Effects of CO2 and H+ on laryngeal receptor activity in the perfused larynx in anaesthetized cats

Authors

  • Z.-H. Wang,

    1. Department of Human Anatomy and Physiology, University College Dublin, National University of Ireland, Dublin 2 and
    Search for more papers by this author
  • A. Bradford,

    Corresponding author
    1. Department of Physiology, Royal College of Surgeons in Ireland, St Stephen's Green, Dublin 2, Ireland
    • Corresponding author
      A. Bradford: Department of Physiology, Royal College of Surgeons in Ireland, St Stephen's Green, Dublin 2, Ireland. Email: abradfor@rcsi.ie

    Search for more papers by this author
  • R. G. O'Regan

    1. Department of Human Anatomy and Physiology, University College Dublin, National University of Ireland, Dublin 2 and
    Search for more papers by this author

Abstract

  • 1Intralaryngeal CO2 reflexly decreases ventilation and increases upper airway muscle activity. Topical anaesthesia of the laryngeal mucosa or cutting the superior laryngeal nerves (SLNs) abolishes these reflexes, indicating that the receptors responsible are superficially located and that their afferent fibres are in the SLN. Intralaryngeal CO2 affects the activity of receptors recorded from the SLN.
  • 2An isolated, luminally perfused laryngeal preparation was developed in anaesthetized, paralysed cats in order to compare the effects of solutions with varying levels of pH and PCO2 on pressure-sensitive laryngeal receptor activity. Since the pH of tracheal surface fluid is reported to be approximately 7.0, two neutral (pH 7.4 and 7.0) and two acidic (pH 6.8 and 6.3) solutions were used.
  • 3Compared with neutral acapnic control solutions, neutral hypercapnic (PCO2 64 mmHg) solutions either excited or inhibited the discharge of 113 out of 211 pressure-sensitive SLN afferents. In 24 receptors, the effects of hypercapnic solutions with either neutral or acidic pH were similar in both direction and magnitude. In 50 receptors affected by neutral hypercapnic solutions, acidic acapnic solutions had no effect on 66 % of units and significantly smaller effects in the remaining units. In 17 receptors, the effects of neutral solutions with a PCO2 of 35 mmHg were significantly less than for neutral solution with a PCO2 of 64 mmHg.
  • 4These results show that the effects of CO2 on laryngeal pressure-sensitive receptors are independent of the pH of the perfusing media, and suggest that acidification of the receptor cell or its microenvironment is the main mechanism of CO2 chemoreception.

Intralaryngeal CO2 reflexly enhances the activity of upper airway dilating muscles and decreases ventilation (Boushey & Richardson, 1973; Lee et al. 1986; Nolan et al. 1990; Bartlett et al. 1992), effects that can also be elicited reflexly by laryngeal negative pressure (Mathew et al. 1982; Mathew & Farber, 1983; Mathew, 1984; Van Lunteren et al. 1984; Sant'Ambrogio et al. 1985; Leiter & Daubenspeck, 1990; Horner et al. 1991; Mezzanotte et al. 1992; Zhang & Mathew, 1992). Topical anaesthesia of the laryngeal mucosa or section of the superior laryngeal nerves (SLNs) abolishes these reflexes, thus indicating that the receptors responsible are superficially located and that their afferent fibres travel in the SLNs. It has been proposed that laryngeal CO2 and negative pressure may work synergistically to promote upper airway patency (Anderson et al. 1990; Nolan et al. 1990). Indeed, intralaryngeal CO2 has been shown to affect the activities of laryngeal pressure-sensitive receptors as recorded from afferent units of the SLN (Anderson et al. 1990; Bartlett & Knuth, 1992; Bradford et al. 1993, 1998; Ghosh & Mathew, 1994). In particular, receptors stimulated by negative pressure, which have a low discharge rate when the larynx is unventilated, tend to be excited by increased levels of intralaryngeal CO2, whereas pressure-sensitive receptors of other categories are mainly inhibited (Bradford et al. 1993, 1998; Ghosh & Mathew, 1994).

To elucidate the mechanisms of CO2-induced effects on laryngeal receptor activity, an isolated, in situ, luminally perfused laryngeal preparation was developed in which the superficially located laryngeal receptors were subjected to different solutions of known compositions, and to sinusoidal fluctuations of pressure simulating a breathing cycle. To test the hypothesis that CO2 alters receptor activity by acidifying the receptor's microenvironment, the effects of changes in PCO2 and pH of the perfusing media on the activities of laryngeal pressure-sensitive receptors were studied. To do this, four different solutions were used: (1) neutral acapnic solution with a pH of 7.4 or 7.0, equilibrated with air, (2) acidic hypercapnic solution with a pH of 6.8 or 6.3, equilibrated with 9 % CO2, (3) neutral hypercapnic solution with a pH of 7.4 or 7.0, equilibrated with 9 % CO2, and (4) acidic acapnic solution with a pH of 6.8 or 6.3, equilibrated with air.

METHODS

Experiments were performed on 35 cats of either sex weighing 1.0-3.5 kg. Anaesthesia was induced by i.p. injection of pentobarbitone sodium (Sagatal, May & Baker, Dagenham, Essex, UK; 40-48 mg kg−1). Surgical anaesthesia was maintained by supplementary i.v. injections of 6-12 mg pentobarbitone sodium every hour. The animals were secured in the supine position and rectal temperature maintained at 37°C using a rectal probe and a thermostatically controlled heating blanket (Harvard Apparatus Ltd, Edenbridge, Kent, UK). Arterial blood pressure was constantly monitored (Statham P23, Hato Rey, Puerto Rico) and arterial blood samples were withdrawn periodically for determination of blood gas and acid-base status (Ciba Corning Diagnostics Ltd, Halstead, Essex, UK). The larynx and the cervical trachea were exposed and a low tracheostomy was performed for the ventilation of the animal's lungs. The animals were paralysed with pancuronium bromide (Sigma, 0.8 mg i.v. as required) and artificially ventilated to maintain arterial blood gas tension and pH values within the physiological range (pH 7.3-7.4, PO2 75-85 mmHg, PCO2 30- 40 mmHg) and to maintain end-tidal CO2, monitored using an infrared CO2 meter (Engstrom Eliza Duo, Gambro, Sweden), at approximately 4 %. The depth of anaesthesia was assessed by continuously monitoring blood pressure and heart rate. Animals were killed at the end of experiments with an overdose of the anaesthetic.

Surgical preparation for laryngeal perfusion

A cannula was inserted rostrally into the trachea via a higher tracheostomy, secured by a tie and its tip positioned just below the cricoid cartilage (Fig. 1). It served as the inflow cannula for the perfusion of the larynx. A fluid-filled polyethylene catheter was inserted through a side arm of this cannula to pass through the glottic opening so that its tip lay in the laryngeal lumen cranial to the vocal cords. This catheter served to continuously monitor intralaryngeal pressure (Statham P23, Hato Rey, Puerto Rico). The oesophagus was tied at its junction with the pharynx. An oral cannula (inside diameter 0.7 cm) was inserted so as to position its distal end just below the tip of the epiglottis, and secured in place by a tie encircling the pharynx. To position this tie, the transverse vein was cut on either side between ligatures. The lateral wall of the pharynx was then carefully dissected free from surrounding tissues at and above the level of the hyoid bone. The back of the pharyngeal wall was approached by dissecting along the lateral pharyngeal wall dorsally and cranially to reach the base of the skull in the region of the tympanic bulla. When the dissection was completed on both sides of the neck, the posterior wall of the pharynx could be readily separated from the base of the skull behind. These procedures allowed a tie to be passed around the pharyngeal wall at the level of the hyoid bone and minimized distortion of the supralaryngeal airway when the tie sealed the pharyngeal wall closely around the the tip of the oropharyngeal cannula. In preliminary experiments, efforts were made to seal the oral cannula into the pharynx by inflating a Foley catheter. However, this did not produce an effective seal and resulted in much tissue distortion. When the larynx was perfused in a cranial direction, the solutions passed through the upper tracheal cannula, entered the larynx and part of the hypopharynx and were drained away through the oral cannula. During the initial perfusion, the larynx and hyoid region were manipulated in order to remove bubbles during the replacement of air by fluid.

Figure 1.

Diagrammatic representation of the method of perfusion and pressure application of the isolated in situ laryngeal preparation

During perfusion, clamps C1 and C2 were open, while C3 was closed. During pressure application, C1 and C2 were closed, while C3 was open. The modified ventilator provided sinusoidal waves of pressure to the larynx and C4, an adjustable screw clip, was used to control the pressure magnitude. Intralaryngeal pressure was monitored by a pressure transducer (P) through a catheter inserted into the larynx.

Arrangements for laryngeal perfusion and pressure application

The perfusing solutions were maintained at 37°C in an organ bath. The delivery tubes for the perfusates had outer tubings that acted as Baker coils, within which water at 38°C was pumped constantly to prevent cooling of the solutions on their way to the larynx. The temperature within the larynx was monitored in two experiments using a thermistor microprobe (Physitemp Instruments Inc., Clifton, NJ, USA) inserted through the outflow tube and caudally into the larynx. The registered temperature was 35-37.5°C during trial periods. When changing perfusion solutions, a period of 10-15 s was required for the solution to reach the larynx, as determined by timing the arrival of Methylene Blue dye at the entrance to the larynx.

The oral cannula was also connected through a side arm to a rodent ventilator (Model 683, Harvard Apparatus Ltd), which was modified to subject the larynx to sinusoidal waves of pressure oscillating around zero (see Figs 1, 2, 3 and 4). The ventilator was modified by connecting its outlet and inlet ports to a T-tube (Fig. 1). The degree of opening of the side arm of the T-tube could be altered by an adjustable screw (C4, Fig. 1). The amplitude of the pressure waves provided by the ventilator could be controlled by adjusting this screw and/or by adjusting the stroke volume of the ventilator. During laryngeal perfusion, the connection between the ventilator and the oral cannula (C3) was clamped and laryngeal pressure was maintained at atmospheric pressure. During pressure wave application (± 3-10 cmH2O), the perfusion was momentarily interrupted by clamping both the inflow (C1) and outflow (C2) cannulae, and the connection between the outflow cannula and the ventilator was opened (C3).

Figure 2.

Responses of a quiescent fibre stimulated by negative pressure to perfusion with neutral hypercapnic solution

A, neutral acapnic solution (pH 7.4, PCO2 0 mmHg). B, neutral hypercapnic solution (pH 7.4, PCO2 64 mmHg). C, neutral acapnic solution again. Signals are laryngeal airway pressure (Paw) and neurogram (NG). Perfusion with neutral hypercapnic solution consistently increased nerve discharge during pressure wave application.

Figure 3.

A comparison of the effects of neutral hypercapnic and acidic acapnic solution on a tonically active fibre stimulated by negative pressure

A, perfusion with neutral acapnic solution (pH 7.4, PCO2 0 mmHg). B, perfusion with neutral hypercapnic solution (pH 7.4, PCO2 64 mmHg). C, perfusion with neutral acapnic solution again. D, perfusion with acidic acapnic solution (pH 6.8, PCO2 0 mmHg). E, perfusion with neutral acapnic solution again. Signals are laryngeal airway pressure (Paw) and neurogram (NG). The inhibitory effect of acidic acapnic solution on receptor activity was slight, and much less than that of neutral hypercapnic solution.

Figure 4.

Responses of a quiescent fibre stimulated by negative pressure to perfusion with acidic hypercapnic and neutral hypercapnic solution

A, neutral acapnic solution (pH 7.4, PCO2 0 mmHg). B, acidic hypercapnic solution (pH 6.8, PCO2 64 mmHg). C, neutral acapnic solution again. D, neutral hypercapnic solution (pH 7.4, PCO2 64 mmHg). E, neutral acapnic solution again. Signals are laryngeal airway pressure (Paw) and neurogram (NG). Perfusion with acidic hypercapnic solution and neutral hypercapnic solution caused similar increases in nerve discharge during pressure wave application.

Composition of the perfusing solutions

Overall, two sets of experiments were performed using identical protocols but different perfusion solutions. The making of the solutions was based on Krebs-Henseleit solution. In the first set of experiments (Set 1), a pH value of 7.4 was taken as neutral with reference to the pH of blood plasma. In the second set of experiments (Set 2), a pH value of 7.0 was taken as neutral. This was done because laryngeal receptors may have an environment with a pH value of approximately 7.0. The pH of tracheal airway surface fluid has been reported to be close to 7.0 (Kyle et al. 1990). The pH of the test solutions was achieved by adjusting the amount of bicarbonate buffer used while their isotonicity was maintained by varying the amount of NaCl. In neutral acapnic solution, however, the sodium salt of Hepes (N-2-hydroxyethyl piperazine-N‘-2-ethanesulphonic acid) replaced bicarbonate buffer. The per cent CO2 and PCO2 of the equilibration gas and the pH of the various solutions are shown in Table 1, and the composition of the solutions is given in Table 2 and Table 3. In Set 1 experiments, the four solutions used were: (1) neutral acapnic solution with a pH of 7.4, equilibrated with air, (2) acidic hypercapnic solution with a pH of 6.8, equilibrated with 9 % CO2, (3) neutral hypercapnic solution with a pH of 7.4, equilibrated with 9 % CO2, and (4) acidic acapnic solution with a pH of 6.8, equilibrated with air. In Set 2 experiments, the four solutions were: (1) neutral acapnic solution with a pH of 7.0, equilibrated with air, (2) acidic hypercapnic solution with a pH of 6.3, equilibrated with 9 % CO2, (3) neutral hypercapnic solution with a pH of 7.0, equilibrated with 9 % CO2, and (4) acidic acapnic solution with a pH of 6.3, equilibrated with air. In some experiments, a solution equilibrated with 5 % CO2 at pH 7.0 and a solution equilibrated with air at pH 4.9 were also tested. The pH, PCO2 and PO2 values of the solutions were measured frequently throughout each experiment using a blood gas analyser (Ciba Corning Diagnostics Ltd) and deviations from the desired values were corrected. However, corrections were rarely necessary since values usually remained stable once equilibrium had been reached.

Table 1. The pH, per cent CO2 and PCO2 of the equilibration gas of the perfusing solutions
 Neutral acapnicAcidic hypercapnicNeutral hypercapnicAcidic acapnic
Set 1    
  pH7.46.87.46.8
  Per cent CO20990
  PCO2 (mmHg)064640
Set 2    
  pH7.06.37.06.3
  Per cent CO20990
  PCO2 (mmHg)064640
Table 2. Composition of solutions (Set 1)
 Neutral acapnicAcidic hypercapnicNeutral hypercapnicAcidic acapnic
  1. Values are in millimoles, except for osmotic pressure, which is in milliosmoles.

NaCl137130100  138
KCl  4.7  4.7  4.7  4.7
CaCl2  2.2  2.2  2.2  2.2
MgSO4  1.2  1.2  1.2  1.2
NaH2PO4  1.2  1.2  1.2  1.2
Dextrose11.111.111.111.1
NaHCO310.039.0  1.5
Hepes  2.5
Osmotic pressure311.9311.9309.9310.9
Table 3. Composition of solutions (Set 2)
 Neutral acapnicAcidic hypercapnic (9% CO2)Neutral hypercapnic(9% CO2)Neutral hypercapnic(5% CO2)Acidic acapnic
  1. Values are in millimoles, except for osmotic pressure, which is in milliosmoles.

NaC137136124130138
KCl  4.7  4.7  4.7  4.7  4.7
CaCl2  2.2  2.2  2.2  2.2  2.2
MgSO4  1.2  1.2  1.2  1.2  1.2
NaH2PO4  1.2  1.2  1.2  1.2  1.2
Dextrose11.111.111.111.111.1
NaHCO3  2.815.0  8.5  0.5
Hepes  1.5
Osmotic pressure308.9309.5309.9308.9308.9

Laryngeal receptor activity

Laryngeal receptor activity was recorded from the internal branch of the SLN using conventional techniques. Fine filaments of nerve fibres were dissected from the distal cut end and placed on a pair of fine silver electrodes mounted on a micromanipulator. The nerve impulse signal was filtered and amplified (Grass P16, Grass Instrument Company, Quincy, MA, USA). Nerve filaments exhibiting discharge were further dissected to yield activity from a single fibre, or a few fibres easily distinguished by the differences in spike heights. Nerve activity was recorded, together with laryngeal pressure, on an electrostatic recorder (Type ES1000, Gould Electronics, Cleveland, OH, USA).

Protocols

Nerve discharge was recorded when the larynx was perfused with neutral acapnic solution, and fibres were classified as quiescent or tonic according to their discharge frequency. Quiescent fibres had sparse, irregular or no discharge, while tonic fibres had constant regular discharge during perfusion. The larynx was then subjected to sinusoidal pressure waves to assess the receptor's responsiveness to pressure. Since some fibres had no discharge during perfusion, each fibre that was placed on the electrodes was tested with pressure waves so that quiescent fibres with no baseline activity were included in the study. Fibres showing no consistent responses to pressure were discarded. The following groups of receptors were thus identified: quiescent units stimulated by negative pressure (QN) or positive pressure (QP), and tonically active units stimulated by negative pressure (TN) or positive pressure (TP). The responses of laryngeal receptors to CO2 and to alterations of pH of the perfusates were assessed by comparing nerve discharge during the application of a test solution with that during the control periods when neutral acapnic solution was perfusing the larynx. Each test trial consisted of successive perfusions with neutral acapnic solution, a test solution and neutral acapnic solution again. Pressure waves with a magnitude of ±3-10 cmH2O and a frequency of 30 cycles min−1 were applied following each period of perfusion. The second control period allowed recovery of receptor activity from any test-induced effect, and prepared the receptor for another test trial. The perfusion of neutral acapnic solution lasted between 2 and 5 min and the perfusion of test solutions lasted for up to 3 min.

Data analysis

The mean discharge of a receptor, in impulses per pressure cycle during a test period was compared with the control period immediately before and after the test. Differences in the discharge were assessed by Student's unpaired, two-tailed t tests. An observed change in fibre activity was considered to be due to the effect of the test solution only when the differences in receptor activity between the test and the two control periods were in the same direction and both significant (P < 0.05). Per cent changes in activity were calculated by comparing the test activity with the average of the activity of the two control periods. When a receptor was tested with two test solutions, the change in nerve discharge from control caused by each solution was calculated. For all the fibres that were tested with the same two test solutions, the difference in excitation or inhibition caused by the two solutions was assessed by a paired t test.

RESULTS

Effects of neutral hypercapnic solutions

The effects of neutral hypercapnic solution were studied in 211 SLN afferent fibres, 113 of which showed significant changes to increased intralaryngeal CO2 levels. These CO2-sensitive afferents consisted of the following: (1) 57 QN units, 39 of which were excited and 18 inhibited by CO2; (2) 22 QP units, 10 of which were excited and 12 inhibited by CO2; (3) 16 TN units, all of which were inhibited by CO2; (4) 18 TP units, of which 17 were inhibited and one excited by CO2. Two or more CO2 trials were performed on each fibre in order to confirm its CO2 sensitivity or to compare the effects of different solutions. For the majority of CO2-responsive receptors, the CO2 effects were not obvious during perfusion of the CO2-containing solution alone, but became apparent only during the pressure wave application that followed. The quantitative effects of neutral hypercapnic solution on laryngeal afferents are summarized in Table 4. Since the magnitude of the responses from Set 1 and Set 2 experiments was similar, these data were pooled together. An example of the excitatory effect of neutral hypercapnic solution on a QN unit is shown in Fig. 2 and an example of the inhibitory effect of neutral hypercapnic solution on a TN unit is shown in Fig. 3.

Table 4. Effect of neutral hypercapnic solution (pH 7.4 or 7.0, equilibrated with 9% CO2) on fibre activity compared with neutral acapnic solution (pH 7.4 or 7.0, equilibrated with air)
 Excited by CO2Inhibited by CO2Unaffected by CO2
Receptor n Per cent increase in discharge n Per cent decrease in discharge n
  1. For each of the fibres excited or inhibited by CO2, the discharge frequency of the fibre during laryngeal application of neutral hypercapnic solution was significantly higher or lower than during the application of neutral acapnic solution. QN, quiescent units stimulated by negative pressure; QP, quiescent units stimulated by positive pressure; TN, tonically active units stimulated by negative pressure; TP, tonically active units stimulated by positive pressure. Values are means with s.e.m. in parentheses.

QN3991.6 (10.5)1841.7 (4.1)52
QP1081.9 (18.5)1250.9 (9.0)18
TN1633.1 (7.4)16
TP111.61725.3 (5.8)12

The effects of neutral hypercapnic solution on laryngeal receptors were directly related to the CO2 concentration of the solutions. In 17 receptors (6 experiments), the CO2 effects were compared between neutral solution equilibrated with 9 % CO2 and neutral solution equilibrated with 5 % CO2, and were found to be significantly different. Thus for the 6 units excited by CO2 (5 QN and 1 QP), neutral solution equilibrated with 9 % CO2 caused an increase in discharge of 70.5 ± 17.2 % (mean ±s.e.m.), while neutral solution equilibrated with 5 % CO2 caused an increase of 33.7 ± 10.5 %. For 11 receptors inhibited by CO2 (1 QN, 3 QP, 4 TN and 3 TP), neutral solution equilibrated with 9 % CO2 caused a reduction in activity of 28.3 ± 9.3 %, while neutral solution equilibrated with 5 % CO2 caused a reduction of 15.1 ± 8.7 %.

Effects of acidic hypercapnic solutions

The effects of the neutral hypercapnic solution and the acidic hypercapnic solution were compared in 24 fibres obtained from 14 experiments. Of these fibres, 18 were obtained from Set 1 experiments while the remaining 6 were obtained from Set 2 experiments. As the results from Set 1 and Set 2 experiments were similar, they were pooled together. The responses of each fibre to the neutral and acidic solutions were similar in both direction and magnitude. The differences in the changes in activity caused by the two solutions were not statistically significant. The per cent changes (mean ±s.e.m.) in activity caused by acidic hypercapnic solution and neutral hypercapnic solution were 78.2 ± 16.9 and 79.6 ±15.7 %, respectively, for fibres excited by CO2, and 38.9 ± 11.8 and 42.4 ± 11.9 % for fibres inhibited by CO2. An example of the responses of a QN unit to acidic hypercapnic solution and neutral hypercapnic solution is shown in Fig. 4.

Effects of acidic acapnic solutions

The effects of the acidic acapnic solution (pH 6.8 or 6.3) were compared with that of the neutral hypercapnic solution (pH 7.4 or 7.0) on 25 CO2-sensitive fibres.

Fifteen receptors were obtained from Set 1 experiments and 10 from Set 2 experiments. All of these fibres discharged significantly more (excitation by CO2) or less (inhibition by CO2) impulses during laryngeal application of neutral hypercapnic solution than during the control periods with neutral acapnic solution. However, acidic acapnic solution had no effect on 17 of these 25 CO2-sensitive fibres. For the remaining 8 units, acidic acapnic solution did affect receptor activity, but the degree of excitation or inhibition caused by acidic acapnic solution was significantly smaller than for neutral hypercapnic solution. An example of this is shown in Fig. 3.

In another 25 CO2-sensitive laryngeal fibres, the effect of a more acidic acapnic solution with pH 4.9 was compared with that of the neutral hypercapnic solution with pH 7.0. The activity of 16 of these fibres that responded to neutral hypercapnic solution was unaffected by the more acidic acapnic solution. For the remaining 9 units, the more acidic acapnic solution and the neutral hypercapnic solution had directionally similar effects but quantitatively, the effects of the more acidic acapnic solution were significantly smaller than for the neutral hypercapnic solution.

DISCUSSION

In this study, an isolated, luminally perfused laryngeal preparation has been developed in order to expose superficially located laryngeal receptors to solutions whose composition could be varied precisely to compare the effects of solutions of varying PCO2 and pH. The reliability of this preparation was demonstrated by the consistent behaviour of each receptor throughout the long recording periods, and by the consistency of such behaviour compared with that observed by other workers in the open or ventilated closed larynx, as will be discussed below. The preparation offers a unique opportunity to manipulate the composition of the environment to which laryngeal receptors are exposed.

Responses of laryngeal receptors to CO2

The present data show that the activity of approximately 50 % of pressure-sensitive laryngeal receptors of the quiescent and tonic groups were sensitive to CO2. Among the quiescent, pressure-sensitive receptors responding to CO2, 68 % of the receptors stimulated by negative pressure were excited by CO2, while more than half of the receptors stimulated by positive pressure were inhibited by CO2. The responses of tonic pressure-sensitive receptors to CO2 were remarkably consistent in that 33 of 34 fibres, either stimulated by negative or positive pressure, were inhibited by CO2. This is also largely in agreement with that reported for the ventilated larynx in anaesthetized, paralysed cats (Bradford et al. 1993). Similarly, in anaesthetized, spontaneously breathing cats, Ghosh & Matthew (1994) reported that fibres inhibited by CO2 generally had high resting discharge frequencies (tonic), while fibres excited by CO2 had low resting discharge frequencies (quiescent). Generally similar findings were also reported in another study on the ventilated larynx in paralysed cats (Bartlett & Knuth, 1992).

Effects of CO2 independent of the pH of the perfusing media

The mechanisms of the action of CO2 on laryngeal receptors are not clear, but it would be expected that acidification of the receptor cell or its microenvironment would be responsible. The fact that the CO2 sensitivity of laryngeal receptors appears to be diminished or abolished by carbonic anhydrase inhibition (Wang, 1994; Coates et al. 1996) supports this hypothesis. In the present study, it was shown that the responses of CO2-sensitive fibres to two hypercapnic solutions with the same PCO2 but different pH values were similar in both direction and magnitude. It is noteworthy that acidic acapnic solutions with pH values of 6.8, 6.3 or 4.9 had no effect on a majority of the CO2-responsive fibres tested and significantly smaller effects in the remaining fibres. These results indicate that the effects of CO2 on laryngeal receptors are independent of the pH of the extracellular fluid, and suggest that the intracellular space or the receptor's microenvironment is a likely site of CO2 action. Since CO2 is highly soluble and permeates biological membranes readily, it would rapidly equilibrate between the inside and outside of superficially located laryngeal receptors once CO2 is exposed to the mucosal surface. The presence of carbonic anhydrase in laryngeal surface epithelium and in sensory nerve terminals, as shown by Wang et al. (1994), could facilitate CO2 hydration and the subsequent release of H+, leading to acidification. It is this acidification that may modify receptor activity. The reason that a reduction in extracellular pH without an increase in PCO2, which occurred when the larynx was perfused with acidic acapnic solutions, did not produce similar effects as CO2 on laryngeal receptor activity is, presumably, that H+ cannot enter the cell and cell microenvironment as rapidly as CO2 in order to generate a significant change in H+ concentration. When the larynx is perfused with an acidic solution, only receptors that are directly exposed to the luminal surface may have immediate access to this acidic environment. The extracellular fluid surrounding deeper receptors would probably equilibrate with the acidic perfusing media at a somewhat later stage. It is envisaged, however, that at the usual perfusion time in the present experiments of up to 3 min, H+ in the intralaryngeal perfusing media should be well equilibrated with the extracellular fluid of the superficially located laryngeal receptors. The fact that one-third of laryngeal CO2-sensitive receptors also showed some response to acids indicates that H+ ions have access to the receptors. It is also apparent that the weaker response to acid is not due to H+ ion dilution in the extracellular fluid, since the responses of receptors to acids of pH 4.9 and 6.3-6.8 were similar. A stronger acid solution should reach receptors faster, at higher concentrations, and consequently it would be expected to affect more receptors and exert stronger effects if the diffusion of H+ was a limiting factor in this study.

Evidence for the ability of H+ in laryngeal perfusing media to reach laryngeal receptors at a reasonably fast rate also comes from other studies. Wetmore (1993) reported that application of saline solutions with pH values of 4.0 or 2.0 to the laryngeal surface of the epiglottis of infant rabbits resulted in apnoea within 30 s. Kovar et al. (1979) elicited apnoeic responses in newborn lambs within 20 s of laryngeal application of saline solutions with pH values of 5.0 to 3.0. While the receptors responsible for the initiation of these acid-induced reflexes are likely to be of a different group to those examined in the present study, the ability of acid applied to the laryngeal mucosa to stimulate some laryngeal receptors promptly would indicate that H+ ions can penetrate the barrier provided by mucosal extracellular fluid. Thus the lack of dependence of CO2 responses of laryngeal receptors on the pH of the perfusing media in this study may be an indication of the intracellular location of the chemosensing process. However, whether this chemosensing takes place initially in SLN sensory nerve endings, from which the receptor activity was recorded, or is assisted by other nearby structures in the epithelium, remains to be elucidated.

The role of laryngeal CO2-sensitive receptors

Intralaryngeal CO2 reflexly increases the activity of the genioglossus muscle and decreases ventilation, and both effects are abolished by SLN section (Nolan et al. 1990). Carbon dioxide delivered to the larynx during expiration alone also causes marked changes in receptor activity in anaesthetized, paralysed cats (Bradford et al. 1998), suggesting that CO2 may participate in the regulation of breathing pattern and upper airway muscle activity by altering receptor baseline activity and sensitivity to pressure and other stimuli related to normal breathing. The concurrence of the effects of laryngeal CO2 and negative pressure on ventilation and upper airway dilating muscle activity suggests that CO2 and negative pressure in the larynx may work jointly to maintain upper airway patency or to counteract airway obstruction under pathological conditions as suggested previously (Anderson et al. 1990; Nolan et al. 1990; Bradford et al. 1993, 1998). When the pharyngeal airway becomes occluded, such as happens in obstructive sleep apnoea, the larynx is subjected to increased negative pressure and, possibly, elevated levels of CO2. The combined effects of CO2 and negative pressure on laryngeal pressure-sensitive receptors would reflexly favour the reopening of the occluded airway and limit further occlusion.

To what extent each group of laryngeal receptors contribute to each of the CO2-induced reflexes remains to be elucidated. The different patterns of responses of laryngeal receptors to CO2 suggest that the reflex regulation of breathing and upper airway patency by laryngeal CO2 is complex. In the present experiments, CO2 had effects that depended on the receptor type examined. In previous work, we have shown that CO2 also affects the activity of a variety of other receptors including cold (Bradford et al. 1993, 1994, 1998) and ‘drive’ receptors (Bradford et al. 1994). The reflex effects of specific stimulation of cold receptors have been elucidated (Orani et al. 1991) but reflex responses to specific stimulation of ‘drive’ and pressure receptors are unknown. It is difficult to devise appropriate experiments to examine this question because of the variety of pressure and ‘drive’ receptors described and the overlap in sensitivity of these receptors. It is therefore difficult to ascribe a particular receptor response to the reflex effects produced by intralaryngeal CO2.

In summary, this study shows that an isolated laryngeal preparation perfused through its lumen with physiological saline solution can be used to examine laryngeal receptor function allowing control of the composition of the receptor's environment. This preparation confirms and extends the findings by other workers on the CO2 responses of laryngeal pressure-sensitive receptors in that: (1) a significant proportion of laryngeal receptors from each functional group are CO2 sensitive; (2) tonic laryngeal receptors are generally inhibited by CO2 while quiescent receptors are either inhibited or excited by CO2; (3) among the CO2-sensitive quiescent fibres, a majority of the receptors stimulated by negative pressure are excited by CO2, while over half of the receptors stimulated by positive pressure are inhibited by CO2. In addition, this study demonstrates that the effects of CO2 on a majority of the CO2-sensitive pressure receptors, especially the quiescent type, are not obvious under resting conditions but become apparent at an active state. More importantly, this study has shown that the effects of CO2 on laryngeal receptors are independent of the pH of the media perfusing the larynx, thus suggesting that acidification of the receptor cell or its microenvironment is the mechanism for laryngeal CO2 chemoreception.

Acknowledgements

This study was supported by the Health Research Board, Ireland.

Ancillary