Catecholamine Release and Excretion in Rats with Immunologically Induced Preganglionic Sympathectomy

Abstract: Plasma and urinary catecholamines were quantified to assess global sympathoadrenal function in rats with preganglionic lesions caused by antibodies to acetyl-cholinesterase (AChE). Rats were given intravenous injections of normal mouse IgG or murine monoclonal anti-acetylcholinesterase IgG (1.5 mg). Five or 16 days afterward, basal blood samples were taken through indwelling arterial cannulae. A few hours later, the rats were immobilized for 10 min in padded restrainers, and another blood sample was drawn. HPLC determinations showed low basal levels of norepinephrine and epinephrine (<0.2 ng/ml in all rat plasma samples). In control rats, immobilization stress increased levels of plasma catecholamines up to 35-fold. In rats tested 5 days after injection of antibody, the norepinephrine response was much smaller (15% of control), and (he epinephrine response was nearly abolished (5% of control). There was some recovery at 16 days after antibody treatment, but stress-induced catecholamine release was still markedly impaired. Reduced stress-induced release: was not accompanied by major changes in tissue epinephrine or norepinephrine (heart, spleen, adrenal glands, and brain), although adrenal dopamine content dropped by 60%. Urinary excretion was studied in parallel experiments to gain insight into the effects of AChE anti-bodies on basal sympathoadrenal activity. Epinephrine, norepinephrine, dopamine, and selected metabolites were quantified in 24-h urine samples collected at frequent intervals for 30 days after antibody injection. No statistically gnificant changes were detected in the urinary output of dopamine, 3-methoxytyramine, normetanephrine, or 3-methoixy-4-hydroxyphenylglycol. On the other hand, epinephrine and norepinephrine output increased sharply at the time of antibody injection and then fell significantly below control levels. Norepinephrine output returned to normal after 2 weeks, but epinephrine output remained depressed. These results are consistent with previous evidence of widespread and persistent antibody-mediated βmade to the preganglionic sympathetic system.     

Adrenergic response to intense muscular activity in sedentary subjects as a function of emotivity and training]

Seven male sedentary human subjects were studied during intense muscular work (80% of maximal oxygen uptake) performed either for 15 min or until exhaustion (mean duration: 47 +/- 2 min). Plasma catecholamines were estimated before and after the experiment by means of an original fluorimetric assay. Epinephrine or norepinephrine were individually isolated from plasma and assayed in single extracts by a highly sensitive fluorimetric method. Epinephrine and norepinephrine levels as low as 15 ng per liter were detectable by this procedure in human plasma. The adrenergic pattern was found to be greatly different from one subject to another and related to emotivity: the effect of this factor was revealed by the predominance of epinephrine in plasma at rest or under exercise (ratio NA/A less than 1). In nonemotive subjects (ratio NA/A greater than 1 at rest) plasma epinephrine and norepinephrine increased progressively during exercise. Increments after exercise were higher for norepinephrine changes; however, the fact that epinephrine concentrations correlated significantly with norepinephrine suggests a simulataneous and coordinated stimulation of adrenal glands and orthosympathetic nervous system. In emotive subjects (ratio NA/A less than 1 at rest) the apprehension of muscular work promoted a difference in catecholamine responses: norepinephrine release was not affected by subject's anxiety, while epinephrine secretion, already elevated before the test, reached a high degree of magnitude in the first minutes of muscular work, remaining nearly constant until exhaustion. Physical training of nonemotive subjects, during 2 months with two intense exercises by a week, reduced strongly norepinephrine release after exhaustive muscular work. In the same conditions, the adrenal-medullary response was not significantly modified when compared with untrained subjects. Our results suggest that the adrenergic behaviour during exercise is a function of effort intensity to be supplied; catecholamines seem to be important factors in regulating body homeostasy during muscular work in man. In addition, emotive subjects exhibit amplified adrenal-medullary response, which may be related to psychological stimuli.     

Heat and water flux in the intrathoracic airways and exercise-induced asthma

To explore the relationship between the flux of heat and water within the respiratory tract during exercise and recovery to the development of exercise-induced asthma (EIA), we recorded airstream temperature at multiple points throughout the tracheobronchial tree in 10 normal and 10 asthmatic subjects before, during, and after cycle ergometry. In both groups, the intra-airway temperature fell progressively as ventilation increased, and there were no significant differences between the thermal profiles of the two populations at rest or during exercise. Calculation of water losses and the osmolality of the airway surface fluid failed to demonstrate significant airway drying in either group. With cessation of the work load, the airstream temperature increased abruptly, rising two times more rapidly in the asthmatics than the normals. Since the major source of heat in these experiments is the bronchial circulation, our findings suggest a reaction sequence consisting of vasoconstriction and airway cooling during exercise followed by a rapid resupply of heat when exercise ceases. The latter may cause the hyperplastic capillary bed in the airways of asthmatics to develop an exaggerated rebound hyperemia which may lead to airway edema and EIA.     

Effect of increased central blood volume with water immersion on plasma catecholamines during exercise

To examine the influence of an increase in central blood volume with head-out water immersion (WI) on the sympathoadrenal response to graded dynamic exercise, nine healthy men underwent upright leg cycle exercise on land and with WI. Plasma norepinephrine and epinephrine concentrations were used as indexes of overall sympathoadrenal activity. Oxygen consumption (VO2), heart rate, systolic blood pressure, and plasma concentrations of norepinephrine, epinephrine, and lactate were determined at work loads corresponding to approximately 40, 60, 80, and 100% peak VO2. Peak VO2 did not differ on land and with WI. Plasma norepinephrine concentration was reduced (P less than 0.05) at 80 and 100% peak VO2 with WI and on land, respectively. Plasma epinephrine and lactate concentrations were similar on land and with WI at the three submaximal work stages, but both were reduced (P less than 0.05) at peak exertion with WI. Heart rate was lower (P less than 0.05) at the three highest work intensities with WI. These results suggest that the central shift in blood volume with WI reduces the sympathoadrenal response to high-intensity dynamic exercise.     

Effect of dietary proteins and carbohydrates on urinary and sympathoadrenal catecholamines

We previously observed that administration of tyrosine to rats or humans elevated urinary dopamine, norepinephrine and epinephrine levels. The present studies examine the effects on these urinary catecholamines of varying the ratio of protein to carbohydrate in the diets.Rats consumed diets containing 0, 18 or 40% protein (76, 58 and 36% carbohydrate respectively) for 8 days. The stress of consuming the protein-free food was associated with a 16% weight reduction, and with significantly lower serum, heart and brain tyrosine levels than those noted in rats eating the 18 or 40% protein diets. Absence of protein from the diet also decreased urinary levels of dopamine and DOPA but increased urinary norepinephrine and epinephrine, probably by increasing sympathoadrenal discharge; it also increased the excretion of DOPA in animals pretreated with carbidopa, a DOPA decarboxylase inhibitor. Carbidopa administration decreased urinary dopamine, norepinephrine and epinephrine as expected; however, among carbidopa-treated rats urinary norepinephrine and epinephrine concentrations were highest for animals consuming the protein-free diet, again suggesting enhanced release of stored catecholamines from sympathoadrenal cells. The changes in urinary catecholamines observed in animals eating the protein-free diet were similar to those seen in rats fasted for 5 days: dopamine levels fell sharply while norepinephrine and epinephrine increased.These data indicate that the effects of varying dietary protein and carbohydrate contents on dopamine secretion from peripheral structures differ from its effects on structures secreting the other two catecholamines. Protein consumption increases dopamine synthesis and release probably by making more of its precursor, tyrosine, available to peripheral dopamine-producing cells; it decreases urinary norepinephrine and epinephrine compared with that seen in protein-deprived animals, probably by diminishing the firing of sympathetic neurons and adrenal chromaffin cells.     

Plasma free and sulfoconjugated catecholamines during sustained exercise

Previous research established a relationship between circulating sulfoconjugated norepinephrine (NE-SO4) and oxygen consumption at various exercise intensities. In this study, the stability of the NE-SO4 response was examined during sustained exercise at a constant relative intensity. Seven trained men bicycled at 78 +/- 3% of their maximal O2 consumption for 28 min and then rested on the ergometer for a comparable duration. After a 30-min rest, plasma samples were collected through an indwelling catheter at 7-min intervals during the exercise and recovery periods. Free NE and epinephrine increased sixfold during exercise. These changes were accompanied by increases in sulfoconjugated catecholamines, but only NE-SO4 achieved statistical significance (rest, 712 +/- 602; exercise, 1,329 +/- 1,163 pg/ml). This occurred at three collection periods (14, 21, and 28 min). Approximately 35, 52, and 95% of NE, epinephrine, and dopamine, respectively, existed as sulfoconjugated during exercise. Subject variation was present in the sulfoconjugated catecholamine response that could not be attributed to corresponding differences in circulating free catecholamine release. These findings implicate blood flow as a factor in the sulfoconjugation of NE, but not epinephrine or dopamine.     

Plasma free and sulfate conjugated catecholamine levels during acute physiological stimulation in man

The responses of plasma free and sulfate-conjugated catecholamines to acute physiological stimulation was examined in normal male subjects. Catecholamines were measured with a sensitive radioenzymatic assay incorporating simultaneous hydrolysis of sulfate conjugates and O-methylation of free norepinephrine and epinephrine. Following 20 minutes recumbency after venepuncture 30 +/- 3% of norepinephrine and 16 +/- 5% of epinephrine was in thr free form. Free catecholamines generally increased during standing, cold immersion and isometric handgrip, but sulfates did not change. Bicycle ergometry markedly increased free catecholamines which rapidly returned to basal levels at the end of exercise. In contrast, sulfated norepinephrine decreased substantially with exercise in all subjects but returned to basal levels 3 minutes after stopping exercise. Epinephrine sulfate varied considerably between subjects but showed a similar, although smaller, fall with exercise. Thus, during physiological stimulation, which caused increases in free norepinephrine and epinephrine levels in plasma, the only consistent change in sulfated catecholamines was a marked fall in norepinephrine sulfate after bicycle exercise. This may indicate saturation of sulfotransferase activity, substrate inhibition or impaired tissue conjugation.     

Evaluation of pulmonary resistance and maximal expiratory flow measurements during exercise in humans.

To evaluate methods used to document changes in airway function during and after exercise, we studied nine subjects with exercise-induced asthma and five subjects without asthma. Airway function was assessed from measurements of pulmonary resistance (RL) and forced expiratory vital capacity maneuvers. In the asthmatic subjects, forced expiratory volume in 1 s (FEV1) fell 24 +/- 14% and RL increased 176 +/- 153% after exercise, whereas normal subjects experienced no change in airway function (RL -3 +/- 8% and FEV1 -4 +/- 5%). During exercise, there was a tendency for FEV1 to increase in the asthmatic subjects but not in the normal subjects. RL, however, showed a slight increase during exercise in both groups. Changes in lung volumes encountered during exercise were small and had no consistent effect on RL. The small increases in RL during exercise could be explained by the nonlinearity of the pressure-flow relationship and the increased tidal breathing flows associated with exercise. In the asthmatic subjects, a deep inspiration (DI) caused a small, significant, transient decrease in RL 15 min after exercise. There was no change in RL in response to DI during exercise in either asthmatic or nonasthmatic subjects. When percent changes in RL and FEV1 during and after exercise were compared, there was close agreement between the two measurements of change in airway function. In the groups of normal and mildly asthmatic subjects, we conclude that changes in lung volume and DIs had no influence on RL during exercise. Increases in tidal breathing flows had only minor influence on measurements of RL during exercise. Furthermore, changes in RL and in FEV1 produce equivalent indexes of the variations in airway function during and after exercise.     

Effects of clentiazem (TA-3090) and nifedipine on basal circulating catecholamine levels and on stimulation-evoked adrenal catecholamine secretion in anesthetized dogs.

The effects of TA-3090 (clentiazem) and nifedipine on basal sympathoadrenal activity and on the adrenal medullary response during splanchnic nerve stimulation were studied in dogs anesthetized with sodium pentobarbital. Plasma concentrations of epinephrine and norepinephrine were measured in aortic and adrenal venous blood before and after acute administration of the drugs, as well as during left splanchnic nerve stimulation before and after administration of drugs. Following intravenous injections, TA-3090 (30, 100, and 300 micrograms/kg) did not affect basal circulating catecholamine levels, whereas nifedipine (10, 30, and 100 micrograms/kg) markedly increased aortic epinephrine and norepinephrine concentrations in a dose-dependent manner in correlation with progressive decreases in mean arterial pressure. The changes in aortic epinephrine and norepinephrine concentrations were inversely related to those in mean arterial pressure (r = 0.603, p < 0.01; r = 0.536, p < 0.01; respectively). In response to direct splanchnic nerve stimulation (2 Hz, 2 ms, 1 min, 12 V), adrenal venous epinephrine and norepinephrine concentrations significantly increased, with a high degree of reproducibility. The catecholamine responses to splanchnic nerve stimulation were not affected by either TA-3090 or nifedipine at any dose tested. The present results suggest that the increases in circulating catecholamine levels following nifedipine administration are due to baroreflex activation secondary to the drug-induced hypotension. The study indicates that both TA-3090 and nifedipine did not significantly affect L-type Ca2+ channels related to catecholamine release in the adrenal medulla under the present experimental conditions.     

Effects of exercise on plasma catecholamine levels in the toad, Bufo paracnemis: role of the adrenals and neural control

Resting plasma epinephrine (E) and norepinephrine (N) concentrations for intact toads (Bufo paracnemis) were 5.57+/-1.0 and 0.88+/-0.38 ng/ml, respectively. Exercise induced a significant increase in heart rate, blood pressure and plasma epinephrine (about 4.3 times), whereas norepinephrine remained unchanged. The resting [E]/[N] ratio was 6.3 and increased to 32.9 during exercise. Adrenal denervation did not alter the basal plasma catecholamine or norepinephrine levels after exercise, but prevented the increase in epinephrine during exercise, suggesting that in the intact toad this increase is due to adrenal secretion whereas resting norepinephrine may be liberated by extra-adrenal chromaffin tissues. This also suggests that the adrenal glands can release selectively the two catecholamines. The increases in heart rate and blood pressure in denervated toads were not significantly different from those of intact animals, suggesting that during exercise the sympathetic nerves play the main role in inducing cardiovascular responses. Spinal transection induced a significant increase in basal norepinephrine levels, which remained elevated after exercise. Since spinal toads are unable to perform spontaneous movements it is possible that this increase may be caused by this stressful condition. The increases in heart rate and blood pressure observed in spinal toads during exercise may be due to direct mechanical effects of venous return on the heart.     

Catecholamine and blood pressure regulation by gonadotropin-releasing hormone analogs in amphibians

Analogs of gonadotropin-releasing hormone (GnRH) occur in the brain, plasma, and sympathoadrenal system of anuran amphibians. The present experiments studied the effects of GnRH and [Trp7, Leu8]-GnRH on plasma catecholamines and cardiovascular function in conscious adult bullfrogs (Rana catesbeiana) and cane toads (Bufo marinus). Both GnRH analogs elicited dose-dependent (0.1-1 nmol.kg-1) increases in arterial norepinephrine, epinephrine, and blood pressure levels when injected intravenously into toads. In bullfrogs, [Trp7, Leu8]-GnRH (1 nmol.kg-1) increased arterial norepinephrine concentration approximately 10-fold without affecting the concentrations of norepinephrine sulfate, norepinephrine glucuronide, epinephrine, epinephrine sulfate, or epinephrine glucuronide. The noradrenergic response of bullfrogs to [Trp7, Leu8]-GnRH was specific to the neurohormone because it could be inhibited by [D-pGlu1, D-Phe2, D-Trp3,6]-GnRH. The sympathomimetic activities of the GnRH analogs did not depend on changes in temperature, which occur seasonally in natural habitats, because similar noradrenergic responses were observed at 4 and 22 degrees C. GnRH and [Trp7, Leu8]-GnRH (0.01-10 nmol.kg-1) did not raise arterial blood pressure in bullfrogs despite their pressor actions in toads. This interspecific difference was remarkable because cardiovascular responses to norepinephrine, angiotensin II, and vasotocin in bullfrogs were similar to those in toads. The parallels between catecholamine and blood pressure responses suggest that epinephrine is the principal mediator of the blood pressure response to native GnRH analogs in toads. In bullfrogs, [Trp7, Leu8]-GnRH mobilizes norepinephrine but not epinephrine, and the noradrenergic effect is insufficient to raise blood pressure. These observations are consistent with a physiological role for native GnRH analogs in the regulation of the sympathoadrenal system in anuran amphibians.     

Plasma catecholamine and lactate response during graded exercise with varied glycogen conditions.

The relationships between the lactate threshold (TLa), plasma catecholamines, and ventilatory threshold (TVE) were examined under normal and glycogen-depleted conditions. Nine male subjects performed a graded exercise test on a bicycle ergometer in a normal glycogen (NG) state and in a glycogen-depleted (GD) state to determine if manipulation of muscle glycogen content would affect their ventilatory, lactate, and catecholamine responses. High correlations were found between plasma lactate and the two catecholamines, epinephrine (r = 0.964) and norepinephrine (r = 0.965) under both conditions. The GD protocol resulted in a shift in the TLa to a later work rate; inflections in epinephrine and norepinephrine shifted in a coordinated manner. TVE and TLa occurred at similar work loads under NG conditions [67.2 +/- 1.5 and 65.6 +/- 2.3% maximal oxygen consumption (VO2max), respectively], but TLa occurred at a later work load (75.3 +/- 1.9% VO2max) compared with TVE (68.3 +/- 1.6% VO2max) under GD conditions. These results suggest a causal relationship between plasma lactate and epinephrine during a graded exercise test under the glycogen conditions studied. Although an association existed between ventilation and lactate, this relationship was not as strong.     

Bronchial reactivity in asthmatic adults with normal spirometric values.

A comparison was made between airways obstruction produced by exercise. Hyperventilation of cold air, hyperventilation of room air, and inhalation of ultrasonically nebulised distilled water. Seven asthmatics with no airflow obstruction were studied together with seven normal controls. All four provocations produced comparable mean reductions in forced expired volume in one second in the asthmatic patients, but no reduction was obtained in the controls. Whereas exercise, hyperventilation of cold air, and hyperventilation of room air were associated with heat loss from airways, inhalation of nebulised distilled water was not. Inhalation of ultrasonically nebulised distilled water offers a simple additional diagnostic test for asthma. Changes in the osmolality of airway surface liquid may possibly be the common mechanism by which "fog" and hyperventilation produce bronchoconstruction.     

Plasma catecholamines in the hypoxaemic fetal rhesus monkey

To assess the response of the sympathoadrenal system of the primate fetus to oxygen deprivation, we measured plasma catecholamines in 8 chronically catheterized fetal rhesus monkeys. A range of fetal hypoxaemia was produced by having the mother inspire 15, 10, or 9% oxygen mixtures while tranquilized with ketamine. Catecholamines from fetal carotid and maternal femoral arteries were measured by radioenzymatic assay. Fetal plasma norepinephrine and epinephrine concentrations increased significantly at all levels of hypoxaemia, but dopamine increased only at very low fetal oxygen tensions. Norepinephrine levels exceeded those of epinephrine and dopamine under all conditions. Relatively more severe hypoxaemia was necessary to elevate concentrations of epinephrine above baseline as compared with norepinephrine. A negative exponential correlation (P less than 0.001) was found between both fetal arterial PO2 and oxygen content and plasma norepinephrine and epinephrine, which was qualitatively similar to that observed previously in the sheep fetus. Maternal catecholamines were found to increase during hypoxaemia as well, but to a lesser degree than in the fetus.     

Upregulation of Phagocyte-Derived Catecholamines Augments the Acute Inflammatory Response

Following our recent report that phagocytic cells (neutrophils, PMNs, and macrophages) are newly discovered sources of catecholamines, we now show that both epinephrine and norepinephrine directly activate NFκB in macrophages, causing enhanced release of proinflammatory cytokines (TNFα, IL-1β, IL-6). Both adrenal-intact (AD+) and adrenalectomized (ADX) rodents were used, because ADX animals had greatly enhanced catecholamine release from phagocytes, facilitating our efforts to understand the role of catecholamines released from phagocytes. Phagocytes isolated from adrenalectomized rats displayed enhanced expression of tyrosine-hydroxylase and dopamine-β-hydroxylase, two key enzymes for catecholamine production and exhibited higher baseline secretion of norepinephrine and epinephrine. The effects of upregulation of phagocyte-derived catecholamines were investigated in two models of acute lung injury (ALI). Increased levels of phagocyte-derived catecholamines were associated with intensification of the acute inflammatory response, as assessed by increased plasma leak of albumin, enhanced myeloperoxidase content in lungs, augmented levels of proinflammatory mediators in bronchoalveolar lavage fluids, and elevated expression of pulmonary ICAM-1 and VCAM-1. In adrenalectomized rats, development of ALI was enhanced and related to α₂-adrenoceptors engagement but not to involvement of mineralocorticoid or glucocorticoid receptors. Collectively, these data demonstrate that catecholamines are potent inflammatory activators of macrophages, upregulating NFκB and further downstream cytokine production of these cells. In adrenalectomized animals, which have been used to further assess the role of catecholamines, there appears to be a compensatory increase in catecholamine generating enzymes and catecholamines in macrophages, resulting in amplification of the acute inflammatory response via engagement of α₂-adrenoceptors.     

Low plasma concentrations of ionized calcium in patients with asthma

It has been suggested that calcium homeostasis is abnormal in the vascular smooth muscle of hypertensive patients and in the bronchial smooth muscle in asthmatics. We have found the mean baseline concentration of plasma ionized calcium to be significantly lower both in 12 asthmatics with exercise-induced asthma (EIA) [1.16 +/- 0.01 (SE) mmol/l, P less than 0.001] and in 20 asthmatics without EIA (1.16 +/- 0.01; P less than 0.001) compared with 42 healthy subjects (1.24 +/- 0.01). The mean concentrations of plasma ionized calcium were not significantly different in asthmatics with and without EIA when measured either before treadmill exercise, during the last seconds of this exercise, or 10 or 20 min after exercise but were significantly lower than in another seven healthy subjects who undertook the same exercise protocol. Total plasma calcium concentrations in the three exercising groups were not significantly different at any point in time. The results suggest that in bronchial asthma an alteration of calcium metabolism may be important, but they also suggest that there is no simple relationship between the plasma ionized calcium concentration and acute exercise-induced bronchoconstriction.     

Effects of pre-exercise listening to slow and fast rhythm music on supramaximal cycle performance and selected metabolic variables

We examined the effect of listening to two different types of music (with slow and fast rhythm), prior to supramaximal cycle exercise, on performance, heart rate, the concentration of lactate and ammonia in blood, and the concentration of catecholamines in plasma. Six male students participated in this study. After listening to slow rhythm or fast rhythm music for 20 min, the subjects performed supramaximal exercise for 45 s using a cycle ergometer. Listening to slow and fast rhythm music prior to supramaximal exercise did not significantly affect the mean power output. The plasma norepinephrine concentration immediately before the end of listening to slow rhythm music was significantly lower than before listening (p < 0.05). The plasma epinephrine concentration immediately before the end of listening to fast rhythm music was significantly higher than before listening (p < 0.05). The type of music had no effect on blood lactate and ammonia levels or on plasma catecholamine levels following exercise. In conclusion, listening to slow rhythm music decreases the plasma norepinephrine level, and listening to fast rhythm music increases the plasma epinephrine level. The type of music has no impact on power output during exercise.     

The effect of a MAO inhibitor on reserpine-induced secretion of catecholamines from the adrenal medulla of the rose-ringed parakeet

The effect of MAO inhibitor on reserpine-induced release of catecholamines from the adrenal medulla of the rose-ringed parakeet was investigated in this biochemical, histochemical, and ultrastructural study. The MAO inhibitor provided some blockade of epinephrine secretion and was ineffective in blocking norepinephrine secretion. Differences between birds and mammals were noted, suggesting that different mechanisms of catecholamine release may be operating in the two classes.     

Specificity of Exercise in Exercise-induced Asthma

Ventilatory function after three types of exercise—running, cycling, and swimming—was studied in 10 control subjects and 40 asthmatic patients. All performed eight minutes of submaximal aerobic exercise during each of the programmes, which were conducted in a randomly selected order. Biotelemetric monitoring of heart rates was used to equate the intensity of the exertion undertaken during the three systems of exercise. No control subject showed any significant variation in ventilatory capacity after exercise, and the responses after the three forms of exercise did not differ.In asthmatics exercise-induced asthma was observed after 72·5% of running tests, 65% of cycling tests, and 35% of swimming tests. In addition, those patients who developed exercise-induced asthma after swimming were noted to have significantly smaller falls in FEV₁ levels than were recorded after running and cycling. These results were statistically significant (P <0·01).The unexplained aetiology of increased airways resistance after exercise in asthmatics is discussed. This study indicates that swimming should be recommended in preference to running or cycling as an exercise programme for adults and children with asthma.