Vagal afferent and reflex responses to changes in surface osmolarity in lower airways of dogs T. E. PISARRI, Cardiovascular
A. JONZON, Research Institute
San Francisco, California
H. M. COLERIDGE, and Department
AND
of Physiology,
J. C. G. COLERIDGE University
of California,
94143-0130
are denervated below the suture line, just above the carina (16). However, although pulmonary stretch receptors, rapidly adapting receptors, and pulmonary and bronchial C-fibers are all believed to be involved in the cough reflex (18) and are known to influence bronchomotor tone (9), their sensitivity to changes in airway surface osmolarity has not been studied. Because rapidly adapting receptors and the endings of pulmonary and bronchial C-fibers are stimulated by injection of hyperosmotic saline into the blood supply to the endings (23) and because these vagal afferents are also accessible to stimulants delivered via the airways (4, 6,8), we hypothesized that they would respond to deviations from isosmolarity of the fluid lining the airway lumen. We undertook the present study to test this hypothesis. In spontaneously breathing dogs, we examined the vagally mediated reflex changes evoked by injection of small volumes of water or hyperosmotic saline into a lobar bronchus. In artificially ventilated dogs, we recorded the response of vagal afferents in the lower airways to the same solutions. We compared the reflex and afferent responses to nonisosmotic solutions with those evoked by similar volumes of isosmotic saline. To determine whether the vagal endings were sensitive to changes in airway C-fibers; airway defensereflexes; effects of near-drown- ionic content rather than osmolarity per se, we compared ing; evaporative water loss from airways; exercise-induced the afferent response to water with that to isosmotic gluasthma; inhalation of nonisosmotic liquids; pulmonary vagal cose solution. PISARRI, T.E., A. JONKIN, H.M. COLERIDGE, AND J.C.G. COLERIDGE. Vagal afferent and reflex responses to changes in surface osmolarity in Zower airways of dogs. J. Appl. Physiol. 73(6): 230%2313,1992.--In anesthetized dogswe examinedthe sensitivity of afferent vagal endingsin the lungs to changesin airway fluid osmolarity. Injection of 0.25-0.5 ml/kg water or hyperosmotic sodium chloride solutions (1,200-2,400 mmol/l) into a lobar bronchus causedbradycardia, arterial hypotension, apneafollowed by rapid shallow breathing, and contraction of tracheal smooth muscle. All effects were abolished by vagotomy. We examined the sensory mechanismsinitiating these effects by recording afferent vagal impulsesarising from the lung lobe into which the liquids were injected. Water stimulated pulmonary and bronchial C-fibers and rapidly adapting receptors; isosmoticsaline and glucosesolutions were ineffective. Hyperosmotic saline (1,200-9,600mmol/l, 0.25-I ml/kg) stimulated these afferents in a concentration-dependent manner. Stimulation began l-10 s after the injection and sometimes continued for several minutes. Responsesof slowly adapting stretch receptorsvaried. Our resultssuggestthat nonisosmoticfluid in the lower airways initiates defensereflexes by stimulating pulmonary and bronchial C-fibers and rapidly adapting receptors. Conceivably, stimulation of these afferents asa result of evaporative water lossfrom airway surface liquid could contribute to exercise-inducedasthma.
afferents; rapidly adapting receptors
METHODS RESPIRATORYDEFENSE REFLEXES triggeredbystimula-
tion of sensory nerve endings in the respiratory tract serve to protect the lungs from the injurious effects of inhaled foreign substances (5,9,11). In animals, stimulation of the laryngeal mucosa with water causes coughing, bronchoconstriction, apnea, and adduction of the vocal chords (32). In humans, inhalation of aerosols of nonisosmotic solutions evokes cough and, particularly in hyperresponsive individuals, bronchoconstriction (2, 14, 18). Under deep anesthesia, the cough evoked by water may be replaced by apnea or rapid shallow breathing (22). Afferent input from vagal endings in the lower airways could contribute to the cough and other reflex effects triggered in humans by inhalation of nonisosmotic fluids. For example, cough evoked by inhaling water aerosol of small-diameter droplets is much less frequent in recipients of heart-lung transplants, whose lungs and airways
General Dogs (12-28 kg) were given promazine hydrochloride (Sparine, 2.5 mg/kg im, Wyeth Laboratories) 30 min before they were anesthetized. In the reflex studies, four dogs were anesthetized with thiopental sodium (25 mg/ kg iv) followed by cr-chloralose (80 mg/kg iv). Supplemental doses of cu-chloralose (10 mg/kg iv) were given hourly to maintain surgical anesthesia. In the afferent studies, 35 dogs were anesthetized with 0.25 ml/kg iv of a 1~1mixture of solutions of Dial compound (allobarbital 100 mg/ml, urethan 400 mg/ml; Sigma Chemical) and pentobarbital sodium (50 mg/ml). Supplemental doses of anesthetic were given as necessary to maintain surgical anesthesia. The trachea was cannulated low in the neck. A catheter was inserted through a rubber stopper in the tracheal cannula and with the aid of a bronchoscope its tip was directed into a left or right lobar bronchus. Tidal CO, was monitored by a Beckman LB-l gas ana-
016L7567/92$2.00Copyright 0 1992 the AmericanPhysiological Society
2305
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 20, 2019.
2306
AIRWAY
OSMOLARITY
AND LUNG AFFERENTS
220
HR (beatshin) 60 200
ABP (mm)-lg) 0
Tr T (9)
0
1 min 300 2600 FIG. 1. Reflex effects of injecting sodium chloride solutions (0.25 ml/kg) of various osmolarities in random order into left lower lobar bronchus. HR, heart rate; ABP, arterial blood pressure; VT, tidal volume; Tr T, tracheal smooth muscle tension, measured from baseline of 75 g. Note marked response to water (0 mmol/l) and trivial response to isosmotic saline (300 mmol/l); effects of hyperosmotic saline increased with increasing deviation from isosmolarity. Osmolarity -of injected solutions is indicated in mmol/l beneath event marks.
lyzer. Arterial blood samples were withdrawn periodically, and blood gas and acid-base values were determined with an automatic analyzer (Corning 175); base deficit was corrected by intravenous administration of sodium bicarbonate solution. Reflex Studies
The dog breathed spontaneously through a pneumotachograph (Fleisch no. 1) attached to the tracheal cannula. A constant flow of 0, (1 Urnin) was directed across the inlet of the pneumotachograph to enrich the inspired air. Measurements. Femoral arterial blood pressure was measured by a Statham P23Gc strain gauge. An electrocardiogram (lead II) was recorded. Heart rate was measured by a cardiotachometer (Grass 7P4C) triggered by the arterial pressure signal. Respiratory airflow was measured by the pneumotachograph and a differential pressure transducer (Statham PMlSTC), the signal of which was integrated (Grass polygraph 7PlOA) to provide a record of tidal volume. The signals representing tidal CO,, blood pressure, heart rate, airflow, tidal volume, and tracheal smooth muscle tension (see below) were recorded by a Grass model 7 polygraph. Tracheal smooth muscle tension. A segment of trachea (4-5 cm) immediately caudal to the larynx was incised ventrally in the midline and transversely across both ends of the midline incision. The posterior wall was left intact. Each midline cut edge was retracted laterally by nylon threads attached to a stationary bar on one side and to a force-displacement transducer (Grass FT03) on the other. The segment was stretched to a baseline tension of 75 g. Threads were placed around the recurrent and pararecurrent laryngeal nerves to faeilitate their transection before cervical vagotomy. (Cervical vagotomy interrupts both afferent fibers from the lower airways and the recurrent nerves, which supply some of the motor innervation of tracheal smooth muscle. With the recurrent nerves cut, the motor innervation of the cra-
nial portion of the trachea is exclusively from the superior laryngeal nerves, which are not affected by infranodose vagotomy. Therefore postvagotomy responses were always compared with control responses recorded after the recurrent nerves were cut.) Experimental protocols. We recorded changes in heart rate, blood pressure, tidal volume, and tracheal smooth muscle tension evoked when small volumes (0.25-0.5 ml/ kg) of test solutions at room temperature (M25OC) were injected through the bronchial catheter into a lobar bronchus. Isosmotic saline was injected as a control. Once the cardiovascular and respiratory variables had returned to control or reached new steady-state values after an injection, we aspirated any accumulated fluid from the lung through the bronchial catheter and hyperinflated the lung to restore compliance. Because the volume of aspirated fluid suggested that water produced the least accumulation of fluid in the lung, it was generally the first solution injected. We then injected NaCl solutions with osmolarities of 300 (isosmotic, O-9%), 1,200 (3.6%), and 2,400 (7.2%) mmol/l (mosmolll). NaCl solutions with an osmolarity of 2,400 mmolfl were usually injected last; solutions of intermediate osmolarity were injected in random order (Fig. 1). Finally, we cut the recurrent laryngeal nerves and repeated the injection of water or 2,400 mmol/l saline before and after cutting the cervical vagus nerves. Afferent Studies
The chest was opened in the midsternal line, and the lungs were ventilated with 50% 0, in air by a Harvard respirator (model 613), the expiratory outlet of which was placed under 3-5 cm of water. Tidal volume was set at -15 ml/kg and ventilation frequency at Xi-20 cycles/ min. End-tidal PCO, was kept at -35 Torr by adjusting ventilatory frequency. When impulses were to be recorded from bronchial C-fibers (see below), the chest was also opened in the fifth right intercostal space, and a fine polyethylene
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 20, 2019.
AIRWAY
HR (beats/mi
OSMOLARITY
n)
-
**
Tr T (9)
-i 30-
**
TT (9
0
0
2. Changes in HR, ABP, Tr T, and duration of respiratory cycle (TT) evoked by injection of 0.5 ml/kg water and NaCl solutions of various osmol&iti& into a lobar bronchus. 0, Control; l , peak response. Peak response in TT indicates duration of reflex apnea. Values are means -t SE-from 2 trials in each of 4 dogs, except that Tr T was measured in only 3 dogs, and effects of injecting 2,400 mmol/l saline were examined only once in each dog. Probability that change is same as that after isosmotic solution (300 mmol/l): * P < 0.05; ** P < 0.01. FIG.
catheter was inserted retrogradely in the intercostal artery from which the right bronchoesophageal artery originated. The catheter was tied in place with its tip just downstream to the origin of the bronchial artery, so that solutions injected sloWly through the catheter entered the bronchial circulation (19). Indocyanine green dye (Cardio-Green) was injected to confirm that the artery perfused the bronchial tree; arterial branches supplying the esophagus were ligated. Measurements. Arterial blood pressure, heart rate, and an electrocardiogram were recorded as described above. Tracheal pressure was recorded from a sidearm of the tracheal cannula by a Statham PR23-6G-300 strain gauge. Signals representing tidal CO,, tracheal pressure, blood pressure, heart rate, and a rate meter record of vagal impulse frequency (see below) were recorded by a Grass model 7 polygraph. The electrocardiogram, tracheal pressure, blood pressuret heart rate, and nerve action potentials were also recorded by a Gould 1000 electrostatic recorder. Nerve recording and identificatiun fibers. Using conventional techniques,
of ufferent vugal
we dissected fine slips from the otherwise intact cervical vagus nerves and recorded impulses arising from endings in the lungs. Pulmonary C-fibers were identified by their sparse and irregular discharge, their short latency response to injection of capsaicin into the right atrium, their conduction velocity, and their response to hyperinflation of the lungs (4). Bronchial C-fibers, which were dissected from the right
AND LUNG AFFERENTS
2307
vagus nerve, were identified by their sparse and irregular discharge, their conduction velocity, and their response to injection of small doses of capsaicin into the right bronchial artery (4, 19). Rapidly adapting receptors and slowly adapting stretch receptors were identified by their characteristic patterns of discharge and their rate of adaptation to hyperinflation of the lung (3 X tidal volume). We determined the location of afferent endings by probing the lungs and selected fibers whose endings were situated in the lobe in which the tip of the bronchial catheter was positioned. However, C-fiber endings are sometimes hard to locate. We discarded four pulmonary and six bronchial C-fibers, provisionally selected on the basis of their responses to capsaicin injected into the right atrium and bronchial artery, because we were not certain that their endings were in the lobe receiving the injected solutions, inasmuch as none of them was stimulated by any of the solutions. However, we cannot exclude the possibility that some of these receptors were in fact not osmotically sensitive. Afferent impulses were counted (in 0.1-s or l-s bins or by a counter reset at the beginning of each ventilator cycle) by rate meters with window discriminators set to count potentials of a particular amplitude. Experimental protocols. We selected fibers with sensory endings located in the lung lobe in which the tip of the bronchial catheter was positioned. We recorded control activity over a period of at least five ventilator cycles. We then injected 0.25-1.0 ml/kg of distilled water or of one or more solutions of sodium chloride with osmolarities of 300 (0.9%), 1,200 (3.6%), 2,400 (7.2%), 4,800 (14.4%), and 9,600 (28.8%) mmol/l. Initially, solutions (including control solutions of isosmotic saline) were injected at room temperature; in later experiments, all solutions were warmed to 37OC. Generally, the first test on a fiber was with water, and the most concentrated solutions were injected last; however, solutions were sometimes injected in random order. Some fibers were also tested by injection of 5% glucose, an isosmotic solution free of sodium and chloride ions. After each injection, we aspirated as much liquid as possible through the bronchial catheter and hyperinflated the lung to restore compliance. Decreases in compliance after injections of water or isosmotic saline were small and readily reversed by hyperinflation. Water and isosmotic saline were tested on most fibers. Hyperosmotic solutions were used more sparingly, because repeated injections irreversibly decreased lung compliance. The more concentrated solutions were used only occasionally because they tended to produce lung edema, and the experiments had to be terminated. Analysis of Results
In the reflex studies, control heart rate, blood pressure, tracheal tension, and duration of respiratory cycle before bronchial injection were compared with the corresponding values at the point of maximal change. In the afferent studies, control impulse frequencies were averaged over five ventilator cycles; maximum firing rates evoked by the solutions were averaged over one ventilator cycle at the peak of the response; both were expressed as im-
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 20, 2019.
2308
AIRWAY
Pulmonary
20
OSMOLARITY
AND LWNG AFFERENTS
B *
r
C-fibers IF
(imp/s)
Stimulated
i
4
0
0
7 17
7 8
0 8
*I
Bronchial C-fibers IF
1
(impts)
/
d Stimuiated Tested
Rapidly Adapting Receptors
3
5 9
9
7
9
PIG. 3. Responses (means t SE) of pulmonary C-fibers, bronchial C-fibers, and rapidly adapt.ing receptors to paired injections into a lobar bronchus of water (0 mmol/l) and isosmotic saline (A) and of hyperosmotic saline solutions and isosmotic saline (B). IF, impulse frequency. No. of fibers stimulated and tested and osmolarity of solutions (in mm01 /u are given beneath each column. 0, Control; l , peak response. Probability that change is same as that to isosmotic solution (300 mmol/l): * P < 0.05; ** P < 0.01.
c
50 ,F
(imp/s) T t+
0 Stimulated Tested
-/
5
18 19
13 19
9
6 9
9 9
0
300
300
1200
2400
pulses per second. No change in discharge was recorded if the peak frequency was not greater than the frequency during the most active cycle in the control period. Latenties of afferent and reflex responses were measured from the beginning of the bronchial injection. Results are expressed as means t SE. Because the parametric statistics used to analyze the data assume normal distribution and common variances, we analyzed the square roots of impulse frequencies to more closely approximate these requirements. The effect of each solution was compared with that of isosmotic saline by use of analysis of variance for repeated measures, followed when appropriate by the multiple-stage F test of Ryan, Elliot and Gabriel, and Welsch with CRUNCH statistical software. Statistical significance was accepted if P < 0.05. RESULTS
Reflex Effects
Injection of water or hyperosmotic sodium chloride solution (0.25-0.5 ml/kg) into a lobar bronchus in four dogs invariably decreased heart rate and arterial blood pressure and evoked apnea, usually followed by rapid shallow breathing (Fig. I). In each of the three dogs in which
tracheal tension was recorded, bronchial injection also caused contraction of tracheal smooth muscle (Fig. 1). These changes began 3-10 s after the beginning of the injection, cardiovascular and respiratory effects being maximal after lo-20 s and tracheal contraction after ZO40 s. The amplitude of the effects varied with the deviation of the injected solution from isosmolarity (Figs. 1 and 2). Isosmotic saline (300 mmol/l) had little or no effect on any of the variables except tracheal smooth muscle tension, which increased but by much less than after any of the other solutions (Figs. 1 and 2). We examined the effects of vagotomy on the responses to bronchial injection in three dogs. After the recurrent and pararecurrent laryngeal nerves were cut, injection of water or 2,400 mmol/l sodium chloride solution invariably had conspicuous effects on heart rate, blood pressure, breathing, and tracheal tension. After cervical vagotomy, injection of the same solutions had no effect on any of the recorded variables. Afferent Effects
We recorded the activity of 35 nonmyelinated and 50 myelinated afferent vagal fibers arising from endings located in the pulmonary lobe to which the bronchial ca th-
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 20, 2019.
AIRWAY
USMOLARITY
AND LUNG AFFERENTS
2309
AP
4. Responses of a pulmonary (A-D) and a bronchial (E) C-fiber to bronchial injection of solutions of various osmolarities. Pulmonary C-fiber stimulated by 1.0 ml/kg water (A) or hyperosmotic NaCl solutions (C and D) injected into left lower lobar bronchus. Stimulation by hyperosmotic solutions increased with deviation from isosmolarity, stimulation by 1,200 mmol/l saline lasting -12 s, that by 2,400 mmol/l saline 1 min. Note that same volume of isosmotic saline (300 mmol/l) had no effect (B). E: bronchial C-fiber stimulated by 0.6 ml/ kg of 1,200 mmol/l saline injected into right lower lobar bronchus. AP, action potentials recorded from strand of left (A-D) and right. (27) vagus nerve. Osmolarity of injected solutions is given in mmol/l beneath event marks. Note reflex decrease in HR and ABP in E. FIG.
300
C AP :mHPO)
20 o E1200
AP
2400
HR (beats/min)
ABP (mm&d
5s
180
,4o 150
50 was c rectea ana examinea tnelr response to nqection of water and hyperosmotic solutions. The concentration-dependence of the response was established by paired tests with 300,1,200, and 2,400 mmol/l saline and, in a few fibers, by tests with more concentrated solutions. Isosmotic saline was used as a control for the mechanical and temperature effects of the injections. Pulmonary C-fibers. We recorded the activity of 22 pulmonary C-fibers. Injection of water into a lobar bronchus stimulated all 17 pulmonary C-fibers tested (Fig. 3A). The evoked discharge, which consisted of a continuous or intermittent burst of impulses without respiratory modulation (Fig. 4A), began 7.7 t 1.6 s after the injection (the most rapid response beginning within 1 s) and lasted 19 t 3 s (range 9-44 s). Stimulation by water was not due only to the mechanical effects of the injected volume, because a similar volume of isosmolar saline (300 mmolll) had only minor effects (Fig. 3A), evoking a small increase in 7 of the 17 fibers and none in the others (Fig. 4B). Stimulation by water appeared to be due to hyposmolarity rather than to an absence of sodium or chloride ions. Thus the response of seven pulmonary C-fibers to isosmotic glucose was not significantly different from that to isosmotic saline, both being considerably less than that to water (Fig. 5A). Hyperosmotic saline stimulated 11 of 12 pulmonary C-fibers tested, evoking a continuous discharge (latency 5.5 + 1.1 s) that reached maximum intensity within 10 s of its onset. The discharge was often prolonged (Figs. 4D and 6A), its duration averaging 112 t 31 s (range 7-180 s) when 2,400 mmolll saline were injected. The intensity and duration of the evoked discharge usually increased
eter
witn tne concentration of the injected solution (Figs. 3B, 4, C and D, and 6A). In paired tests on eight pulmonary C-fibers, firing increased by 10.3 t 3.5 impulses/s in response to injection of 1,200 mmol/l saline and by 18.0 t 4.8 impulses/s in response to 2,400 mmol/l saline; it was virtually unaffected by isosmotic saline (Fig. 3B). Peak tracheal pressure usually increased by -1 cmH,O after injection of hyperosmotic saline, but the afferent response was unrelated to the ventilator cycle (Fig. 4, C and D) and often began before or in the absence of any change in pressure. Even when tracheal pressure increased progressively in the few experiments in which fluid was not aspirated from the airways between injections, the afferent response was determined by the osmolarity of the injected fluid rather than by the level of tracheal pressure (Fig. 6A). Bronchid C-fibers. Water stimulated 9 of the 13 bronchial C-fibers investigated, the latency and duration of the response being similar to that of pulmonary C-fibers; activity in these 13 fibers increased from 0.2 t 0.1 to 9.2 t 2.6 impulses/s. In some fibers the evoked discharge was continuous but irregular; in others it took the form of sporadic bursts. Effects of water and isosmotic saline were compared in paired tests on 10 bronchial C-fibers; isosmotic saline had little effect, causing a small increase in discharge in only three of the fibers (Fig. 3A). Hyperosmotic saline stimulated 10 of 12 bronchial Cfibers (Figs. 4E and 6B), the response being similar in pattern to that evoked by water but, as was the case with pulmonary C-fibers, generally more prolonged (duration 44 +- 9 s, maximum 200 s). Of the 10 bronchial C-fibers stimulated by hyperosmotic saline, six were stimulated
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 20, 2019.
2310
AIRWAY
OSMOLARITY
AND LUNG AFFERENTS
of 1,200 and 2,400 mmol/l saline on nine receptors indicated that the response was concentration dependent (Figs. 3B and 6C). Water was by far the strongest stimulus to rapidly adapting receptors; in paired tests on 14 receptors, water IF increased discharge from 0.4 t 0.2 to 33.0 t 4.6 impulses/ (imp/s) s, whereas 2,400 mmol/l saline increased it from 0.9 t 0.4 to 18.6 t 4.3 impulses/s (P < 0.01). Slowly adapting pulmonary stretch receptors. In conl trast to their effects on other lower airway afferents, nei4 0 0 ther water nor hyperosmotic saline (1,200 and 2,400 mmol/l) had significant effects on most of the 26 slowly 5 adapting stretch receptors examined. Water stimulated c only 4 of 23 receptors, which began to discharge throughout deflation within l-9 s of the injection, thereby increasing mean activity (from 22 t 5 to 57 t 10 impulses/ s) while reducing the ventilatory modulation (Fig. 8A); IF the control pattern of discharge returned in 22-96 s. Two (imp/s) of the receptors stimulated by water were also stimulated, to a lesser extent, by isosmotic saline. Hyperosmotic saline (1,200 and 2,400 mmol/l) had 0 only minor effects on 21 stretch receptors, the changes consisting at most of small increases in peak frequency Glucose Water NaCI 300 mmol/l (
A. JONZON, Research Institute
San Francisco, California
H. M. COLERIDGE, and Department
AND
of Physiology,
J. C. G. COLERIDGE University
of California,
94143-0130
are denervated below the suture line, just above the carina (16). However, although pulmonary stretch receptors, rapidly adapting receptors, and pulmonary and bronchial C-fibers are all believed to be involved in the cough reflex (18) and are known to influence bronchomotor tone (9), their sensitivity to changes in airway surface osmolarity has not been studied. Because rapidly adapting receptors and the endings of pulmonary and bronchial C-fibers are stimulated by injection of hyperosmotic saline into the blood supply to the endings (23) and because these vagal afferents are also accessible to stimulants delivered via the airways (4, 6,8), we hypothesized that they would respond to deviations from isosmolarity of the fluid lining the airway lumen. We undertook the present study to test this hypothesis. In spontaneously breathing dogs, we examined the vagally mediated reflex changes evoked by injection of small volumes of water or hyperosmotic saline into a lobar bronchus. In artificially ventilated dogs, we recorded the response of vagal afferents in the lower airways to the same solutions. We compared the reflex and afferent responses to nonisosmotic solutions with those evoked by similar volumes of isosmotic saline. To determine whether the vagal endings were sensitive to changes in airway C-fibers; airway defensereflexes; effects of near-drown- ionic content rather than osmolarity per se, we compared ing; evaporative water loss from airways; exercise-induced the afferent response to water with that to isosmotic gluasthma; inhalation of nonisosmotic liquids; pulmonary vagal cose solution. PISARRI, T.E., A. JONKIN, H.M. COLERIDGE, AND J.C.G. COLERIDGE. Vagal afferent and reflex responses to changes in surface osmolarity in Zower airways of dogs. J. Appl. Physiol. 73(6): 230%2313,1992.--In anesthetized dogswe examinedthe sensitivity of afferent vagal endingsin the lungs to changesin airway fluid osmolarity. Injection of 0.25-0.5 ml/kg water or hyperosmotic sodium chloride solutions (1,200-2,400 mmol/l) into a lobar bronchus causedbradycardia, arterial hypotension, apneafollowed by rapid shallow breathing, and contraction of tracheal smooth muscle. All effects were abolished by vagotomy. We examined the sensory mechanismsinitiating these effects by recording afferent vagal impulsesarising from the lung lobe into which the liquids were injected. Water stimulated pulmonary and bronchial C-fibers and rapidly adapting receptors; isosmoticsaline and glucosesolutions were ineffective. Hyperosmotic saline (1,200-9,600mmol/l, 0.25-I ml/kg) stimulated these afferents in a concentration-dependent manner. Stimulation began l-10 s after the injection and sometimes continued for several minutes. Responsesof slowly adapting stretch receptorsvaried. Our resultssuggestthat nonisosmoticfluid in the lower airways initiates defensereflexes by stimulating pulmonary and bronchial C-fibers and rapidly adapting receptors. Conceivably, stimulation of these afferents asa result of evaporative water lossfrom airway surface liquid could contribute to exercise-inducedasthma.
afferents; rapidly adapting receptors
METHODS RESPIRATORYDEFENSE REFLEXES triggeredbystimula-
tion of sensory nerve endings in the respiratory tract serve to protect the lungs from the injurious effects of inhaled foreign substances (5,9,11). In animals, stimulation of the laryngeal mucosa with water causes coughing, bronchoconstriction, apnea, and adduction of the vocal chords (32). In humans, inhalation of aerosols of nonisosmotic solutions evokes cough and, particularly in hyperresponsive individuals, bronchoconstriction (2, 14, 18). Under deep anesthesia, the cough evoked by water may be replaced by apnea or rapid shallow breathing (22). Afferent input from vagal endings in the lower airways could contribute to the cough and other reflex effects triggered in humans by inhalation of nonisosmotic fluids. For example, cough evoked by inhaling water aerosol of small-diameter droplets is much less frequent in recipients of heart-lung transplants, whose lungs and airways
General Dogs (12-28 kg) were given promazine hydrochloride (Sparine, 2.5 mg/kg im, Wyeth Laboratories) 30 min before they were anesthetized. In the reflex studies, four dogs were anesthetized with thiopental sodium (25 mg/ kg iv) followed by cr-chloralose (80 mg/kg iv). Supplemental doses of cu-chloralose (10 mg/kg iv) were given hourly to maintain surgical anesthesia. In the afferent studies, 35 dogs were anesthetized with 0.25 ml/kg iv of a 1~1mixture of solutions of Dial compound (allobarbital 100 mg/ml, urethan 400 mg/ml; Sigma Chemical) and pentobarbital sodium (50 mg/ml). Supplemental doses of anesthetic were given as necessary to maintain surgical anesthesia. The trachea was cannulated low in the neck. A catheter was inserted through a rubber stopper in the tracheal cannula and with the aid of a bronchoscope its tip was directed into a left or right lobar bronchus. Tidal CO, was monitored by a Beckman LB-l gas ana-
016L7567/92$2.00Copyright 0 1992 the AmericanPhysiological Society
2305
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 20, 2019.
2306
AIRWAY
OSMOLARITY
AND LUNG AFFERENTS
220
HR (beatshin) 60 200
ABP (mm)-lg) 0
Tr T (9)
0
1 min 300 2600 FIG. 1. Reflex effects of injecting sodium chloride solutions (0.25 ml/kg) of various osmolarities in random order into left lower lobar bronchus. HR, heart rate; ABP, arterial blood pressure; VT, tidal volume; Tr T, tracheal smooth muscle tension, measured from baseline of 75 g. Note marked response to water (0 mmol/l) and trivial response to isosmotic saline (300 mmol/l); effects of hyperosmotic saline increased with increasing deviation from isosmolarity. Osmolarity -of injected solutions is indicated in mmol/l beneath event marks.
lyzer. Arterial blood samples were withdrawn periodically, and blood gas and acid-base values were determined with an automatic analyzer (Corning 175); base deficit was corrected by intravenous administration of sodium bicarbonate solution. Reflex Studies
The dog breathed spontaneously through a pneumotachograph (Fleisch no. 1) attached to the tracheal cannula. A constant flow of 0, (1 Urnin) was directed across the inlet of the pneumotachograph to enrich the inspired air. Measurements. Femoral arterial blood pressure was measured by a Statham P23Gc strain gauge. An electrocardiogram (lead II) was recorded. Heart rate was measured by a cardiotachometer (Grass 7P4C) triggered by the arterial pressure signal. Respiratory airflow was measured by the pneumotachograph and a differential pressure transducer (Statham PMlSTC), the signal of which was integrated (Grass polygraph 7PlOA) to provide a record of tidal volume. The signals representing tidal CO,, blood pressure, heart rate, airflow, tidal volume, and tracheal smooth muscle tension (see below) were recorded by a Grass model 7 polygraph. Tracheal smooth muscle tension. A segment of trachea (4-5 cm) immediately caudal to the larynx was incised ventrally in the midline and transversely across both ends of the midline incision. The posterior wall was left intact. Each midline cut edge was retracted laterally by nylon threads attached to a stationary bar on one side and to a force-displacement transducer (Grass FT03) on the other. The segment was stretched to a baseline tension of 75 g. Threads were placed around the recurrent and pararecurrent laryngeal nerves to faeilitate their transection before cervical vagotomy. (Cervical vagotomy interrupts both afferent fibers from the lower airways and the recurrent nerves, which supply some of the motor innervation of tracheal smooth muscle. With the recurrent nerves cut, the motor innervation of the cra-
nial portion of the trachea is exclusively from the superior laryngeal nerves, which are not affected by infranodose vagotomy. Therefore postvagotomy responses were always compared with control responses recorded after the recurrent nerves were cut.) Experimental protocols. We recorded changes in heart rate, blood pressure, tidal volume, and tracheal smooth muscle tension evoked when small volumes (0.25-0.5 ml/ kg) of test solutions at room temperature (M25OC) were injected through the bronchial catheter into a lobar bronchus. Isosmotic saline was injected as a control. Once the cardiovascular and respiratory variables had returned to control or reached new steady-state values after an injection, we aspirated any accumulated fluid from the lung through the bronchial catheter and hyperinflated the lung to restore compliance. Because the volume of aspirated fluid suggested that water produced the least accumulation of fluid in the lung, it was generally the first solution injected. We then injected NaCl solutions with osmolarities of 300 (isosmotic, O-9%), 1,200 (3.6%), and 2,400 (7.2%) mmol/l (mosmolll). NaCl solutions with an osmolarity of 2,400 mmolfl were usually injected last; solutions of intermediate osmolarity were injected in random order (Fig. 1). Finally, we cut the recurrent laryngeal nerves and repeated the injection of water or 2,400 mmol/l saline before and after cutting the cervical vagus nerves. Afferent Studies
The chest was opened in the midsternal line, and the lungs were ventilated with 50% 0, in air by a Harvard respirator (model 613), the expiratory outlet of which was placed under 3-5 cm of water. Tidal volume was set at -15 ml/kg and ventilation frequency at Xi-20 cycles/ min. End-tidal PCO, was kept at -35 Torr by adjusting ventilatory frequency. When impulses were to be recorded from bronchial C-fibers (see below), the chest was also opened in the fifth right intercostal space, and a fine polyethylene
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 20, 2019.
AIRWAY
HR (beats/mi
OSMOLARITY
n)
-
**
Tr T (9)
-i 30-
**
TT (9
0
0
2. Changes in HR, ABP, Tr T, and duration of respiratory cycle (TT) evoked by injection of 0.5 ml/kg water and NaCl solutions of various osmol&iti& into a lobar bronchus. 0, Control; l , peak response. Peak response in TT indicates duration of reflex apnea. Values are means -t SE-from 2 trials in each of 4 dogs, except that Tr T was measured in only 3 dogs, and effects of injecting 2,400 mmol/l saline were examined only once in each dog. Probability that change is same as that after isosmotic solution (300 mmol/l): * P < 0.05; ** P < 0.01. FIG.
catheter was inserted retrogradely in the intercostal artery from which the right bronchoesophageal artery originated. The catheter was tied in place with its tip just downstream to the origin of the bronchial artery, so that solutions injected sloWly through the catheter entered the bronchial circulation (19). Indocyanine green dye (Cardio-Green) was injected to confirm that the artery perfused the bronchial tree; arterial branches supplying the esophagus were ligated. Measurements. Arterial blood pressure, heart rate, and an electrocardiogram were recorded as described above. Tracheal pressure was recorded from a sidearm of the tracheal cannula by a Statham PR23-6G-300 strain gauge. Signals representing tidal CO,, tracheal pressure, blood pressure, heart rate, and a rate meter record of vagal impulse frequency (see below) were recorded by a Grass model 7 polygraph. The electrocardiogram, tracheal pressure, blood pressuret heart rate, and nerve action potentials were also recorded by a Gould 1000 electrostatic recorder. Nerve recording and identificatiun fibers. Using conventional techniques,
of ufferent vugal
we dissected fine slips from the otherwise intact cervical vagus nerves and recorded impulses arising from endings in the lungs. Pulmonary C-fibers were identified by their sparse and irregular discharge, their short latency response to injection of capsaicin into the right atrium, their conduction velocity, and their response to hyperinflation of the lungs (4). Bronchial C-fibers, which were dissected from the right
AND LUNG AFFERENTS
2307
vagus nerve, were identified by their sparse and irregular discharge, their conduction velocity, and their response to injection of small doses of capsaicin into the right bronchial artery (4, 19). Rapidly adapting receptors and slowly adapting stretch receptors were identified by their characteristic patterns of discharge and their rate of adaptation to hyperinflation of the lung (3 X tidal volume). We determined the location of afferent endings by probing the lungs and selected fibers whose endings were situated in the lobe in which the tip of the bronchial catheter was positioned. However, C-fiber endings are sometimes hard to locate. We discarded four pulmonary and six bronchial C-fibers, provisionally selected on the basis of their responses to capsaicin injected into the right atrium and bronchial artery, because we were not certain that their endings were in the lobe receiving the injected solutions, inasmuch as none of them was stimulated by any of the solutions. However, we cannot exclude the possibility that some of these receptors were in fact not osmotically sensitive. Afferent impulses were counted (in 0.1-s or l-s bins or by a counter reset at the beginning of each ventilator cycle) by rate meters with window discriminators set to count potentials of a particular amplitude. Experimental protocols. We selected fibers with sensory endings located in the lung lobe in which the tip of the bronchial catheter was positioned. We recorded control activity over a period of at least five ventilator cycles. We then injected 0.25-1.0 ml/kg of distilled water or of one or more solutions of sodium chloride with osmolarities of 300 (0.9%), 1,200 (3.6%), 2,400 (7.2%), 4,800 (14.4%), and 9,600 (28.8%) mmol/l. Initially, solutions (including control solutions of isosmotic saline) were injected at room temperature; in later experiments, all solutions were warmed to 37OC. Generally, the first test on a fiber was with water, and the most concentrated solutions were injected last; however, solutions were sometimes injected in random order. Some fibers were also tested by injection of 5% glucose, an isosmotic solution free of sodium and chloride ions. After each injection, we aspirated as much liquid as possible through the bronchial catheter and hyperinflated the lung to restore compliance. Decreases in compliance after injections of water or isosmotic saline were small and readily reversed by hyperinflation. Water and isosmotic saline were tested on most fibers. Hyperosmotic solutions were used more sparingly, because repeated injections irreversibly decreased lung compliance. The more concentrated solutions were used only occasionally because they tended to produce lung edema, and the experiments had to be terminated. Analysis of Results
In the reflex studies, control heart rate, blood pressure, tracheal tension, and duration of respiratory cycle before bronchial injection were compared with the corresponding values at the point of maximal change. In the afferent studies, control impulse frequencies were averaged over five ventilator cycles; maximum firing rates evoked by the solutions were averaged over one ventilator cycle at the peak of the response; both were expressed as im-
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 20, 2019.
2308
AIRWAY
Pulmonary
20
OSMOLARITY
AND LWNG AFFERENTS
B *
r
C-fibers IF
(imp/s)
Stimulated
i
4
0
0
7 17
7 8
0 8
*I
Bronchial C-fibers IF
1
(impts)
/
d Stimuiated Tested
Rapidly Adapting Receptors
3
5 9
9
7
9
PIG. 3. Responses (means t SE) of pulmonary C-fibers, bronchial C-fibers, and rapidly adapt.ing receptors to paired injections into a lobar bronchus of water (0 mmol/l) and isosmotic saline (A) and of hyperosmotic saline solutions and isosmotic saline (B). IF, impulse frequency. No. of fibers stimulated and tested and osmolarity of solutions (in mm01 /u are given beneath each column. 0, Control; l , peak response. Probability that change is same as that to isosmotic solution (300 mmol/l): * P < 0.05; ** P < 0.01.
c
50 ,F
(imp/s) T t+
0 Stimulated Tested
-/
5
18 19
13 19
9
6 9
9 9
0
300
300
1200
2400
pulses per second. No change in discharge was recorded if the peak frequency was not greater than the frequency during the most active cycle in the control period. Latenties of afferent and reflex responses were measured from the beginning of the bronchial injection. Results are expressed as means t SE. Because the parametric statistics used to analyze the data assume normal distribution and common variances, we analyzed the square roots of impulse frequencies to more closely approximate these requirements. The effect of each solution was compared with that of isosmotic saline by use of analysis of variance for repeated measures, followed when appropriate by the multiple-stage F test of Ryan, Elliot and Gabriel, and Welsch with CRUNCH statistical software. Statistical significance was accepted if P < 0.05. RESULTS
Reflex Effects
Injection of water or hyperosmotic sodium chloride solution (0.25-0.5 ml/kg) into a lobar bronchus in four dogs invariably decreased heart rate and arterial blood pressure and evoked apnea, usually followed by rapid shallow breathing (Fig. I). In each of the three dogs in which
tracheal tension was recorded, bronchial injection also caused contraction of tracheal smooth muscle (Fig. 1). These changes began 3-10 s after the beginning of the injection, cardiovascular and respiratory effects being maximal after lo-20 s and tracheal contraction after ZO40 s. The amplitude of the effects varied with the deviation of the injected solution from isosmolarity (Figs. 1 and 2). Isosmotic saline (300 mmol/l) had little or no effect on any of the variables except tracheal smooth muscle tension, which increased but by much less than after any of the other solutions (Figs. 1 and 2). We examined the effects of vagotomy on the responses to bronchial injection in three dogs. After the recurrent and pararecurrent laryngeal nerves were cut, injection of water or 2,400 mmol/l sodium chloride solution invariably had conspicuous effects on heart rate, blood pressure, breathing, and tracheal tension. After cervical vagotomy, injection of the same solutions had no effect on any of the recorded variables. Afferent Effects
We recorded the activity of 35 nonmyelinated and 50 myelinated afferent vagal fibers arising from endings located in the pulmonary lobe to which the bronchial ca th-
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 20, 2019.
AIRWAY
USMOLARITY
AND LUNG AFFERENTS
2309
AP
4. Responses of a pulmonary (A-D) and a bronchial (E) C-fiber to bronchial injection of solutions of various osmolarities. Pulmonary C-fiber stimulated by 1.0 ml/kg water (A) or hyperosmotic NaCl solutions (C and D) injected into left lower lobar bronchus. Stimulation by hyperosmotic solutions increased with deviation from isosmolarity, stimulation by 1,200 mmol/l saline lasting -12 s, that by 2,400 mmol/l saline 1 min. Note that same volume of isosmotic saline (300 mmol/l) had no effect (B). E: bronchial C-fiber stimulated by 0.6 ml/ kg of 1,200 mmol/l saline injected into right lower lobar bronchus. AP, action potentials recorded from strand of left (A-D) and right. (27) vagus nerve. Osmolarity of injected solutions is given in mmol/l beneath event marks. Note reflex decrease in HR and ABP in E. FIG.
300
C AP :mHPO)
20 o E1200
AP
2400
HR (beats/min)
ABP (mm&d
5s
180
,4o 150
50 was c rectea ana examinea tnelr response to nqection of water and hyperosmotic solutions. The concentration-dependence of the response was established by paired tests with 300,1,200, and 2,400 mmol/l saline and, in a few fibers, by tests with more concentrated solutions. Isosmotic saline was used as a control for the mechanical and temperature effects of the injections. Pulmonary C-fibers. We recorded the activity of 22 pulmonary C-fibers. Injection of water into a lobar bronchus stimulated all 17 pulmonary C-fibers tested (Fig. 3A). The evoked discharge, which consisted of a continuous or intermittent burst of impulses without respiratory modulation (Fig. 4A), began 7.7 t 1.6 s after the injection (the most rapid response beginning within 1 s) and lasted 19 t 3 s (range 9-44 s). Stimulation by water was not due only to the mechanical effects of the injected volume, because a similar volume of isosmolar saline (300 mmolll) had only minor effects (Fig. 3A), evoking a small increase in 7 of the 17 fibers and none in the others (Fig. 4B). Stimulation by water appeared to be due to hyposmolarity rather than to an absence of sodium or chloride ions. Thus the response of seven pulmonary C-fibers to isosmotic glucose was not significantly different from that to isosmotic saline, both being considerably less than that to water (Fig. 5A). Hyperosmotic saline stimulated 11 of 12 pulmonary C-fibers tested, evoking a continuous discharge (latency 5.5 + 1.1 s) that reached maximum intensity within 10 s of its onset. The discharge was often prolonged (Figs. 4D and 6A), its duration averaging 112 t 31 s (range 7-180 s) when 2,400 mmolll saline were injected. The intensity and duration of the evoked discharge usually increased
eter
witn tne concentration of the injected solution (Figs. 3B, 4, C and D, and 6A). In paired tests on eight pulmonary C-fibers, firing increased by 10.3 t 3.5 impulses/s in response to injection of 1,200 mmol/l saline and by 18.0 t 4.8 impulses/s in response to 2,400 mmol/l saline; it was virtually unaffected by isosmotic saline (Fig. 3B). Peak tracheal pressure usually increased by -1 cmH,O after injection of hyperosmotic saline, but the afferent response was unrelated to the ventilator cycle (Fig. 4, C and D) and often began before or in the absence of any change in pressure. Even when tracheal pressure increased progressively in the few experiments in which fluid was not aspirated from the airways between injections, the afferent response was determined by the osmolarity of the injected fluid rather than by the level of tracheal pressure (Fig. 6A). Bronchid C-fibers. Water stimulated 9 of the 13 bronchial C-fibers investigated, the latency and duration of the response being similar to that of pulmonary C-fibers; activity in these 13 fibers increased from 0.2 t 0.1 to 9.2 t 2.6 impulses/s. In some fibers the evoked discharge was continuous but irregular; in others it took the form of sporadic bursts. Effects of water and isosmotic saline were compared in paired tests on 10 bronchial C-fibers; isosmotic saline had little effect, causing a small increase in discharge in only three of the fibers (Fig. 3A). Hyperosmotic saline stimulated 10 of 12 bronchial Cfibers (Figs. 4E and 6B), the response being similar in pattern to that evoked by water but, as was the case with pulmonary C-fibers, generally more prolonged (duration 44 +- 9 s, maximum 200 s). Of the 10 bronchial C-fibers stimulated by hyperosmotic saline, six were stimulated
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.081.226.078) on January 20, 2019.
2310
AIRWAY
OSMOLARITY
AND LUNG AFFERENTS
of 1,200 and 2,400 mmol/l saline on nine receptors indicated that the response was concentration dependent (Figs. 3B and 6C). Water was by far the strongest stimulus to rapidly adapting receptors; in paired tests on 14 receptors, water IF increased discharge from 0.4 t 0.2 to 33.0 t 4.6 impulses/ (imp/s) s, whereas 2,400 mmol/l saline increased it from 0.9 t 0.4 to 18.6 t 4.3 impulses/s (P < 0.01). Slowly adapting pulmonary stretch receptors. In conl trast to their effects on other lower airway afferents, nei4 0 0 ther water nor hyperosmotic saline (1,200 and 2,400 mmol/l) had significant effects on most of the 26 slowly 5 adapting stretch receptors examined. Water stimulated c only 4 of 23 receptors, which began to discharge throughout deflation within l-9 s of the injection, thereby increasing mean activity (from 22 t 5 to 57 t 10 impulses/ s) while reducing the ventilatory modulation (Fig. 8A); IF the control pattern of discharge returned in 22-96 s. Two (imp/s) of the receptors stimulated by water were also stimulated, to a lesser extent, by isosmotic saline. Hyperosmotic saline (1,200 and 2,400 mmol/l) had 0 only minor effects on 21 stretch receptors, the changes consisting at most of small increases in peak frequency Glucose Water NaCI 300 mmol/l (
