Published ahead of print on April 7, 2005, doi:10.1164/rccm.200405-649OE
© 2005 American Thoracic Society doi: 10.1164/rccm.200405-649OE
A Century of Control of BreathingUniversity of Calgary, Calgary, Alberta, Canada Correspondence and requests for reprints should be addressed to John E. Remmers, Department of Medicine, Heritage Medical Research Building, 3330 Hospital Drive NW, Room 293, Calgary, AB, T2N 4N1 Canada. E-mail: jeremmer{at}ucalgary.ca Compared with other segments of respiratory physiology, control of breathing was poorly understood at the beginning of the twentieth century. In the first half of the century, knowledge slowly accumulated, and by midcentury, a nascent field of investigation was recognizable. In the second half of the century, this field blossomed with an array of basic discoveries, so that by the end of the century, study of control of breathing emerged as a multidisciplinary research focus that held substantial promise for contribution to understanding disease. THE TWENTIETH CENTURY OPENS WITH A THEORY At the end of the nineteenth century, investigators knew that an increase in blood CO2 or a decrease in O2 provoked an increase in ventilation. Haldane's groundbreaking experiments at the turn of the century showed that small increases in inspired CO2 concentration stimulated breathing, whereas a comparable ventilatory response to hypoxia was achieved only by a large decrease in inspired O2 concentration (1). In 1911, Winterstein (2) (Figure 1) suggested that CO2 stimulated breathing by acidifying extracellular fluid near the "respiratory centers," and Gesell (3) speculated that the intracellular concentration of hydrogen ion of respiratory neurons was the ultimate stimulus to breathing. Comroe (4) (Figure 2) and others searched for chemosensory areas by injecting into the brain solutions, which increased local PCO2 in inspiratory and expiratory centers. The results were inconsistent, however, and this approach was abandoned for 50 years.
Haldane's pioneering work led to the tidy scenario that CO2, acting exclusively in the brain, was the dominant chemoreflex stimulus and that hypoxia stimulated breathing by acidifying the brain. This was the dominant view in the first quarter of the twentieth century, but in the 1920s, two key observations indicated that this scenario was flawed. First, arterial pH was found to increase, not decrease, during hypoxia, indicating that something other than hydrogen ion was driving the ventilatory response to hypoxia. The glimmer of an entirely new dimension in chemoreflex control of breathing appeared in 1926, when de Castro (5) presented histologic evidence of a chemoreceptive function of the carotid body (Figure 3). In the 1930s, Heymans and Heymans made their revolutionary observation that anoxemia or hypercapnia of the aortic or carotid regions stimulated breathing. Although the ideas of Haldane and Winterstein did not fade easily, Gesell (6) enthusiastically heralded the new perspective:
If ever there was a conviction firmly entrenched in physiology, it was the monopoly of the chemical control of breathing by the respiratory center. But ... no one had actually demonstrated that either excess of carbon dioxide or lack of oxygen, restricted solely to the medulla, is capable of augmenting pulmonary ventilation. Such central stimulation was simply taken for granted. Indirect evidence was accepted as direct proof. It was a shaky foundation upon which all of us worked. So when Heymans et al. produced hyperpnea by a lack of oxygen or an excess of carbon dioxide confined to the aortic and carotid chemoreceptors, they gave to us a new outlook on respiration for which physiology is deeply indebted. A NEW SYNTHESIS BY MIDCENTURY At the end of World War II, Nielsen and Smith (7) returned to Haldane's approach, but now, independently controlling alveolar CO2 and O2. Extensive data in two subjects showed a linear dependence of ventilation on alveolar PCO2, and the slope of this relationship was increased by hypoxia. Dejours and coworkers (8) demonstrated in humans that an important ventilatory drive arises from O2 under normal conditions, raising questions about Haldane's view of the dominance of CO2 in chemoreflex control. The interaction of the hypoxic and hypercapnic stimuli in humans was quantitated by Cunningham (9) in the mid-1960s. Nerve traffic traveling from the peripheral chemoreceptors to the brain was recorded in the 1930s to 1950s and showed that anoxic, stagnant and histotoxic anoxia all increased impulse traffic (10–12). Comroe and Schmidt (10) proposed that a decrease in arterial PO2 rather than O2 content was the effective hypoxic stimulus to the peripheral chemoreceptors. Hornbein and coworkers (11) demonstrated that stimulus interaction between PO2 and PCO2 of arterial blood actually occurred at the carotid body. The requisite role of the peripheral chemoreceptor in the hypoxic ventilatory response was firmly established when Heymans and Neil (12) demonstrated that hypoxic hyperventilation was eliminated by interrupting the chemoreceptor afferents acutely in anesthetized animals. Gautier and Bonora (13) demonstrated chronic alveolar hypoventilation in awake cats after denervating the carotid body. CENTRAL CHEMORECEPTION: WAS WINTERSTEIN RIGHT? Leusen (14) initiated the modern investigation of central chemoreception in 1950, using perfusion of the cerebral ventricles with artificial cerebrospinal fluid in anesthetized dogs (14). He noted that increases in the hydrogen ion concentration of cerebrospinal fluid stimulated breathing. These seminal experiments stimulated Pappenheimer and others (15) (Figure 4) to perform similar experiments a decade later, but using awake, chronically instrumented goats. Ventilation of these animals responded briskly to changes in pH of cerebrospinal fluid caused by acute changes in bicarbonate concentration of the cerebrospinal fluid, and Pappenheimer found that ventilation correlated with a calculated hydrogen ion concentration of the brain extracellular fluid (16). Pappenheimer's synthesis was consistent with the discredited Winterstein theory and, surprisingly, revealed no role of the peripheral chemoreceptors in the ventilatory response to CO2.
Loeschcke and coworkers (17) also built on Leusen's results and, in 1958, located respiratory responsive areas in the lateral recesses of the fourth ventricle. In collaboration with Mitchell and colleagues (18), Severinghaus and Schlaefke areas on the surface of the ventral medulla were identified where an increase in hydrogen ion concentration stimulated breathing. In 1990, Issa and Remmers (19) resumed Comroe's search for specific regions within the brain sensitive to CO2. They found that injections of hypercapnic cerebrospinal fluid solutions into the ventral medulla produced prompt neuroventilatory responses. Nattie (20) identified a number of areas in the medulla and pons where increases in CO2 or H+ produced increases in ventilation. Amazingly, many of the responsive areas corresponded to known locations of respiratory neurons. The Winterstein hypothesis was further resuscitated by evidence that CO2 could directly excite synaptically isolated respiratory neurons (21). Thus, the story came full circle and Winterstein's ideas were substantiated. We are still uncertain regarding the relative importance of surface and deep receptors in the brainstem. THE PERIPHEREAL CHEMORECEPTION: WAS DE CASTRO RIGHT? De Castro (5) documented that the carotid body contained type I and II cells and sensory nerve endings. He proposed that the type I cells sense PO2 and PCO2 and release a neurotransmitter that activates the sensory nerve ending (Figure 3). However, dissenting views appeared over the next 50 years, some proposing that the type II cells or the nerve endings sensed changes in arterial PO2 and PCO2. In the last 10 years of the century, new methods were applied, yielding convincing evidence for the de Castro view—that is, that CO2 and O2 are sensed by the type I cell, which releases transmitters that depolarize neighboring sensory nerve endings, which then transmit neural impulses to the brain. A variety of possible neurotransmitters were examined, including acetylcholine, dopamine, and substance P, but unequivocal identification of key transmitters eluded investigators. How the type I cell senses a decrease in PO2 and an increase in PCO2/H+ ion concentration was also debated over the past 60 years. Winder (22) proposed that hypercapnia and hypoxia both increase hydrogen ion concentration around the sensory element. Like the Winterstein hypothesis, this idea proved correct for CO2 but not for hypoxia. Hypoxia may depolarize type I cells by inhibiting specific K+ channels of type I cells (23), or by acting on a heme protein, which triggers a cascade of intracellular events that ultimately decrease ATP levels. CO2 has been shown to decrease the intracellular pH of the type I cell, which may activate the Na+/H+ exchange, trigger Ca2+ influx and cause transmitter release (24). Thus, the century ends with an understanding that CO2 and hypoxia act on the type I cell by different mechanisms. How the interaction of the two stimuli occurs at the carotid body remains an enigma. HOW DO WE GENERATE TIDAL BREATHING? The twentieth century opened with knowledge that the brainstem (pons and medulla) generates regular breathing. In the 1920s and 1930s, Lumsden (25) discovered that the medulla alone produced an irregular respiratory rhythm and this was converted to a normal respiratory pattern by the action of two centers in the pons. Lumsden also found that the anoxic medulla generated "gasping," large abrupt bursts of inspiratory activity in thoracic and upper airway muscles, which Lawson and Thach (26) later showed to mediate recovery from anoxic apnea. Early in the twentieth century, stimulation of specific medullary sites led to the idea that breathing arose from discrete inspiratory and expiratory centers of the medulla (4), and this appeared to be confirmed by the recordings of bulbar respiratory neurons by Pitts (27). However, subsequent studies failed to show clear clustering of inspiratory and expiratory neurons. Cohen (28) showed that stimulation of the identified regions of the rostral pons could terminate inspiration or expiration, and he recorded phase-spanning discharge of respiratory neurons in these areas that plausibly explained the action of the rostral pons to produce a normal respiratory pattern. The century saw a great deal of interest in afferent traffic from the lungs traveling to the brain in the vagus nerve. Adrian (29), using single-fiber recordings, identified slowly adapting receptors, and Knowlton and Larrabee (30) provided evidence that these receptors sense lung volume. Sant' Ambrogio and Mortola (31) found them to be located in walls of large airways. These receptors accounted for the well-known dependence of inspiratory and expiration duration on lung volume. Knowlton and Larrabee (30) also inferred that rapidly adapting receptors, later found to be located in the submucosa, triggered the spontaneous deep breath, and Widdicombe (32) showed that these receptors could trigger cough and reactions to noxious agents. Finally, Paintal (33) discovered that nonmyelinated vagal fibers had sensory endings in the walls of alveoli that sense lung water. These afferents were subsequently shown to provoke rapid, shallow breathing and increase lung volume during pulmonary vascular congestion (34), and to play a vital role in newborns by preventing atelectasis and ensuring adequate alveolar ventilation (35). In the 1970s, a variety of investigators provided a picture of how the ventilatory pattern is determined (36). They showed that the tidal volume and the duration of inspiration were determined by the interaction of an "off-switch" threshold, with neural feedback derived either externally from sensory input or internally through the rostral pons. WHERE ARE THE "REAL" RESPIRATORY NEURONS? At midcentury, investigators understood little of the cellular mechanisms generating the respiratory rhythm. The methods for exploring such questions became available in the second half of the century, and deciphering the mysteries of respiratory rhythmogenesis became a search for the metaphoric Holy Grail. However, for technical reasons and because of the complexity of the neural circuitry, the search would prove long and arduous. The search began with pioneering contributions from Merrill (37) and Richter and colleagues (38). Merrill identified and located various types of bulbar respiratory neurons and identified the dorsal and ventral groups. He noted that the latter was segregated into two expiratory concentrations, one caudal and one rostral with inspiratory neurons lying in between. Richter and colleagues (38) provided the first high-quality, stable intracellular recordings from identified bulbar respiratory neurons. They showed that these neurons have abundant synaptic interconnections, and the patterns of synaptic activities indicated that the respiratory cycle consisted of three respiratory phases: inspiration and two sequential phases of expiration. Despite these advances, fundamental understanding remained elusive: too many interacting respiratory neurons in too many locations. Merrill (37) succinctly queried, "Where are the ‘real’ respiratory neurons?" In the closing decade of the century, Feldman's and Homma's studies in a simpler model, the in vitro rat brainstem, proved helpful. Smith and colleagues discovered a region of the medulla that contained pacemaker inspiratory neurons essential for rhythmogenesis (39). Onimatu and colleagues found an essential region containing preinspiratory neurons (40). Whether the respiratory rhythm generated by the in vitro brainstem is normal or gasping is uncertain. Nonetheless, the two bulbar oscillators seem to be important because local lesions cause apnea and respiratory failure in awake rats. Thus, the twentieth century closed with considerable understanding of respiratory rhythmogenesis. The Holy Grail? Maybe not. But, a very complex, vital rhythmogenic system has, at some level, been partially deciphered. No small achievement! IS THE "HIGHER" NERVOUS SYSTEM IMPORTANT? The complexity imposed by the "higher" nervous system became apparent through the remarkable clinical studies of Plum (41). Orem (42) recorded the activity of the bulbar respiratory neurons in intact, unanesthetized cats and found that they differed substantially from those in anesthetized or decerebrate animals. He also showed that during a voluntary, cortically commanded breath-hold, the brainstem respiratory oscillator is arrested, disproving the original concepts that corticospinal projections were the dominant pathway by which the cortex influenced breathing. St. John (43) showed that apneustic breathing caused by rostral pontine lesions disappeared when the animal awoke. Thus, neural structures above the pons can exert a dominant influence on the brainstem respiratory oscillator. Perhaps the most dramatic evidence of the respiratory role of the higher nervous system in humans was provided by Fink (44), who identified a "wakefulness factor" controlling breathing. Dempsey and coworkers (45) showed that, during sleep, but not during wakefulness, hypocapnia produces apnea. This led to the concept manifest in the work of Orem (42), that the cortex is postulated to excite and regulate bulbar respiratory mechanisms. RESPIRATORY SENSATIONS AND DYSPNEA Voluntary breath-holding was used early in the century to evaluate the role of the cortex in control of breathing. Hill and Flack (46) found that "air hunger" induced by breath-holding depended on the intensity of chemical stimulation. In addition, they observed that rebreathing from a bag containing high CO2 and low O2 satisfied this air hunger, indicating that the lack of breathing movements was also a major determinant of dyspnea during breath-holding. The interaction of chemical stimuli and breathing sensations was subsequently confirmed by Fowler (47) and later extended to continuous, controlled breathing (48). Experience with mechanical loading of breathing revealed that the restriction of the act of breathing plays a major role in the feeling of urge to breathe, shortness of breath, or air hunger. Experiments on awake humans showed that most individuals experience dyspnea while breathing CO2, and the tolerance of this unpleasant sensation increases when tidal volume or respiratory frequency increases either voluntarily or by mechanical ventilation (49). Thus, investigators came to understand that the awake brain continuously monitors afferent information generated by the motor act of breathing. These perceptive inferences about conscious perceptions of respiratory sensory information were provided a rigorous scientific foundation by Bakers and Tenney (50) in 1970 (Figure 5). They reported that normal subjects can accurately scale perceptions related to voluntary ventilatory acts, such as change in volume, pressure, or muscle force. This has been confirmed and extended by others and has led to investigations of the cortical projections from mechanoreceptors in the lung and chest wall. Thus, breathlessness was believed to be related to a relative deficiency of sensory information from the chest wall and lung in relation to a demand for ventilation. Campbell and Howell (51) proposed a related concept called length–tension inappropriateness, where dyspnea was caused by a mismatch between outgoing motor command to the respiratory muscles and incoming afferent information. This hypothesis is now generally accepted and has been supplemented by evidence from Guz (52) that vagal afferents play a role in dyspnea. Experiments on normal subjects and on patients with a variety of pulmonary diseases suggest that, under a certain level of respiratory drive, the brain "expects" a certain level of ventilation and its associated afferent traffic. A deficiency of the latter causes or intensifies dyspnea. Thus, the century closed with an understanding that cortical systems monitoring respiratory movements and effort are undoubtedly involved in dyspnea, but the sensory receptors and pathways responsible for this sensation are unknown.
SLEEPING AND BREATHING Certainly, one of the most important applications of the new knowledge of control of breathing was in the field of sleep. Bulow (53) was the first to systematically explore the effect of sleep on breathing. He showed that sleep depressed the hypercapnic ventilatory response, which led to the idea that sleep depressed respiratory chemoreflexes. However, Bulow's pioneering work antedated a major advance in understanding of sleep—namely, the discovery that sleep comprises two fundamentally different states, REM and non-REM sleep, the former being associated with generalized muscle atonia. Phillipson and colleagues (54) examined chemoreflex responses in REM and non-REM in dogs. They observed that non-REM and REM had drastically different effects on the chemoreflex control of breathing. The ventilatory responses to hypercapnia and hypoxia changed little during non-REM compared with wakefulness. By contrast, they noted a profound reduction in chemoreflex responsiveness during REM. This may correlate with Nattie's (20) studies, which showed that the chemosensitivity of discrete regions of the brainstem differs between wakefulness and sleep. Thus, the nervous system may regulate regional chemosensitivity in a state-specific way, even though the global hypercapnic ventilatory response does not differ greatly between wakefulness and non-REM sleep. In humans, Skatrud and Dempsey (55) indicated the observed reduction in chemoresponsiveness during non-REM may result from mechanical loading caused by increased upper airway resistance. That is, in humans, narrowing of the pharynx during sleep appears to be an important determinant of ventilatory responses. Phillipson and colleagues (54) arrived at a stunning synthesis: they proposed that, in non-REM, metabolic control is dominant, whereas in REM, a behavioral type of ventilatory control is driven by the cortex and nonrespiratory motor control systems. SLEEP APNEA: A NEW GROUP OF RESPIRATORY DISORDERS Early in the twentieth century, sleep was known to cause periodic breathing at altitude, and Cheyne-Stokes breathing was known to be an ominous sign of deteriorating heart failure. In the mid-1950s, two types of alveolar hypoventilation in patients with normal lung function were described: one, which was in the setting of normal body habitus (Ondine's curse), was discovered to be a form of central sleep apnea (56); the other occurred in the setting of obesity (Pickwickian syndrome) and was associated with daytime sleepiness. The causes of alveolar hypoventilation were obscure in both syndromes. Thus, by midcentury, little evidence indicated that sleep might pose a challenge for breathing in a large segment of the population. A breakthrough in understanding sleep apnea occurred in 1978, when Remmers and others (57) described the pathogenesis of airway occlusion in obstructive sleep apnea. Occlusion was shown to occur in the pharynx and was attributed to a reduction in the dilator activity of the pharyngeal muscles together with negative pharyngeal luminal pressure caused by inspiratory inflow. Structural narrowing of the pharynx was demonstrated to be the underlying cause because patients with obstructive sleep apnea had increased closing pressure and decreased maximum area of the passive pharynx. Young and coworkers (58) provided the most surprising finding—namely, that obstructive sleep apnea was extremely common in the general population (4% of females and 9% of males), and that only half of all individuals with apnea experience daytime symptoms. Sullivan and coworkers (59) made a key discovery in this field, showing that nasal continuous positive airway pressure provided benign and effective treatment of obstructive sleep apnea. Thus, a common disease was discovered that compromised quality of life and was treatable. It was shown to cause hypertension, to be a risk factor for vehicular accidents, and to have a role in cardiovascular disease. A new field, sleep-disordered breathing, had been created. Younes (60) presented a crystalline, comprehensive synthesis of factors contributing to ventilatory instability during sleep, one which integrated central and obstructive processes and provides a conceptual foundation for the new field. REFLECTIONS AND PERSPECTIVES At the end of the twentieth century, we are left with detailed knowledge about components of the respiratory control system but with large gaps in understanding how these components interact to produce normal or abnormal function. In this regard, Tenney (Figure 5) urged us to be mindful of the functionality of the entire system: The physiologist keeps the whole always in mind. He accepts the tactical necessity of reductionism to understand the parts, but, once done, it is for him only the beginning, never the end. Synthesis is his overriding strategy. (Marsh Tenney, 1969) Although the twenty-first century will continue to see progress in understanding elemental function, the need for synthesis is clear if we are to understand how the system functions in disease states. FOOTNOTES Conflict of Interest Statement: J.E.R. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form May 19, 2004; accepted in final form April 5, 2005 REFERENCES
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
