Ventilatory CO₂ Drive in the Tortoise Testudo Horsfieldi

That CO₂ breathing induces hyperventilation in chelonians is well established (Lumsden, 1924; Randall, Stullken & Hiestand 1944; Frankel et al. 1969; Jackson, Palmer & Meadow, 1974; Glass, Burggren & Johansen, 1978). Nevertheless very little is known about the receptors that mediate these responses. Frankel et al. (1969) give inconclusive evidence for the presence of peripheral chemoreceptors at the bifurcation of the common carotid artery in Pseudemys scripta but the histological evidence was established in only a few of the animals examined. Subsequently, Milsom & Jones (1975, 1976) concluded that P. scripta possesses intrapulmonary CO₂-sensitive re-ceptors while Hitzig & Jackson (1978) provide evidence that there are central chemon receptors sensitive to changes in the acid-base status of the cerebrospinal fluid. In a previous study (Benchetrit, Armand & Dejours, 1977), we found some evidence for the existence of Heymans-type chemoreceptor structures, perfused in pulmonary arteries, and which are innervated by branches of the vagus nerves. Oxygen tests and sodium cyanide injections were used to demonstrate the ventilatory effect of chemoreceptor stimulation.

The present study investigates: (1) whether the inspired oxygen tension influences the level of CO₂-induced hyperventilation and (2) how vagotomy affects CO₂-induced hyperventilation. To this purpose, the ventilatory effects of normoxic and hyperoxic hypercapnia were observed in intact and bivagotomized conscious tortoises.

The experiments were carried out on 4 conscious tortoises (Testudo horsfieldi), two males and two females weighing 550–620 g. One day before experimentation, the animals were fitted with experimental devices (Benchetrit, Armand & Dejours, 1977) to record the pneumotachogram, the spirogram and the intrapulmonary pressure (Fig. 1). A mid-neck bivagotomy was performed under general anaesthesia (subcutaneous injection of 50 mg.kg⁻¹ of pentobarbital). Recordings were made from the bivagotomized animals after total recovery from anaesthesia, that is at least 24 h later. One animal was kept for 3 weeks before showing signs of pulmonary oedema. The other three animals were killed by an overdose of anaesthetic at the end of the experimental period 8–10 days following bivagotomy. They did not lose weight during this period.

The animals were placed in a plastic box with a hole drilled in front of the animal’s head. The distal end of the pneumotachograph protruded from the box and the inspired gas flowed in through a T-tube fitted to it. The animals were loosely restrained. Their state of relaxation could be judged from the intrapulmonary pressure recorded via an implanted catheter. The animals were maintained and the experiments carried out at an ambient temperature of 23–25 °C; the temperature of the box was continuously recorded. Ventilation was recorded for 1 h while the animals inhaled air or oxygen and then for 1 h while the hypercapnic mixture flowed in the T-tube, after which the animals again breathed the original CO₂-free gas. A single experiment was performed per day on each animal so that a minimum of 20 h separated two successive hypercapnic exposures.

The amplified pneumotachographic signal was digitalized and processed on line by a laboratory mini-computer. For each cycle the available data were: V_T (tidal volume), T_E (expiratory duration), T_Ap (inspiration duration) and (apneic plateau duration). From the total cycle duration T_T=T_E+ T₁+ T_Ap, the respiratory frequency f = 60/T and the minute ventilation were computed. These data were continuously displayed during the experiment and then stored for off-line processing. The tidal volume was the measured inspiratory volume. This was not always equal to the immediately preceding expiratory volume, but the mean values over several cycles were similar.

An examination of the different measured variables during CO₂ inhalation showed that they reached a steady state after 30 min. Thus the mean values of the variables during the second h of inhalation of CO₂-enriched mixtures were compared to those of the preceding period of air-or O₂-breathing.

The expiratory phase was immediately followed by an inspiratory phase and then by an apneic plateau (Fig. 1). Table 1 shows the characteristics of the breathing pattern in one tortoise breathing air. The variations in T_T, T_E and V_T were about 30%, judged by the coefficient of variation (s.D./mean), but the apneic period showed greater fluctuations. The inspiratory phase was always longer than the expiratory one. The relative contribution of each of these three phases to the total period is given in Table 1. Inhalation of pure O₂ did not change this breathing pattern. There was, however, a very slight decrease in ventilation in all animals, due to decrease in tidal volume.

The breathing of hypercapnic mixtures led to an increase in ventilation. With a 2 % CO₂ mixture, ventilation increased because of a rise in respiratory frequency. With a higher inhaled CO₂ concentration, however, both V_T and/contributed to the observed increase in ventilation (Fig. 2). The increase in respiratory frequency was achieved mainly by the shortening of the apneic plateau durations (Table 2), and during the steady state of response to hypercapnia very long-lasting apneic periods of 300 s or more were not observed. In a steady state response to 4% CO₂ breathing, extreme values of 3·3 and 41·7 s were recorded, compared to 8·7 and 341 s during the preceding control period.

Although the ventilatory responses to breathing O₂-CO₂ mixtures were less than those to breathing air-CO₂ mixtures (Fig. 3) the pattern of the response was similar: at 2% CO₂ only respiratory frequency increased, but at higher concentration both V_T and f increased.

Bilateral vagotomy led to slow deep breathing in tortoises. The respiratory frequency was between and of the control value. The tidal volume was, however 2–3 times greater, thus causing a slight decrease in ventilation. The lower respiratory frequency was associated with a lengthening of T_1,T_E and, especially, of T_Ap. Apneic periods as long as 1200 s were observed. The relative contribution of T_Ap to the total period was greater than in intact animals (Table 1). Pure oxygen inhalation did not change this pattern of breathing; however, ventilation was slightly higher during hyperoxia because of a greater tidal volume.

The breathing of hypercapnic mixtures led to an increase in ventilation in bi-vagotomized animals. This increase resulted primarily from an increase in respiratory frequency, for an increase in tidal volume was obvious only in 4% CO₂ (Fig. 2). Similarly, during hyperoxic hypercapnia the respiratory frequency rose in 2 and 3 % CO₂, and both V_T and/were increased in 4% CO₂ (Fig. 2 and Table 2). Bivagotomy may induce lung damage and so modify lung mechanics. Comparison between responses to hypercapnia in intact and bivagotomized animals may therefore be inappropriate. It thus seems more pertinent to compare hypercapnic responses in bivagotomized animals during normoxia and hyperoxia. Our results (Figs. 2, 3 and Table 2) show that the CO₂-induced hyperventilation was the same in normoxic and hyperoxic bivagotomized animals.

The observation that a CO₂-induced hyperventilation is lower in hyperoxia than during normoxia indicates that O₂ and CO₂ have opposite effects on the control of ventilation in the tortoise. Such opposite effects are observed at the level of the peripheral chemoreceptors in mammals (Heymans & Neil, 1958) and birds (Bouverot & Leitner, 1972). Similar receptors, sensitive to O₂ and to CO₂ may also exist in tortoises, but the possibility of separate receptors cannot be ruled out. Even if Heymans-type receptors exist in the tortoise, it is probable that they are not the only CO₂-sensitive, vagally-innervated, receptors. The observation that hyperventilation in hyperoxic intact animals was higher than in bivagotomized animals, suggests that not all the thoracic component of the CO₂ drive is O₂-sensitive. However, as we have emphasized, the pulmonary mechanics may be altered following bivagotomy. Furthermore, the difference between the ventilatory responses to CO₂ in the intact and bivagotomized animals is not sufficiently evident to allow unequivocal conclusions as to the type of receptor involved.

It is conceivable then that mammalian-like O₂-and CO₂-sensitive chemoreceptors may exist in the tortoise. Oxygen-dependent, CO₂-induced, hyperventilation has also been observed in chelonians by Glass et al. (1978) and in lizards by Nielsen (1962).

The persistence of a CO₂-induced hyperventilation after bilateral vagotomy suggests that in the tortoise the ventilatory reactions to CO₂ result from two drives: one of thoracic origin, and vagally mediated, the other extra-thoracic and probably central. The central drive does not seem to be affected by O₂. These results are comparable to the reactions to CO₂ in mammals (Bernards, Dejours & Lacaisse, 1966), there being an interplay of two drives (one peripheral and one central); the peripheral component being affected by the level of oxygen and the central one being O₂-independent. The existence of central chemoreceptors has been reported recently by Hitzig & Jackson (1978) in semi-aquatic turtles.

The breathing pattern in tortoises includes a particular phase: the apneic plateau. Changes in tortoise ventilation are mainly due to changes of the duration of the apneic plateau. The hyperventilation of intact animals breathing 2% CO₂ in air resulted primarily from the shortening of the apneic plateau. At higher CO₂ concentrations, the increase in f was always greater than the increase in V_T. T₁ and T_E were very little changed by CO₂ breathing (Table 2) in hyperoxic or bivagotomized animals. According to Gans & Hughes (1967), there is no inspiratory activity during the apnea apart from closure of the glottis. But glottal closure is not necessary for the apnea: we have observed apnea with an open glottis in deeply anaesthetized animals. However, comparison of the breathing patterns of intact and bilaterally vagotomized animals shows an apparent relationship between the vagally-mediated information and the length of the apnea: The fall in respiratory frequency following bivagotomy resulted mainly from an increase in T_Ap, and, to a lesser degree, in T₁ and T_E (table 1).

As some chelonians are aquatic, it could be that the apneic plateau corresponds to a need for coordination between periods of air breathing and immersion. This may also explain the differences in the pattern of hyperventilation between tortoises and turtles (e.g. Pseudemys scripta) at nearly the same temperature (Jackson et al. 1974). The latter authors observed a more prominent contribution of V_T to the ventilatory increase. However, the sensitivity to CO₂ in our tortoises can hardly be compared to that given in other studies, as we have only measured the inspiratory fraction of CO₂ (Dejours et al. 1965). Moreover, as emphasized by Jackson, Allen & Strupp, (1976), non-pulmonary surfaces contribute to CO₂ loss in a way which can be correlated with the habitat of the various species. It should also be emphasized that the control of ventilation in chelonians is temperature-dependent (Jackson et al. 1974; Robin, 1962). Our study has been carried out at a constant temperature, but the temperature dependence of the peripheral chemoreceptor responses has been demonstrated in tortoises (Benchetrit et al. 1977), and Hitzig & Jackson (1978) reported that the central chemoreceptor sensitivity is temperature-dependent in Pseudemys scripta.