 02 to 5 % CO2/95 % 02, (arrow) heart rate (HR) decreases, respiratory frequency (RR) decreases, ventilation (V) temporarily increases. The effects reverse on switching inspired gas back to 100 % 02 (second arre,w). The record of respiratory frequency shows frequent occurrence of sighs. Interruption of record of arterial pressure indicates times of arterial sampling.
proved intermediate between those of hypocapnia and hypercapnia (table 2). Thus respiratory frequency fell from a mean of 90.5 in rats breathing 100% 02 (Paco: 30.9 mmHg) to 82.1 in rats breathing 2-3 % CO2 in 02 (Paco2 46.4 mmHg) and to 78.3 in rats breathing 5% CO2 (Paco2 56.7 mmHg). TABLE 2 Mean values ( _ SEM) for tidal volume (VT) frequency (f} and ventilation (VI) in 6 decerebrate rats breathing 100% 02, 2-3% CO2 in 02, and 5% CO2 in 02. The Paco2 of dccerebrate rats breathing 2-3% CO2 is similar to that of intact rats breathing 100% 02; see table 1. Inhaled gas
100% 0 2 2-3% CO., in 02 5% CO:, in 02
1.95 + 0.19 2.36 + 0.20 2.67 + 0.30
90.5 ± 3.3 82.1 + 3.4 78.3 + 2.2
181 + 21 193 + 16 208 + 22
30.9 + 1.7 46.4 + 2.4 56.7 .+. !.1
F. HAYASHI AND J.D. SINCLAIR
The combined results from the two series of experiments show that in a total of 19 decerebrate animals, breathing 5 % COz in 0 2 after breathing 100~o 02 decreased ffrom 87.4 + 3.4 to 76.5 + 4.8 (P < 0.01), increased VT from 2.15 + 0.11 to 2.76 + 0.14 ml (P < 0.01) and increased (~q) from 184.5 + 8.5 to 205.3 + 11.3 ml/min (P < 0.05). Responses to hypoxia. In intact animals exposed to 10% 0 2 in N 2 after breathing 100% 02, the mean level of arterial BP fell from 105 + 6 to 60 + 5 mmHg and HR increased from 411 + 8 to 449 + 13. Arterial Po., fell to a mean level of 53 mmHg and arterial Pco2 to 30.2 mmHg. ('¢I) increased by 56% from 132.2 to 206.1 ml/min; by an increase of 61Y/o in f without significant change of tidal volume (table 1). When decerebrate rats were switched from hyperoxic to hypoxic conditions, B P fell from 105 + 6 to 63 + 8 mmHg, HR increased from 479 + 8 to 486 + 14 per min, arterial Po2 fell to 61.6 mmHg and arterial Pco2 fell to 23.0 mmHg. (VI) in this state increased by 13% from a mean level of 188.4 to 212.3 ml/min (P < 0.05), fincreased by 34% and mean VT decreased by 18% (table 1). The differences in responses off, VT and ('~/I) to hypoxia, after decerebration compared with )he intact state, were all significant (P < 0.01) but, because of their high levels in hyperoxic conditions in the decerebrate animals, the mean levels during hypoxia were not significantly different in intact and decerebrate animals. Effect of carotid sinus nerve section. Bilateral carotid body denervation eliminated the frequent sighs seen after decerebration and reduced the level ofhyperventilation but ('~/I) did not return to the control value. The reduction of (~'l) resulted from the reductions of f, with VT not significantly altered. Responses to inhaled 5 ~o CO2 were not altered significantly by carotid body denervation but the hypoxic response was almost eliminated. Details are presented in table 3. Circulatory responses to the denervation comprised small decreases of BP and increases of HR in rats breathing air, 100% 02, or 5% CO2 in Oa; in 10% 02, mean HR increase was also small (from 474 + 8 to480 + 9) but arterial BP fell from 100 + 11 mmHg in 02 to 51 + 4 mmHg in 10% Oz compared with a fall to 70 + 8 mmHg in the decer~brate state prior to denervation.
Discussion These experiments showed that in the rat the combined ligation of the common carotid and basilar arteries produced ischemia of the forebrain, midbrain and rostral pons. As long as animals were well oxygenated they survived the procedure without a major change of heart rate or arterial pressure. In a stable state an hour after decerebration, compared with the intact state, their respiration was characterised by hyperpnoea, with increases of frequency and tidal volume, and an associated hypocapnia. There was a striking attenuation of the response to inhaled carbon dioxide in all cases. The normal increase of ventilation was replaced by a reduction in 5 of 13 animals; and while the normal mediation of the CO2 response was by increased VT, in the decerebrate state
C H E M I C A L I N F L U E N C E S ON DECEREBRATE B R E A T H I N G TABLE 3
The effects of anemic decerebration and subsequent carotid sinus nerve section (CSN-X) on the ventilatory variables in 8 rats breathing various inhaled gases. Values are mean + SEM (*P < 0.05, **P < 0.01) Inhaled gas
Decerebrate and CSN-X
Air "~! (ml/min) f (rain)
137.8 + 2.5
200.9 + 6.3**
169.3 + 6.2**
87.1 + 4.4
99.6 + 5.7*
85.8 + 6.6*
VT (ml) 100% 02 I (ml/min) f (min) VT (ml) 5 % CO2/95% 02
VI (ml/min) f (min) VT (ml) 10% 02
VI (ml/min) f (rain) VT (ml)
1.60 + 0.08
2.06 + 0.12"
2.03 + 0.12
129.5 + 3.8 78.5 Z 3.6 !.66 + 0.14
186.4 + 8.5** 84.1 + 7.2 2.29 + 0.15"*
165.4 + 9.1" 81.1 + 6.6 2.08 + 0.09
253.3 + 12.9 100.6 + 8.7 2.60+ 0.19
199.4 + 15.8" 71.9 + 10.0 2.91 + 0.18
188.9 +21.5 74.3 + 11.8 2.70+ 0.16
198.5 + 9.4 115.6 + 7.3 1.78 + 0.17
209.4 + 5.8 106.0 + 7.8 2.06 + 0.18
175.8 + 9.6* 81.8 + 5.0* 2.18 + 0.11
the response was characterised by decreased f, with VT still showing an increase of20%. The response to hypoxia was reduced to a lesser degree and the resulting levels of ventilation during hypoxia were not different in the intact and decerebrate states; but while an increase of f was still present after decerebration, the mean VT was decreased relative to the value in oxygen. The contrast in these effects is illustrated in fig. 5, which demonstrates the consistency of effects among animals. Before considering the interpretation of these results, several methodological points need emphasis. First, the anemic decerebration of Pollock and Davis (1923), while shown here to produce a consistent and stable preparation for respiratory studies and while avoiding the blood loss associated with surgical transection, must still, in acute studies, abruptly eliminate neural input from rostral areas of the brain and produce a gross disturbance of the excitability of residual circuits. In the interpretation of the anatomical extent oflesioning, while it can be assumed that regions subsequently shown to be without perfusion were not contributing to function, it cannot be assumed that perfusion had remained quantitatively normal in all regions in which post-mortem ink-filling indicated patency of capillaries. As a species, the rat has distinct advantages for respiratory studies, including acceptability, economy, and physiological suitability as a model for human respiration (Saether et al., 1987). However, the difficulty of monitoring end tidal or arterial gas levels limits the ability to maintain constant conditions, especially eucapnia in the hypoxic state. Intermittent arterial sampling, as done in the present study, permits
F. HAYASHI AND J.D. SINCLAIR
f (bpm) 2OO
100% 0 2
100% 0 2
100~ O a
10~ O I
10(0, 0 a
Fig. 5. Changes of respiratory frequency (0 and tidal volume (VT) in individual rats in the intact and decerebrate state when inspired gas is switched from 100% O: to 5% COa/95% 02. Thick line indicates mean values. Data for effects of CO: combine studies summarised in tables 1 and 2.
evaluation but not control ofvariables. It might seem that the requirements of a complex control of respiratory level would be different in the rat from the cat and other species. However, most of then neural substrate for such control has now been demonstrated in the rat. Respiratory neurons of the parabrachial and Kolliker-Fuse nuclei of the pens have substantial connections with the premotor nuclei of the medulla, the nucleus of the tractus solitarius, ambiguus, and para-ambiguus, and with hypothalamus, the higher centres, and the spinal cord (see review of Bystrzycka and Nail, 1985). Although variations have been observed, for example in the extent of the dorsal respiratory group in the rat (Ezure et al., 1988), there is little reason to invoke species differences to explain unexpected experimental results. The experiments produced respiratory changes with many of the features seen after transection or iesioning in other species. The hyperventilation seen after decerebration has been attributed to loss of respiratory inhibition effected from rostrai brainstem structures (Hugelin, 1986). Hypocapnia with polypnea has been reported in paralysed vagotomised cats (Cohen, 1964) and in decerebrate unanesthetised cats, in which it was eliminated by mesencephalic transection (Tenney and Ou, 1977). Our results are similar
CHEMICAL INFLUENCES ON DECEREBRATE BREATHING
in many ways to those found in the cat by St John et al. (1975) after intercollicular decerebration and subsequent midpontine decerebration or lesioning of the medial parabrachial complex. The similarities must reflect hindbrain mechanisms that are so basic and robust that they remain manifest despite many differences between studies, for example in species used, in anesthetic agents, in levels of anesthesia, and in technical approaches. In particular, studies of rhythm generation in pens and medulla have usually involved vagotomised cats (St John, 1975, 1979; Hugelin, 1986). Our use of the rat excluded the possibility of bilateral vagotomy, which in this species leads rapidly to pulmonary oedema, preceded by changes in pulmonary mechanics which preclude valid measurements of changes in tidal volume and frequency (Martin-Body and Sinclair, 1987). This complication is avoided by deep anesthesia, but this would be incompatible with the purposes ofthe study. With vagi intact, pulmonary stretch receptor inputs remain intact. These have been shown to be substantial in the rat (Fukuda et al., 1982; Martin-Body and Sinclair, 1985) and are probably more so than in the cat. The vagus nerve also provides inputs of rapidly adapting receptors, of chemosensitive stretch receptors, and probably of secondary thoracic chemoreceptors as well as laryngeal inputs affecting respiration (Martin-Body et al., 1985; Martin-Body and Sinclair, 1985). In view of the substantial general role of vagal inputs in determining respiratory frequency, confirmed in the relevant experiments of Clark and yon Euler (1972), Fukuda et al. (1982) and Martin-Body and Sinclair (1987), the relative constancy of respiration rate demonstrated during hypercapnia is of particular note, especially when contrasted with the tachypnoea characterising hypoxia, even in our intact animals. Our arterial ligations produced substantial reductions of the ventilatory responses to hypercapnia and, to a lesser extent, hypoxia. The effects of CO2 were progressive between hypocapnic, eucapnic and hypercapnic states. St John and Wang (: 76) found similar changes in cats with midpontine transections. Our results give further evidence that the normal excitatory effects of both hypoxia and hypercapnia can be reduced or in some cases reversed to inhibitory effects, even though mechanisms for the generation of increases of f and VT remain operative. The inhibition of respiration in the carotid-body denervated rat exposed to hypoxia has recently been shown to depend on a neural mechanism localised in the rostral pens or mesencephalon (Martin-Body, 1988). Ischemia in our rats generally extended caudal to the critical level and a reduction of VI in hypoxia was seen in only one case. After CSN section, with the usual elimination ofthe response to hypoxia- in the rat, the aortic body and other peripheral chemoreceptors normally have little effect (Martin-Body et al., 1985) the mean response of VT increased, a result compatible with that of Martin-Body. An inhibitory influence of carbon dioxide on the minute volume of respiration has not been reported in comparable circumstances though the low mean levels of change of ventilation with hypercapnia in cats with combined rostral and caudal pontine lesions (St John and Wang, 1976) suggest that depression might have occurred in some cases. Complex inhibitory mechanisms originating in the cortex and Responses to hypercapnia and hypoxia.
F. HAYASHI AND J.D. SINCLAIR
reticular system of the mesencephalon are known to affect the premotor neurons of the medulla (Hugelin, 1986). Loss of these inhibitory mechanisms may account for the tachypnoea seen in our animals after decerebration, and also reported in awake rats after selective hindbrain lesioning (Sinclair et al., 1985) but cannot account for the slowing of respiration seen during hypercapnia. Kobayasi and Murata (1979) concluded that this effect demonstrated the lo~s of a suprapontine accelerating mechanism. This possibility is excluded by the demonstration that in hypoxia, a substantial acceleration of respiration still occurs. It is possible that direct inhibitory effects of hypercapnia on neural tissue (Jodkowski and Lipski, 1986) are producing depression, especially at the very high levels of Paco2 (up to 250 mmHg) used in the experiments of Kobayasi and Murata (1977). What should be noted is that the substantial reduction or loss of the stimulatory action of CO2 seen in these experiments was clearly not the result of ischemia of the ventrolateral medulla, which was specifically examined in all cases and was well away from the border between perfused and unperfused tissues; so that the changes could not be attributed to loss of the respiratory mechanism sited here, whether or not this is chemoreceptor (Loeschcke, 1982; Millhorn and Eldridge, 1986). Patterns of respiratory frequency and tidal volume. A striking result reported here is the variation between the changes of frequency and amplitude occurring in response to chemical stimulation. In the intact state, the resp.onse to hypoxia reflects the longstanding conclusion that hypoxia drives respiration via frequency. The response to hypercapnia further supports the conclusion of Florez and Borison (1967) that in this response the major variable is tidal volume. In our experiments, decerebration further emphasised the independence of response of f and VT, for with the reduced responses, or in a minority of cases with a depression of ventilation, there were mean decreases of VT in hypoxia and of f in hypercapnia. The residual increases of f in hypoxia and VT in hypercapnia indicated that the controlling mechanisms remained intact. This observation excludes the explanation offered by St John (1979) that the loss of f response after decerebration was the consequence of destruction of the parabrachial complex and the retention of the VT response indicated its dependence on neural mechanisms sited more caudally. Separations of changes of f and VT have also been produced recently in studies of the function of the ventral surface of the rabbit medulla (Homma et al., 1988). Our results offer conclusive evidence of a difference in responses to hypoxia and hypercapnia. A possible cause for such differences may lie in other physiological changes associated with one or other chemical disturbance. The increases of cerebral blood flow in mild hypercapnia, for example, must produce further changes in the chemical milieu. The release of endogenous opioids in hypercapnia (Gamble and Milne, 1990) might depress ventilation when excitatory influences are destroyed by decerebration, but such effects depress tidal volume when they occur in severe hypoxia (Neubauer et al., 1987) and thus act in the wrong direction to account for the changes of hypercapnia. The consistency of the effects through the various conditions we report suggest to us that hypoxia and hypercapnia do not act via a common pathway to increase the drive
CHEMICAL INFLUENCES ON DECEREBRATE BREATHING
of a respiratory pattern generator but act, at least in the rat, via two separate pathways. The model of respiratory control proposed by Richter (1986) offers a basis on which to explain our findings and to reconcile them with the findings of others. We propose that hypoxia acts on the reticular activating system and thus leads to excitation of inspiratory- ramp and early inspiratory - neurons (Richter et al., 1986). The immediate result is an increase of activity of inspiratory motoneurons, as can occur selectively in hypoxia (Sears et al., 1982) while the interconnections of medullary neurons (Richter et al., 1986) produce a shortening of the respiratory cycle, measured as an increase of frequency. In contrast, the responses to hypercapnia could be focussed on expiratory neurons of the network, and lead immediately to increased activity of expiratory motoneurons, a feature of the action of CO2 (Sears et ai., 1982 ). The reciprocal inhibition of inspiratory ramp neurons would lead to a prolongation of the respiratory cycle when not suppressed by pontine inputs that are a critical element of the model. The consistency and stability of the decerebrate preparation we describe give it a potential value for further study of these questions.
Acknowledgements. This study was supported by a grant from the Medical Research Council of New Zealand. Professor J. Lipski gave valuable advice concerning the manuscript.
References Bystrzycka, E.K. and B.S. Nail (1985). Brainstem nu~cleiassociated with respiratory, cardiovascular and other autonomic functions. In: The Rat Nervous System. Vol. 2, edited by G. Paxinos. Sydney, Academic Press, pp. 95-110. Clark, F.J. and C. yon Euler (1972). On the regulation of depth and rate of breathing. J. Physiol. (London) 222: 267-295. Cohen, M. !. (1964). Respiratory periodicity in the par alyzed, vagotomized cat: hypocapnic hyperpnea. Am. J. Physiol. 206: 845-854. Ezure, K., M. Manabe and H. Yamada (1988). Distribution ofmedullary respiratory neurons in the rat. Brain Res. 455: 262-270. Florez, J. and H.L. Borison (1967). Tidal volume in CO2 regulation: peripheral denervations and ablation of area postrema. Am. J. Physiol. 212: 985-991. Fukuda, Y., W. R. See and Y. Honda (1982). Effect of halothane anaesthesia on end-tidal Pco.~ and pattern of respiration in the rat. Pfliigers Arch. 392: 244-250. Gamble, G. D. and R.J. Milne (1990). Hypercapnia depresses noeiception: endogenous opioids implicated. Brain Res. 514: 198-205. / Gautier, H. (1976). Pattern of breathing during hypoxia or hypercapnia of the awake or anesthetized cat. Respir. Physiol. 27: 143-206. Homma, I., A. Isobe, M. lwase, A. Kanamaru and M. Sibuya (1988). Two different types ofapnea induced by focal cold block of ventral medulla in rabbits. Neurosci. Lett. 87: 41-45. Hugelin, A. (1986). Forebrain and midbrain influence on respiration. In: Handbook of Physiology. Section 3, Vol. II, edited by Bethesda, Am. Physiol. Soc., pp. 69-92. Jodkowski, J. S. and J. Lipski (1986). Decreased excitability of respiratory motoneurons during hypercapnia in the acute spinal cat. Brain Res. 386: 296-304. Kobayasi, S. and K. Murata (1979). Phrenic activity during severe hypercapnia in vagotomised rabbits. ,J. Appl. Physiol. 47: 91-97.
F. HAYASHI AND J.D. SINCLAIR
Loeschcke, H.H. (1982). Central chemosensitivity and the reaction theory. J. Physiol. (London) 332: 1-24. Martin-Body, R.L. and J.D. Sinclair (1985). Analysis of respiratory patterns in the awake and in the halothane anaesthetised rat. Respir. PhysioL 61: 105-113. Martin-Body, R. L., G. J. Robson and J. D. Sinclair (1985). Respiratory effects ofsectioning the carotid sinus, glossopharyngeal and abdominal vagal nerves in the awake rat. J. Physiol. (London) 361: 35-45. Martin-Body, R.L. and J.D. Sinclair (1987). Differences in respiratory patterns after acute and chronic pulmonary denerva~ion. Respir. Physiol. 70: 205-219. Martin-Body, R. L. (1988). Brain transections demonstrate the central origin of hypoxic ventilatory depression in carotid body denervated rats. J. Physiol. (London) 407" 41-52. Millhorn, D.E. and F.L. Eldridge (1986). Role of ventrolateral medulla in regulation of respiratory and cardiovascular systems. J. Appl. Physiol. 61: 1249-1263. Neubauer, J.A., M.A. Posner, T.V. Santiago and N.H. Edelman (1987). Naloxone reduces ventilatory depression of brain hypoxia. J. Appl. Physiol. 63: 699-706. Paxinos, G. and C. Watson (1982). The Rat Brain in Stereotaxic Coordinales. Sydney: Academic Press. Pollock, L.J. and L.E. Davis (1923). Studies in decerebration. Arch. Neurol. Psychol. 10: 391-398. Richter, D. W., D. Ballantyne and J. E. Remmers (1986). How is the respiratory rhythm generated? A model. News in Physiol. Sci. 1: 109-112. Saether, K., G. Hilaire and R. Monteau (1987). Dorsal and ventral respiratory groups of neurons in the medulla of the rat. Brain Res. 419: 87-96. Sears, T.A., A.J. Berger and E.A. Phillipson (1982). Reciprocal tonic activation of inspiratory and expiratory motoneurones by chemical drives. Nature 229: 728-730. Sinclair, J. D., W. M. St John and D. Bartlett (1985). Enhancement of respiratory response to carbon dioxide produced by lesioning caudal regions of the nucleus of the tractus solitarius. Brain Res. 336:318-320. St John, W.M. (1975). Differing responses to hypercapnia and hypoxia following pneumotaxic center ablation. Respir. Physiol. 23: I-9. St John, W.M. and S.C. Wang (1976). Integration of chemoreceptor stimuli by caudal pontile and rostral medullary sites. J. Appl. Physiol. 41: 612-622.
St John. W.M. (1979). Differential alteration by hypercapnia and hypoxia of the apneustic respiratory pattern in decerebrate cats. J. Physiol. (London) 287: 467-491. Tenney, S. M. and L. C, Ou (1977). Ventilatory response ofdeeorticate and deeerebrate cats to hypoxia and CO,. Respir. Physiol. 29:8 !-92.