Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The control of episodic breathing in the bullfrog (Rana catesbeiana) Kinkead, Richard 1994

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1995-059898.pdf [ 4.19MB ]
JSON: 831-1.0088105.json
JSON-LD: 831-1.0088105-ld.json
RDF/XML (Pretty): 831-1.0088105-rdf.xml
RDF/JSON: 831-1.0088105-rdf.json
Turtle: 831-1.0088105-turtle.txt
N-Triples: 831-1.0088105-rdf-ntriples.txt
Original Record: 831-1.0088105-source.json
Full Text

Full Text

THE CONTROL OF EPISODIC BREATHING IN THE BULLFROG(Rana catesbeiana)byRICHARD KINKEADB.Sc. (lions. Biology), The University of Ottawa, 1988M.Sc. (Biology), The University of Ottawa, 1990A THESIS SUBMITI’ED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIES(Department of Zoology)We accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAJuly, 1995© Richard Kinkead, 1995In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)_____________________________Department of 2ook&The University of British ColumbiaVancouver, CanadaDate_____________DE-6 (2)88)ABSTRACTLike many lower vertebrates, bullfrogs breathe episodically; i.e. breathing occurs in boutsthat are interrupted by periods of apnea. This breathing pattern contrasts with the continuousalternation between inspiration and expiration usually seen in mammals. Little is known aboutthe mechanisms underlying episodic breathing; hence, the objective of this thesis was toinvestigate respiratory control in the bullfrog to determine what causes the onset/termination ofthe episodes of breathing. A variety of preparations, ranging from the “ta (whole) animalto the in vitro brainstem-spinal cord preparation were used. This reductionist approach madeit possible to assess the contribution of the different components of the respiratory control systemtowards pattern formation.The initial studies focused on the role of afferent feedback from different groups ofperipheral receptors. In these experiments, receptor input was either eliminated by denervationor manipulated artificially. The results have shown that each receptor group influences patterndifferently, either by affecting the number of breaths in an episode, or the duration of the apneic(non-ventilatory) period. However, none of the receptor groups investigated were responsiblefor clustering the breaths into episodes. These results were subsequently confirmed by recordingthe respiratory-related motor output from an in vitro brainstem-spinal cord preparation, whichproduced a “fictive” breathing pattern that was episodic, and virtually identical to that of intactfrogs. Because this preparation is essentially devoid of afferent feedback (except for centralchemoreceptors) and descending inputs but still can produce breathing episodes, it was concludedthat the episodic breathing pattern of the bullfrog is an intrinsic property of the central nervoussystem, and can occur without peripheral feedback.The final study assessed the role of the nucleus isthmi in breathing pattern formation.This nucleus is located between the midbrain roof and the base of the cerebellum. Bilateral11microinjections of the neurotoxin kainic acid in the nucleus isthmi area significantly reduced thebreathing frequency, and the breathing pattern consisted mainly of evenly-spaced single breaths.This suggests that the nucleus isthmi provides the tonic drive to breathe, which is a key elementin the production of breathing episodes of more than one breath. It is concluded that themechanisms accounting for the onset/termination of breathing episodes may reflect a propertyof the neuronal circuitry responsible for respiratory rhythmogenesis, and/or burst patternformation rather than an interaction between afferent inputs and the central nervous system.111TABLE OF CONTENTSABSTRACTTABLE OF CONTENTS ivLIST OF TABLES viLIST OF FIGURES viiLIST OF ABBREVIATIONS xiACKNOWLEDGEMENTS xiiiCHAPTER 1:GENERAL INTRODUCTION 1CHAPTER 2:THE ROLE OF CHEMORECEPTORS IN THEBREATHING 29Introduction 30Materials and methods 32Results 38Discussion 59CHAPTER 3:THE ROLE OFCO2-SENSITIVE OLFACTORYCONTROL OF BREATHING PATTERNIntroductionMaterials and methodsResultsDiscussionCHAPTER 4:AN OPEN-LOOP ASSESSMENT OF THE ROLE OF VAGAL AFFERENT INPUT TOCONTROL OF EPISODIC BREATHINGIntroductionMaterials and methodsResultsDiscussionCONTROL OF EPISODICAND PULMONARY RECEPTORS IN THE6667697693THE102103105114140ivCHAPTER 5:INTRINSIC BRAINSTEM GENERATION OF BREATHING PATTERN 155Introduction 156Materials and methods 158Results 166CHAPTER 5 (cont.)Discussion 179CHAPTER 6:IS THE NUCLEUS ISTHMI RESPONSIBLE FOR THE ONSET/TERMINATION OFBREATHING EPISODES 190Introduction 191Materials and methods 195Results 202Discussion 230CHAPTER 7:GENERAL CONCLUSION 236BIBLIOGRAPHY 245VLIST OF TABLESTable 2.1: The effect of unidirectional ventilation on the amplitude of arterial blood gasfluctuations in spontaneously breathing bullfrogs 53Table 2.2: The effect of unidirectional ventilation on selected ventilatory variables inspontaneously breathing bullfrogs 55Table 2.3: The effects of unidirectional ventilation of the lungs with different gasmixtures on ventilatory variables and selected bloou respiratory variables 57Table 3.1: The effects of sham surgery on the breathing pattern of bullfrogs breathingdifferent gases 89Table 3.2: Recovery profile of selected blood gas variables of bullfrogs after exposureto 4.5% Co2 for 30 mm 91Table 4.1: A comparison of the breathing pattern variables between intact anddecerebrate, paralysed bullfrogs 141Table 6.1: Individual values for the number of fictive breaths in a breathing episode thatwere recorded after NI lesions 224viLIST OF FIGTJRESChapter 1Figure 1. 1: Recordings of buccal pressure and lung pressure illustrating differentbreathing patterns observed in the adult bullfrog 6Figure 1.2: Diagram summarizing the sequence of mechanical events observed inthe breathing cycle in Rana catesbeiana 12Chapter 2Figure 2.1: Schematic representation of the experimental system utilized for continuousmeasurement of breathing and blood gases 40Figure 2.2: Continuous traces of blood respiratory variables and breathing patterns . . 42Figure 2.3: The effects of manipulating arterial Pco, on selected breathing patternvariables 44Figure 2.4: Instantaneous and absolute buccal oscillation frequency as a functionof arterial Pco2 or P02 46Figure 2.5: The effects of manipulating arterial P02 on selected breathing patternvariables 48Figure 2.6: Buccal pressure recordings of adult bullfrogs illustrating the effectsof different ventilatory stimulants on the instantaneous breathing and buccaloscillation frequencies 50Chapter 3Figure 3.1: A buccal pressure recording illustrating the breathing pattern ofintact, olfactory denervated and vagotomized bullfrogs 74Figure 3.2: The effects of changing CO2 levels on selected breathing pattern variables.77Figure 3.3: The effects of changing CO2 levels on the instantaneous breathingfrequency, and the instantaneous buccal oscillation frequency 80Figure 3.4: The relationship between breathing activity and the resulting changes inCO2 levels in the buccal cavity under normo- and hypercarbic conditions 85viiChapter 3 (cont.)Figure 3.5: A recording of the changes in breathing activity and the changes in PCO2in the buccal cavity, the gas phase of the chamber, the water phase of thechamber, and in arterial blood that occurred at the transition from hyper- tonormocarbia 87Chapter 4Figure 4.1: Recordings of integrated trigeminal nerve activity illustrating the changes inthe fictive breathing pattern associated with increasing lung pressure at differentlevels of respiratory drive 112Figure 4.2: A comparison between the effects of phasic and tonic vagal afferentfeedback on the fictive breathing pattern of bullfrogs 118Figure 4.3: A comparison of the effects of phasic and tonic vagal feedback on selectedbreathing pattern variables at different levels of hypercarbia 120Figure 4.4: A comparison of the effects of phasic and tonic changes in lung pressure ontrigeminal and vagal burst duration at different levels of hypercarbia 122Figure 4.5: The effects of tonic changes in lung pressure on selected breathing patternvariables under different levels of hypercarbia 124Figure 4.6: The effects of tonic changes in lung pressure on selected breathing patternvariables under normo- and hypoxic conditions 126Figure 4.7: The changes in ENG activity recorded from the vagus and trigeminal nervewhen lung pressure was reduced from 2 cm H20 to 0 cm H20 128Figure 4.8: The effects of tonic changes in lung pressure on vagal and trigeminal burstduration 130Figure 4.9: The effects of tonic changes in lung pressure on instantaneous frequenciesof the two types of fictive buccal movements recorded the bullfrogs 132Figure 4.10: The effects of changing lung pressure on selected breathing patternvariables in vagotomized bullfrogs, and bullfrogs with both vagi intact . . . . 136Figure 4.11: The effects of changing lung pressure on trigeminal and vagal burstduration in bullfrogs with intact vagi, and after bilateral cervical vagotomy. . 138Figure 4.12: ENG recordings that illustrate a fictive lung inflation cycle (type IIepisode) in which the motor output driving each fictive breath within an episodegets progressively larger 143viiiChapter 5Figure 5. 1: Schematic representation of the in vitro brainstem-spinal cord preparation. 160Figure 5.2: Comparison between the breathing pattern recorded from an intact bullfrogbreathing air and motor output recorded from an in vitro brainstem-spinal cordpreparation 164Figure 5.3: Simultaneous ENG recordings from the vagal and trigeminal cranial nerverootlets in the brainstem-spinal cord preparation which illustrate the effects ofadding vagal feedback on the duration or the relative timing of the bursts. . . . 167Figure 5.4: The effects of changing mock CSF/PCO2on the trigeminal and vagal burstduration with and without vagal feedback 169Figure 5.5: The effect of changing mock CSF pH/PCO2 on respiratory related activityrecorded from cranial nerve rootlet V before and during phasic stimulationof the pulmonary branch of the contralateral vagus nerve 173Figure 5.6: The effect of changing mock CSF pH/PC02 via changes in Pco2 on selectedbreathing pattern variables with and without stimulation of the pulmonary branchof the vagus nerve 175Figure 5.7: The effects of changing mock CSF pH/PCO2on the instantaneous breathingand buccal oscillation frequencies with and without vagal feedback 177Figure 5.8: The effect of tonic vagal stimulation on the fictive breathing pattern. . . 184Chapter 6Figure 6.1: Lateral view of the midbrain, medulla oblongata and part of the spinal cordof Rana temporaria 193Figure 6.2: ENG recordings illustrating the short term effects of kainic acid injectionon the vagal and trigeminal motor output 198Figure 6.3: A comparison of the effects of bilateral microinjections of saline, lidocainehydrochloride and kainic acid into the NI on the fictive breathing pattern asshown by a recording of integrated trigeminal nerve activity 204Figure 6.4: The relationship between hypercarbic respiratory drive and selectedbreathing pattern variables in bullfrogs with and without their nucleus isthmiintact 206ixChapter 6 (cont.)Figure 6.5: Simultaneous ENG recordings from the vagal and trigeminal nerves whichillustrate and contrast the bursting activity of a “fictive breath” typically recordedbefore and after bilateral KA microinjections 211Figure 6.6: The effects of changing lung pressure on the duration of the vagal andtrigeminal burst duration at two levels of hypercarbic respiratory drive inbullfrogs with and without their nucleus isthmi intact 213Figure 6.7: The relationship between tonic changes in lung pressure and selectedbreathing pattern variables in bullfrogs with and without their nucleus isthmiintact 215Figure 6.8: The effects of changing lung pressure on instantaneous buccal oscillationfrequency at two levels of hypercarbic respiratory drive in bullfrogs with andwithout their nucleus isthmi intact 217Figure 6.9: Drawing of a cross section of the bullfrog brain showing sites in the nucleusisthmi area where bilateral microinjections of kainic acid were performed 219Figure 6.10: ENG recordings of vagal and trigeminal nerve activity obtained from onebullfrog illustrating the occurrence of breathing episodes after bilateral lesions ofthe nucleus isthmi 226Figure 6.11: A comparison of the relationship between tonic changes in lung pressureand selected breathing pattern variables before and after bilateral vagotomy inbullfrogs without their nucleus isthmi 228xLIST OF ABBREVIATIONSCtg0 Arterial carbon dioxide contentCa02 Arterial oxygen contentCPG Central pattern generatorCRG Central rhythm generatorcm 1120 Centimeters of waterCNS Central nervous systemCSF Cerebrospinal fluidEMG ElectromyogramENG ElectroneurogramfBab Absolute buccal oscillation frequencyfB1 Instantaneous buccal oscillation frequencyFm Carbon dioxide fraction‘-‘--‘2fG Gill ventilation frequencyfLab Absolute lung ventilation frequencyfL Instantaneous lung ventilation frequencyF02 Oxygen fractionh HourKA Kairiic acidLSD Least significant differenceMm Carbon dioxide excretion‘—“-‘2M0 Oxygen consumptionms MillisecondNI Nucleus isthmiC02 Partial pressure of carbon dioxidePaCO2 Arterial partial pressure of carbon dioxidePwCO2 Water partial pressure of carbon dioxidepHa Arterial pHPL Lung pressureP02 Partial pressure of oxygenxiPa02 . Arterial partial pressure of oxygenPw0 Water partial pressure of oxygenPRG Pontine respiratory groupPSR Pulmonary stretch receptorsQ10 Temperature quotientRTN Retrotrapezoid nucleussec SecondsSEM Standard error of the meanTE Expiratory durationTI Inspiratory durationTT0T Total respiratory cycle durationDuration of the non-ventilatory periodUDV Unidirectional ventilationVlndex Minute ventilation indexVTlndex Tidal volume indexxiiACKNOWLEDGEMENTSFirst and foremost, I would like to thank all the members of my supervisory andexamining committee for helping me complete my Ph.D. with such speed, and for understandingthe “unusual” circumstances that I had to deal with. A special word of appreciation goes to Dr.W.K. Milsom for patiently proofreading the numerous drafts of this thesis, and for showing suchsupport. Working under his supervision for the past 5 years has been a fantastic learningexperience. Bill’s open door policy has helped create a friendly relationship, and a stimulatingwork environment. He has taught me many valuable lessons, including the importance of thelittle details that improve the quality of one’s work, and to beware of a good poke-check, as itwill catch the puck carrier by surprise 9 times out of ten! This lesson forced me to develop mystick handling skills over the years. I would also like to thank the physiologists from theBiology department at the University of Ottawa who got me interested in physiology in the firstplace, and more specifically Dr. S.F. Perry, whose enthusiasm and encouragement made itpossible to successfully continue at the Ph.D. level.I also acknowledge the help and support of all my fellow lab mates from the Milsom labwho had to cope with me over the years; however, a special word of appreciation must bementioned for Mike Harris. He truly made a difference in my life by introducing me tonumerous cults: the Data Gods and the art of their worship to succeed in any scientificinvestigation; the Grateful Dead as the key to living on West coast and the Joy of Brewing toblend all of the above and reach a total Nirvana! Thanks Mike, I’m not sure I would havesurvived to it all otherwise.I would like to thank Nick Bernier, Tim West, Les Buck, Jim Staples, Barb Taylor,Joanne Lessard, Cohn Brauner, Deb Donovan and Tom Carefoot who have enriched my life andhelped me maintain my spirits and just for taking the time to share a pitcher of beer, a cup ofxiiicoffee or to play a set or two. Last but not least (despite her height) I thank Joëlle Belanger(a. k. a. Harris) for putting up with all the abuse, practical jokes and lending her ear. A specialword of acknowledgement also for Gillian Dawn Muir (Super Genius) who shared office spacewith me, and yet kept smiling for the past four years.I also thank my parents (especially Mom, who asked me to do this) for theirencouragement and support, even though they did not always understand why, at 30, I still didnot have a job! I also thank my sons Frédéric, Simon and Henri (#16; the Pocket Rocket), whohelped me keep things in perspective, and my wife Chantal to whom I dedicate this thesis. Herlove, support, kindness and understanding through it all, made it possible for me to achieve thisgoal. Mule mercis!Finally, I would like to acknowledge the financial support of NSERC, FCAR and theKillam Trust which have helped me stay focused on my research.xivCHAPTER 1:GENERAL INTRODUCTIONPage 11. Structure-function relationship of gas exchange in amphibians.The emergence of life on land was a pivotal step in the radiation of vertebrates. It isbelieved that the aquatic ancestor of air-breathing vertebrates lived in waters where importantdaily fluctuations in CO2 levels, temperature and oxygen availability occurred. Most likely,hypoxia was the external condition that compelled fish to resort to “air gulping” to supplementbranchial oxygen uptake (M02) in order to satisfy their metabolic needs (Randall et al. 1981).The development of an air receptacle that efficiently exploited the new “02 rich” milieu wasundoubtedly an adaptive asset that facilitated survival in such an environment.During the devoman, crossopterygian fish gave rise to labyrinthodonts, the primitiveamphibians (Norris, 1985). Members of the class Amphibia are characterized by a life cycle thatbegins in water, where they use their gills and skin as gas exchange structures, and concludeswith an adult form that can potentially live on land, and usually breathes using lungs. Thechanges that occur during the development of amphibians with respect to the control of breathingare similar to the changes that are believed to have occurred during the evolutionary transitionfrom water to land (Burggren, 1991).In amphibians, terrestrial life is limited, at least in part, by the dependence on the aquaticmilieu where vital functions, such as reproduction, must be performed, and the fact that the skinof frogs and salamanders cannot sustain desiccation for a prolonged period of time. Hence,these animals must remain in a moist environment. Aquatic urodeles, such as newts andsalamanders, show a wide range of aquatic dependence. The degree of lung development, andtheir ability to perform aerial gas exchange, varies from the lungless plethodontid salamanders,through simple saclike structures in Cryptobranchus and Necturus, to the well-developedunicameral lungs of Amphiuma and Xenopus (For review, see Burggren, 1989). Cryptobranchusrely chiefly on skin and/or external gills to perform gas exchange adequately, even though theirPage 2lungs are functional. In adult toads, on the other hand, the skin is rather impermeable and thusa poor gas exchanger. Consequently, toads rely mostly on their lungs for breathing. Certainspecies of adult frogs and salamanders are situated half-way in the continuum of aquaticdependence. Oxygen uptake in these animals is primarily performed by the lungs since theskin’s contribution, alone, is insufficient (30% at rest at 20-25°C: Gottlieb and Jackson, 1976;Burggren and West, 1982; Burggren, 1984). In salamanders, CO2 elimination typically occursprimarily via the skin, whether they are in water or in air. All animals that utilize both air andwater for gas exchange are referred to as bimodal breathers.The physical and chemical properties of water and air are considerably different withrespect to CO2 and 02 capacitance and diffusion (Dejours, 1988). Hence, the convectionrequirements and the concomitant energy output necessary to fulfil M02 and CO2 excretion(MCO2)will vary significantly depending on whether air or water is breathed. For bimodalbreathers to survive in environments where hypoxic, hypercapnic or even hyperoxic conditionscommonly occur, one would logically predict that natural selection would favor the developmentof control mechanisms that efficiently coordinate gas exchange at different sites. The apparentcomplexity and finesse of this control mechanism has aroused the interest of respiratoryphysiologists for many years, because understanding respiratory control in “primitive” airbreathers could yield valuable insight into the basic principles underlying respiratory control inall vertebrates.With these thoughts in mind, the present chapter will review the literature and summarizeour current knowledge on the control of breathing in vertebrates with a strong focus onamphibians.Page 32. “Resting” breathing pattern in amphibians.2.1. Ontogeny of breathing pattern.In anurans, breathing goes through important developmental changes. Initially, theaquatic breathing of the anuran larva (tadpole), is quite similar to that of water breathing fishes.It consists of a unidirectional flow of water over the internal gills activated by a buccal pumpwhich functions in a continuous and rhythmic fashion (Burggren, 1984). The resting frequencyof buccal pumping for gill ventilation (fG) as well as air breathing frequency (fL) are highlyrelated to developmental stage (Burggren and Doyle, 1986). Immediately after hatching (StageI)’, when the gills are still external extensions of the body, f0 is very high(P—110 cyclesmin’).When the gills become enclosed internally within paired branchial chambers (Stage IV-VII), f0falls sharply to about 50 cyclesmin’. This level of G characterizes older larval stages andpersists as buccal oscillations even in the adult bullfrog (Burggren and Doyle, 1986). Airbreathing under resting normoxic conditions is a rare occurrence up to Stage X, even thoughlungs have developed by at least Stage III. Even in Stages XVI-XXIV, an air breath occurs onlyabout once every 15-20 mm at 20°C. After metamorphosis to the adult, L increases greatly,ranging from 1-3 breathsmin’ (Burggren and Doyle, 1986). Burggren (1991) reports thatbullfrog larvae start to display an “adult” breathing pattern at stages XX-XXIV. This means thatthe larval pattern of infrequent single lung ventilations becomes a pattern more frequent breathsgrouped into episodes.The developmental stages are described according to the method of Taylor and Kollros(1946).Page 42.2. “Resting” breathing pattern in adult amphibians.Figure 1.1. illustrates the normal breathing patterns of the bullfrog. When at rest, thepattern can consist of single breaths separated by a non-ventilatory period of variable duration,which can be influenced by external variables such as the presence of predators (Maclntyre,1975) or a mate (Halliday and Sweatman, 1976). The behavioral component of the apnea, orthe non-ventilatory period, accounts for the inherently high degree of variability of thiscomponent of the breathing pattern. This variability is consistently observed in spite of attemptsto minimize external stimuli during experiments. Sometimes, the single breaths are clustered toproduce breathing episodes, also separated by periods of apnea. At increased levels ofrespiratory drive, the buccal pump is activated in a “ramp-like” fashion such that the lungvolume following each breath within an episode gets progressively larger. This type ofbreathing episode has been referred to as a lung inflation cycle (Maclntyre and Toews, 1976;Vitalis and Shelton, 1990; West and Jones, 1975) (see following section). A notable feature ofmost semi-aquatic and terrestrial anurans is the use of their “buccal pump” to ventilate the buccalcavity between episodes of aerial breathing, even in their adult stage. These buccal ventilationsor “oscillations”, which consist of a simple tidal pattern of flow at low rates through the nostrils,are absent in the aquatic anuran Xenopus laevis (Brett and Shelton, 1979). Their purpose is notknown, but they are believed to help flush the buccal cavity of the residual gas from the previousexpiration, or help in olfaction.Page 5Figure 1. 1: Recordings of buccal pressure and lung pressure illustrating different breathingpatterns observed in the adult bullfrog.Page 6Air AirU=-ti:c It ILl I_iL iBreaths Breathing episodesAir 4% CO2I I [‘ : C________LA 1Lung thflation cycles 30 secPage 73. Breathing mechanics.“The physiologists of the end of the 17th and of the 18th centuries did notfail to observe that there was an essential difference in the mode of respiration ofthe frog and of the mammals... The first detailed account of the mode ofrespiration of the frog was however given by Townson in 1794: his account farexcels in minuteness of description those of any of his predecessors, and issubstantially correct in all points... He pointed out that all the throat movementswere not alike, but that some of them differed in character from the rest, andwere alone accompanied with closure of the external nares and contraction of theflank muscles. He also described and figured the muscles of the throat, and firstshewed that if the mouth of a frog was kept open it could not send air into itslungs...”H. Newell Martin, 1878.Lung ventilation in air breathing fishes and amphibians is, in most cases, performed bya pulse pump. This convective process arose in primitive air-breathers, and is believed to bederived from the normal teleostean respiratory movements, or the cough, which were employedto fill with air various extensively vascularized cavities, such as the mouth, gut or blind sacsfrom the gut walls which function as primordial lungs (Ballintijn, 1987). The basic principleof the pulse pump is that air is first taken into the buccal cavity and subsequently transferred intothe lung (or primitive gas exchange organ) by a rapid and forceful elevation of the floor of thebuccal cavity with the mouth and nares closed.In pulse pumping, the volume of gas that is moved during a single pulse is a function ofthe pressure gradient between the lungs or the air sac and the buccal cavity, and the volumetricchanges that the buccal cavity can accommodate. This places a limitation on the ratio of tidalvolume to lung volume (Gans, 1970). When convective requirements become elevated, frogscircumvent the limitations imposed on tidal volume by the small volume of their buccal cavityby engaging in lung inflation cycles during which the pressure and volume of the lung increaseby increments.Page 8As illustrated in Fig. 1.2, a breathing cycle begins by activation of the buccal depressormuscles which decreases buccal pressure below ambient pressure, and air is aspired into thebuccal cavity via the nostrils. The laryngeal dilator muscles then contract to open the glottis;this allows outflow of pulmonary gas which is retained at pressures above ambient pressure, andwhich exits by the nostrils. Because of the airflow resistance imposed by the nostrils, thisprocess will give rise to an increase in buccal pressure. It should be mentioned that there hasbeen much debate over the air flow pattern during this particular step of the breathing cycle,since it would affect the degree of admixing of inspired gas with the gas expelled from thelungs. Gans (1969) proposed the “jet-stream” hypothesis, which stated that the inspired freshair is confined to the ventral and posterior region of the buccal cavity, and that the gas expiredfrom the lungs flows in a uniform stream in the dorsal portion of the mouth. This air flowpattern would help minimize admixing of the inspired gas, and thus, improve the efficiency ofthe pulse pump. This hypothesis has recently been challenged by the work of Vitalis and Shelton(1 990). Based on simultaneous measurements of pulmonary pressure, buccal pressure and airflow at the nostrils in Rana plpiens, these workers reported that during inflation, the nostrilsclosed simultaneously with glottal opening and almost no gas was expired during this phase.This caused a complete mixing of buccal contents and pulmonary gas and this mixture waspumped back into the lungs. From these results, they concluded that coherent air flow fromglottis to nostrils, as required by the “jet stream” hypothesis, was not likely to occur.When the lungs deflate, pulmonary pressure usually remains above ambient since lunginflation begins before deflation is complete. The lung inflation begins with closure of thenostrils (a process which may or may not be active) and a brisk contraction of the buccallevators which pushes the gas in the buccal cavity, through the glottis and into the lungs. Thisprocess causes a sharp simultaneous rise in buccal and lung pressures which peak when thePage 9glottis closes to “trap” the bolus of gas inside the lungs (there is contention as to whether or notglottal closing is an active process). Lung pressure remains elevated, and buccal pressuredecreases rapidly to ambient when the nostrils reopen. What follows typically is a series of thetidal elevations and depressions of the floor of the buccal cavity with the nares opened and theglottis closed, which are called buccal oscillations. These small-amplitude, low pressure buccalmovements, which are likely to be remnants of gill ventilation from the pre-metamorphic tadpolestages (Taylor, 1989), help flush the buccal cavity from the previous expiration, and minimizeadmixing of gas during the next air breath (De Jongh and Gans, 1969; De Marneffe-Foulon,1962; Gans, 1970; Vitalis and Shelton, 1990; West and Jones, 1975).During lung inflation cycles, the lungs are inflated by a series of lung ventilationmovements of increasing peak pressure. This type of breathing episode usually occurs when theanimals are disturbed, stressed, or when respiratory drive is high. To produce these pulmonarypressure crescendos, the outflow of gas from the lungs is minimized by shortening the timeavailable for deflation; the nares close and the floor of the mouth is elevated soon after theopening of the glottis. It has been reported that, in some instances, buccal levators are activatedbefore the glottis is opened (West and Jones, 1975). This increases buccal pressure rapidly, sothat when the glottis opens, pulmonary and buccal pressure equalize rapidly and reverse thedirection of airflow at the glottis immediately, minimizing outflow of pulmonary gas. Thus, theentire tidal volume is added to the gas already present in the lungs, inflating the lungs to a largervolume. Several such events may occur in succession. Typically, this sequence is followed bya period of apnea of variable duration. The following breath or two deflate the lungs to a lowerpressure before the cycle begins again.The breathing mechanics of the pulse pump make it difficult to define tidal volume. Inmammals, expiration follows inspiration, and the volume expired is usually roughly equal to thePage 10inspired volume. In bullfrogs, however, as we’ve just seen, this may not the case, since thevolume of gas pumped into the lungs may be more or less than the volume expired from thelungs, and both these volumes may differ from the volumes inspired or expired through theflares. Furthermore, the definition of tidal volume can be further complicated depending onwhether one is considering a single breath or a lung inflation cycle. In a lung inflation cycle,the tidal volume could be considered to be the total amount of gas inspired in the bout ofventilation; i.e. the sum of the volumes pushed into the lungs with each buccal movement.Alternately, each buccal movement involved in lung ventilation could be considered as a breath.Most authors use the latter approach.Unlike air-breathing fish, which must open their mouth to aspirate ambient air into theirbuccal cavity at the onset of the breathing cycle, frogs aspirate air via their nostrils. Eventhough this modification imposes a slight resistance to gas flow, it eliminates the energyexpenditure associated with “biting at the air” seen in air-breathing fish (Gans, 1970). In termsof oxidative cost, the pulse pump of frogs is one of the least expensive strategies seen amongstvertebrates; the oxidative cost per liter of gas ventilated in frogs is quite low. However, whenthe differences in metabolic rate and minute ventilation are taken into account, the relative costof ventilation as a fraction of the total 02 consumption of the animal is an order of magnitudegreater in the frog than in mammals (Milsom, 1989)Page 11Figure 1 .2: Diagram summarizing the sequence of mechanical events observed in the breathingcycle in Rana catesbeiana (adapted from De Jongh and Gans, 1969). The time markerin the figure was estimated from the EMG traces reported in that paper.Page 12;Y-)LLNG8JCCALVenhiotion PPoesNOSTRIL MOVEMENTcCLOSE(4kifl1)GLOTTISNOSTRILS_______FLOW AT GLOTTISIN,FLOW AT NTRILS 0UtA1ICLOSED’ 01P E 4OLITf1 secPage 134. The relationship between gill (or buccal) and lung ventilation.In tadpoles (Stages XVII-XIX) and Dipnoi lungfish, the rate and amplitude ofbranchial/buccal pumping is low immediately after an air breath. They then gradually increaseuntil the next air breath is taken (Johansen and Lenfant, 1968; West and Burggren, 1983). Westand Burggren (1983) demonstrated that the decrease in rate and amplitude of buccal oscillationfollowing an air breath was not only due to a decrease of the02-related respiratory drive, butalso to a direct effect of lung inflation mediated by pulmonary mechanoreceptors (West andBurggren, 1983). These authors (West and Burggren, 1983; Burggren, 1984) suggested that this“reflex” prevented loss of 02 at the gills after an air breath; such a reflex would be extremelyadvantageous under hypoxic conditions. Pack et al. (1990) did not believe, however, that theinterpretation of West and Burggren (1983) suited the results they obtained in African lungfish.Instead, they claimed that the decrease in gill ventilation following an air breath was centrallymediated, reflecting the existence of two distinct rhythm generators related to gill and pulmonaryventilation respectively (Pack et at. 1990; Fishman et at. 1989). Their argument was based onobservations that 1) slowing of gill ventilation occurred even when lung inflation with eachattempted lung breath was experimentally prevented and, 2) unlike tadpoles, the inflation oflungs between breaths did not reduce the frequency of gill movement. Furthermore, theyclaimed that since hyperoxia caused a complete cessation of pulmonary ventilation (Jesse et at.,1967), the rhythm generator for pulmonary ventilation was inoperative unless it received afferentstimuli related to hypoxia (Fishman et at. 1989).Fishman and co-workers (1989) were not the first, however, to postulate the existenceof two separate loci for the control of buccal and pulmonary ventilation. Indeed, De MarneffeFoulon (1962) claimed that Rana sp. had two distinct respiratory centres which, under normalPage 14conditions, perfectly coordinated buccal and pulmonary activity to produce efficient breathing.Under certain experimental conditions, the action of these centres could be dissociated. Forinstance, 1) intravenous injection of 0.5 ml of alkaline Ringer (pH = 8.1) provoked a dosedependant increase in buccal oscillation frequency and a concomitant decrease in lung ventilationfrequency. Conversely, intravenous injection of 0.5 ml of acidic Ringer (pH = 5.6) resultedin a significant decrease in the number of buccal oscillations per minute, while lung ventilationfrequency was augmented. Similar observations were reported after injection of acidic oralkaline saline into the CSF, and after pouring the same solutions directly onto the fourthventricle. 2) De Marneffe-Foulon also noticed that although the frequency of buccal oscillationand lung ventilation are both affected by temperature variations, the augmentation (or decrease)of buccal oscillation frequency was not parallel to that of lung ventilation, suggesting a differentQ10 for each variable.Recent investigations on the relationship between gill and lung ventilation performed onreduced tadpole preparations have produced new data relevant to this topic. Addition of -yaminobutyric acid (GABA) to the superfusate (5 mM) of a brainstem-spinal cord preparation ofthe tadpole eliminated fictive gill movements without having much effect on fictive lung breaths(Walker et at., 1990). In the same preparation, removal of chloride ions from the superfusateabolished the bursts corresponding to gill ventilation while the bursts for lung ventilationpersisted (Galante et al., 1992). Furthermore, blockade of small conductance Ca2 activated Kchannels with apamine increased the amplitude of gill-related activity, and prolonged theduration of the lung bursts (Liao et a!., 1994). The sum of these anecdotal results obtained fromthe tadpole preparation shows that both types of breathing events can be modulated differentlyin tadpoles, and thus supports the hypothesis that both types of breathing events share the samemuscles, but are driven by different mechanisms (Pack et at., 1992).Page 155. Control of breathing pattern.“My own observations were on summer frogs... they exhibit verymarkedly a periodic respiration of the Cheyne-Stokes type... It seems well toclass together under the general title periodic respiration all forms of grouprespiration and to reserve the term Cheyne-Stokes phenomenon specially for thatform distinguished by the gradual character of rise and fall of activity, thatCheyne and Stokes were the earliest to comment upon.”Sherrington, 1891.Cheyne-Stokes breathing is a pathological condition in which repeated cycles of irregularbreathing begin with shallow breaths that increase in depth and rapidity, then decrease and ceasealtogether for 15 to 20 seconds (Tortora, 1992). The distinction between Cheyne-Stokesbreathing and periodic breathing was made over 100 years ago; yet the term “Cheyne-Stokesbreathing” has often been used in recent literature to described the breathing pattern of lowervertebrates or mammals experiencing natural depression of metabolism. Although this issomewhat understandable, owing to similarities in the periodic nature of the two patterns, it isa serious misnomer because episodic breathing is a physiological phenomenon, rather than apathological condition.To this day, the mechanisms underlying this intriguing breathing pattern are still poorlyunderstood. Although our knowledge on this topic is still somewhat limited, there are noreasons to believe that the fundamental components of the respiratory control system of episodicbreathers are drastically different from those of animals that breathe continuously. Most of thecurrent information on the topic arises from studies of respiratory reflexes, observing theventilatory responses triggered by exposing animals to a variety of ventilatory stimuli.Therefore, I reviewed that literature briefly before describing what is known about the differentcomponents of the respiratory control system of amphibians.Page 165.1. Ventilatorv responses to hypoxia.An interesting characteristic of the hypoxic ventilatory response of amphibians is itssignificant modification during development. Indeed, the increase in body mass and overallmetabolic demand during development requires the progressive development of more effectiverespiratory structures (gills, lungs) in these developing larvae. These gas exchange structurescontinue their development while the larva is entirely exposed to an external milieu that ispotentially limiting to respiration (Burggren, 1984).Respiratory reflexes to hypoxia are functional during larval development in ranid frogsand salamanders (Feder, 1983; West and Burggren, 1983; Burggren and Pinder, 1991). The Gresponse to hypoxia and hyperoxia varies greatly as a function of development (Burggren andDoyle, 1986). Researchers have reported 3 different types of response. First, Stage I larvaedisplayed an increase in buccal pumping rates in response to decreases in 02 availability (Pw02ranging from 60 to 90 Torr) within one day of hatching, although at this stage, the external gillswere not actively ventilated by buccal movements. More severe hypoxia (Pw0 30 Tort),however, depressed G below resting levels, while hyperoxia (Pw02 > 500 Tort) caused analmost total suppression of gill ventilation. Second, in older larvae (Stage IV-XIV), G increasedprogressively as Pw02 declined; hyperoxia was reported to have no significant effect in theseintermediate larval groups. Third, in larvae of Stage XVI and older, G was no longer affectedby ambient P02.On the other hand, there appear to be only two types of lung frequency responses tohypoxia. In Stages VI or younger, L is completely insensitive to depression of environmentalP02. Hypoxic exposure in Stage IX and older, however, consistently stimulates an increase inL (Burggren and Doyle, 1986). Thus, the effector component of the respiratory reflexesPage 17modulating ventilatory frequency in response to aquatic hypoxia was transferred from the gillsto the lungs with relatively little overlap.It is known that the number of lung ventilations per unit time in Bufo marinus, Bufoparacnemis and Rana catesbeiana is mediated mostly by a decrease in the duration of breath-holding intervals (Boutilier and Toews, 1977; Kruhøffer et at. 1987; West and Burggren, 1982;Jones and Chu, 1988). In addition, it was demonstrated recently that hypoxia (Fl02 rangingbetween 5 and 40%) increased the number of breaths per breathing bout, but had relatively littleeffect on the duration of the non-ventilatory period (TNVP) in lightly anaesthetized unidirectionallyventilated (UDV) Bufo marinus during intracranial perfusion with hypocapnic alkaline mock CSF(Smatresk and Smits, 1991). There are few studies that have reported tidal volume values underhypoxic conditions, but the data indicate that, in Xenopus laevis, hypoxia does cause tidalvolume to increase (Jones and Chu, 1988). Based on the effects of manipulating arterial P02and 02 content (Ca02) independently on the ventilatory response to hypoxia, it would appearthat the reduction in Pa02, rather than 02 carrying capacity, is the stimulus driving this reflexin amphibians.5.2. Ventilatory responses to hypercapma.A wide variety of semi-terrestrial anurans hyperventilate in response to hypercapnia (Bufomarinus: Branco et at., 1992; Maclntyre and Toews, 1976; Boutilier et at., 1979; Smatresk andSmits, 1991; Typhlonectes compressicauda: Toews and Maclntyre, 1978; Rana catesbeiana: DeMarneffe-Foulon, 1962; Sakakibara, 1978; Jackson and Braun, 1979; Infantino; 1989; Xenopuslaevis: Shelton and Boutilier, 1982; Ci’yptobranchus atleganiensis: Boutilier and Toews, 1981).The emergence of this “reflex” is part of the development of the animal since bullfrog larvaePage 18(stages I-X) are almost insensitive to changes in until a late-stage in development (XVIXIX) (Infantino, 1989).In Cryptobranchus, the hyperventilatory response to hypercapnia is mediated by anincrease in lung ventilation and rocking motions, which are though to facilitate convection at thewater-skin interface, as well as aquatic buccal pumping (Boutilier and Toews, 1981). Inunidirectionally ventilated anaesthetized Bufo marinus, the hyperventilatory response tohypercapma was mediated by an increase in the number of breaths per breathing bout (Smatreskand Smits, 1991). In addition to these temporal modifications of breathing pattern, hypercapniaalso affected the way in which gas exchange was performed. For instance, the C02-inducedpulmonary hyperventilation in Bufo marinus was achieved by high pressure inflation-deflationcycles (Maclntyre and Toews, 1976; Boutilier, 1988), which are believed to facilitate lungperfusion and thus, gas transfer (Maclntyre and Toews, 1976). Given that large swings inbuccal pressure were required to produce this type of breathing episode, it was not surprisingthat hypercapnia also caused an increase in tidal volume (Branco et al., 1992).Although peripheral 02 chemoreceptors have been shown to increase their discharge ratewhen they are facing an increase in C02/H, the findings of recent studies suggest that theircontribution to the hypercapnic ventilatory response is secondary to the one produced by centralchemoreceptors. It has been estimated that these latter receptors contribute about 80% of thetotal hypercapnic respiratory drive in Bufo paracnemis (Branco et al., 1992).Page 195.3 Feedback systems5.3.1 Oxygen receptorsNumerous studies have attempted to elucidate the location as well as the role ofperipheral chemoreceptors in the hyperventilatory response to hypoxia in amphibians. Althoughchemoreceptive elements have been identified, the precise role of specific peripheral 02chemoreceptor groups in the regulation of breathing is still poorly understood (Boutilier, 1988).The carotid labyrinth has long been a prime target in the study of 02 chemodetection inamphibians. These receptors are situated at the bifurcation of the internal and external carotidarteries, and are innervated by a branch of the glossopharyngeal nerve which projects its afferentfibres to the solitary tract in the brainstem (Stuess et al., 1984). These receptors arefunctionally similar to the mammalian carotid bodies, as they also respond to hypercapnia andtheir discharge can be modulated by sympathetic stimulation (Ishii and Ishii, 1970; Ishii et at.,1985; van Vliet and West, 1992). In addition, recent studies have shown that the receptors aresensitive to the partial pressure, not the content, of 02 (West and van Vliet, 1992); a findingconsistent with the results of whole animal study of the stimulus modality of the hypoxicventilatory response in toads (Wang et al. ,1994).Carotid labyrinth denervation caused a significant reduction of resting ventilatory activityin comparison to sham-denervated animals, but it had no significant effect on the ventilatoryresponse to hypoxia in Xenopus laevis or Bufo marinus (Evans and Shelton, 1984; West et at.1987; Jones and Chu, 1988). These findings indicate that the carotid labyrinth influencesrespiratory drive but is not essential for the control of ventilation during hypoxia in anurans.Given that hypoxia is a poor ventilatory stimulant in frogs, it is not surprising that the currentconsensus of recent studies performed on whole conscious animals tends to minimize theimportance initially accorded to the carotid labyrinth in the regulation of ventilation.Page 20The existence of other02-chemosensitive structures was subsequently established in theaortic trunk of the toads Bufo vulgaris and Bufo marinus (Ishii et al. 1985; van Vliet and West,1986; 1987). Furthermore, it appears that the pulmocutaneous arteries of anurans may also bethe site of an intra-arterial chemoreceptive zone. It was observed that injection of NaCN intothe pulmocutaneous arches of anaesthetized Rana catesbeiana and conscious Bufo marinusstimulated ventilation (Lillo, 1980; van Vliet and West, 1986). In the absence of recordingsfrom pulmocutanous chemoreceptors, however, the role of the pulmocutaneous artery as achemoreceptive site in amphibians is conjecture (West and van Vliet, 1992). The relativecontribution of the latter two chemoreceptive sites to the hypoxic ventilatory response inamphibians is yet to be investigated.5.3.2 CO/H receptorsLike mammals, C02/H+ levels of anuran amphibians are monitored in the blood and thecerebrospinal fluid (CSF). As it was mentioned in the previous section, increases in arteriallevels of CO2/H will increase discharge rate from the carotid labyrinth. Smatresk and Smits(1991) have shown that, in anesthetized toads, stimulation of these receptors alone can elicit ahypercapnic ventilatory response. These workers have also demonstrated that, as for mostvertebrates (other than fish) investigated thus far, toads also have central chemoreceptors locatedon the ventral surface of the medulla which respond to acidic/hypercapnic challenges. Theseresults have since been repeated in unanesthetized animals (Branco et aL, 1992), and confirmthe early findings of De Marneffe-Foulon (1962) who also reported an increase in pulmonaryventilation when the brain of Rana was bathed in acidic saline.Page 21Although pure chemoreceptors sensitive to CO2 have not been identified in the lungs offrogs, the discharge of stretch receptors is sufficiently modulated by CO2 to have noticeableeffects on the breathing pattern. This topic is discussed in the following section.5.3.3 Pulmonary stretch receptorsPulmonary stretch receptors, also known as mechanoreceptors, constitute anotherimportant feedback contributing to the control of breathing in amphibians. There are threedifferent types of pulmonary mechanoreceptors in amphibians which respond to 1) the degreeof lung inflation, 2) the rate at which lung volume changes or 3) both stimuli (Milsom andJones, 1977; Jones and Milsom, 1982). The mechanoreceptors of amphibians report to the CNSvia afferent fibres in the pulmonary vagii (Milsom and Jones, 1977; Jones and Milsom, 1982),which project to the solitary tract in the brainstem (Stuess et at., 1984). The receptors aremostly slowly adapting and their firing rates decrease when the intrapulmonary CO2concentration is increased (Milsom and Jones, 1977; Kuhlmann and Fedde, 1979).Pulmonary afferent fibres play a key role in the termination of lung inflation in the adultand inhibition of buccal oscillation in the pre-metamorphic tadpoles (see section 2.1). Theevidence is that pulmonary deafferentation by vagotomy in Xenopus results in an increase in thenumber of inspirations in a ventilatory period, and overinflation of the lungs (Evans and Shelton,1984). Consequently, reduction of lung volume and the concomitant depressant effect ofprogressive elevation of alveolar CO2 on receptor firing, are important stimuli to thetermination of a dive (Toews, 1971; Boutilier, 1989).Page 225.4. Central components.5.4.1. Central rhythm generators.For many years, physiologists have been intrigued by the neural mechanisms thatgenerate the continuous alternation of inspiration and expiration. In a series of papers on theregulation of respiration, Lumsden (1923) reported on the results of progressive brainstemtransections in cats, he concluded that“The theory generally held is that the automatic respiratory centre is in thelower part of the bulb, and that the regulation of impulses sent out by it, isdetermined by impulses passing to it.”Since then, his original findings have been confirmed by similar brainstem transectionexperiments in fish (Shelton, 1959; 1961) and frogs (Oka, 1958a,b), which have helped establishthat the basic rhythm generator for breathing is situated in the caudal portion of the brainstem.More recently, the site of rhythmogenesis has been localized in mammals to the pre-Botzingercomplex (Smith et al., 1991b) which is situated in the reticular formation of the medulla at thelevel of the hypoglossal nuclei. The mechanisms underlying respiratory rhythmogenesis are stillunresolved in any species. Recordings from isolated brainstem-spinal cord preparations inlamprey (Rovainen, 1985), bullfrog (McLean, 1992) and turtle (Douse and Mitchell, 1990) haveshown rhythmic respiratory related discharges in spinal and cranial motoneurones. Becausethese preparations can produce a respiratory rhythm in the absence of afferent feedback (withthe possible exception of input from central CO2 chemoreceptors, when present) it would appearthat a CRG is also present in these vertebrates. At the same time, since it is possible to reducebreathing frequency to virtually zero by artificially meeting the convective requirements of ananimal (e.g. external membrane lung, unidirectional gill or lung ventilation; for review, seePage 23Milsom, 1990a), it would appear that in order to function, the CRG requires some externalstimulus to trigger respiratory events.Two models have been proposed to try to explain respiratory rhythmogenesis inmammals. One proposes that the central rhythm generator (CRG) consists of burster orpacemaker neurones, which show spontaneous rhythmic oscillations in membrane potential inthe absence of synaptic inputs or alternatively require a tonic excitatory input before they exhibitrhythmic oscillatory activity. The second one postulates that respiratory rhythm is produced byneural networks that exhibit oscillatory behaviour due to synaptic interactions.Bianchi and co-workers (1995) have recently reviewed the literature on the central controlof breathing in mammals. Because breathing results from the sequential activation of manypopulations of neurones to produce a three-phase motor act (see below) in which each processis conditioned by the previous one and initiates the next, they have suggested that the networkmodel is perhaps a better model to describe how respiratory rhythm is produced. An alternateview would be that the coordination of the groups of respiratory neurones would be performedby an entity separate from the CRG. This entity would be responsible for processing therelevant sensory signals, and ensuring precise spatial and temporal patterns of muscle activationduring each breath so that the respiratory system meets the demands of the organism. It is tofacilitate an understanding of the relationship between respiratory rhythm and pattern that theconcept of a central pattern generator (CPG) has emerged (Feldman et at., 1990; Milsom, 1991).Page 245.4.2 Central pattern generatorsIn vertebrates that display rhythmic breathing, pattern is defined solely in terms of thetime spent in inspiration (T1) and expiration (TE) and the rate of air flow. Combinations of thesevariables produce the more familiar components of breathing, namely, tidal volume (flow/T1),breathing frequency (60/Tl+TE) and minute ventilation. From neurophysiological data, themammalian ventilatory cycle has been divided into three distinct phases in which each phasereflects a “state” of the oscillating network, rather than a particular configuration of the motoroutput. In other words, a cycle phase in this context means a recurring episode when one ormore group of neurones in the network discharge a characteristic pattern of action potentials(Richter, 1982; Richter et al., 1986). These phases have been defined as inspiration, postinspiration and expiration. Note that the post-inspiratory phase is a period of inspiratorybraking, which is also referred to as the first stage of expiration (Kogo and Remmers, 1994).Pattern is more complex in arrhythmic breathers where the components of breathingfrequency also include an apneic or non-ventilatory period of a variable duration (T,) and thenumber of breaths per episode. Kogo and Remmers (1994) have recently discussed thesimilarity of the respiratory phases between amphibians and mammals. Their intra- andextracellular recordings of respiratory neurones in bullfrogs, provide solid evidence to argue thatlower vertebrates also have a three phase respiratory cycle. According to their analysis, the firstphase is expiration, and it occurs when the glottis is first opened. This is then followed byinspiration which is produced by the brisk activation of the buccal levators to push air back intothe lungs. The last phase is a period of breath holding, during which neurones other than thoseinvolved in the production of the two other phases were shown to be active. This phasecorresponds to the post-inspiratory phase described previously for mammals. They concludePage 25their discussion by stating that this analysis is consistent with that of Pack and co-workers whosuggested that lungfish, which also use a buccal force pump, have a post-inspiratory phase.In mammals, the pons co-operates with the pulmonary vagus in switching betweenrespiratory phases and setting the volume threshold for inspiration termination (St-John, 1977).Bilateral vagotomy results in an augmentation of T1, which results in larger tidal volumes andslower breathing. In anaesthetized or decerebrate cats, pontine lesions alone have little effecton the respiratory pattern as long as the phasic vagal input is present. Subsequent vagotomy,however, produces a grossly altered inspiratory pattern. This abnormal breathing pattern,termed apneusis, consists of long bursts of tonic inspiratory activity. These findings suggest aconvergent role of the pons and vagal inputs in the production of normal phase transition, andtheir interaction appears to be an important component of pattern generation (Feldman, 1986;Milsom, 1991).Experiments performed on reptiles demonstrate that mild anaesthesia and brainstemsection at the level of the rostral rhombencephalon (metencephalon) abolish breathing episodes,i.e. these animals now breath in an uninterrupted fashion (Naifeh et a!. 1971a, b). Vagotomyalso affects the breathing pattern by reducing the number of breaths per episode in crocodilians(Naifeh et al. 1971a, b). It is interesting to note, however, that vagotomy has no effect on thebreathing pattern if its performed after episodic breathing has been abolished by a caudalmidbrain transection (Naifeh et al. 197 la).Fish possess a group of neurones with phase switching properties situated in the dorsaltegmentum at the level of the caudal midbrain. This group of respiratory rhythmic neurones(termed type-A neurones) appears to play a key role in the control of episodic breathing.Indeed, type-A neurones fire just before the onset of a breathing bout during intermittentPage 26respiration. Furthermore, stimulation of this area during a ventilatory pause brings forward theonset of the next breathing bout (JUch and Ballintijn, 1983).Taken together, these data suggest that, functionally, a group of cells in the caudalmidbrain act together with vagal afferent input to pattern the output of a CRG and produce thebreathing pattern. It should be mentioned, however, that since central pattern generators aresimply functional concepts for which there are no anatomical correlates, it is possible that thesame cells that compose the CRG are involved in pattern formation.6. RationaleThe thesis was designed to test the hypothesis:“EPISODIC BREATHING IS PRODUCED BY CENTRAL INTEGRATION OF AFFERENT INPUTS.”The relationship between gas exchange and breathing has been the focus of many studieson respiratory control in intermittent breathers, and it has been suggested that the depletion of02 stores and/or accumulation of C02/H+ are directly responsible for the onset of breathingepisodes, and restoration of 02 stores and/or elimination of C02/H are responsible for thetermination of breathing episodes (Shelton and Croghan, 1988). In this scenario, the animal issimply breathing “on demand”, and the contribution of central control is minimal, whereas theinput from arterial chemoreceptors is pivotal. To this end, chapter 2 examines the role ofarterial chemoreceptors in the production of breathing episodes by contrasting the breathingpattern of bullfrogs in which the oscillations of arterial blood gases have been artificiallyeliminated to that of intact animals.The study presented in chapter 3 investigates the role of olfactory and vagal afferentinputs in the control of episodic breathing. Both groups of inputs have been shown to affectventilatory output in other species, however, the effect of removing their respective inputs onPage 27breathing pattern has never been investigated in bullfrogs. In this chapter, the breathing patternof animals in which the olfactory or vagal afferent inputs were surgically eliminated is comparedto that of intact bullfrogs.To investigate the role of pulmonary stretch receptors feedback in respiratory controlspecifically, a decerebrate, paralyzed and unidirectionally ventilated bullfrog preparation wasused in chapter 3 to manipulate PSR feedback directly and relatively independent of other vagalafferent inputs. This “in situ” preparation also made it possible to open both the chemo- andpulmonary stretch receptor feedback loops to assess how the interaction between these inputsmodulates breathing pattern.In chapter 5, the contribution of peripheral feedback is further assessed by recording thebasic respiratory-related motor output that the central nervous system can produce in the absenceof peripheral feedback. This was done using an isolated brainstem-spinal cord preparation.Because the results of this chapter indicated that episodic breathing was an intrinsic property ofthe central respiratory network, the subsequent study (chapter 6) focused on the role of thenucleus isthmi in the control of breathing pattern. This nucleus was chosen because previousworkers have reported important changes in breathing pattern following brainstem transectionsin this area (see section 5.4.2). This chapter evaluates the effects of bilateral lesions of thenucleus isthmi on the “fictive” breathing pattern produced by the in situ preparation. Theimportance of this nucleus in the integration of afferent inputs is also evaluated comparing theeffects of changing PSR and chemoreceptor feedback in animals with and without their nucleusisthmi intact.Page 28CHAPTER 2:THE ROLE OF CHEMORECEPTORS IN THE CONTROL OF EPISODIC BREATHINGPage 29INTRODUCTIONIntermittent breathing patterns are common in lower vertebrates, such as reptiles andamphibians, and contrast with the continuous breathing patterns of birds and mammals by theirapparent lack of constancy and intrinsic rhythm. Many researchers have ascribed the genesisof breathing episodes in amphibians and reptiles to the inherent oscillations of blood 02 and/orC02/pH levels associated with intennittent breathing rather than to the action of a “mammalian-type” central control mechanism. In this model, lung ventilations are turned on when a certainPao2 or Paco2 threshold is reached and breathing is then stopped when the blood gas values havebeen brought back within a certain range (Boutilier, 1988; Shelton and Croghan, 1988; West etat. 1989). The observation that artificial ventilation sufficient to meet respiratory convectionrequirements suppresses ventilation in turtles (Kinney and White, 1977) and toads (West et al.1987; Smatresk and Smits, 1991) supports this concept.Recent studies suggest, however, that episodic breathing does not always reflect thephasic nature of afferent chemoreceptor inputs. Indeed, unidirectionally ventilated toads (Westet at. 1987; Smatresk and Smits, 1991) and alligators (Douse and Mitchell, 1992b) still displayepisodic breathing, although this experimental procedure has been assumed to maintain bloodgases constant and thus produce only tonic chemoreceptor input. Because amphibians andreptiles have intracardiac shunts, however, and some appear to regulate 02 uptake from lungstores during periods of breath holding (Shelton, 1985; Burggren, 1988; West et al. 1992), itis not at all clear that blood gases do remain constant in these animals on unidirectionalventilation (UDV). Thus, it is possible that breathing under such conditions is still due to bloodgas fluctuations, resulting not from intermittent ventilation, but, from intermittent episodes ofblood shunting. The first objective of the present study, therefore, was to monitor arterial P02,Page 30Pco2 and pH continuously with an arterial extracorporeal loop in unidirectionally ventilatedbullfrogs to determine whether intermittent breathing in animals on UDV could still be relatedto arterial blood gas fluctuations. Given that fluctuations in blood gas composition were virtuallyeliminated by UDV in these animals, the second part of the present study analyzed the effectsof phasic vs tonic chemoreceptor input on the components of the breathing pattern. Althoughseveral studies have previously analyzed the effects of UDV on ventilation in anurans, none havespecifically addressed the question of the role of phasic chemoreceptor feedback in the controlof breathing pattern by comparing the breathing patterns of animals with and without phasicchemoreceptor feedback around the same mean levels of blood gases and acid-base status. Thethird part of this study analyzed, in detail, the effects of changing tonic chemoreceptor inputalone on breathing pattern; particularly on the changes in absolute and instantaneous frequenciesof buccal oscillations and lung ventilation. These latter changes have not been documentedpreviously but could potentially offer much insight into the mechanisms underlying the genesisof episodic breathing.Page 31MATERIALS AND METHODSAdult bullfrogs (Rana catesbeiana) of either sex weighing between 277g and 489g (meanmass = 367 ± 28g), were obtained from a commercial supplier. The animals were maintainedindoors in fiberglass sinks continuously supplied with city of Vancouver tap water. The holdingtemperature was 20-22°C. Frogs were exposed to a 12L: 12D photoperiod and fed live locustsat least once a week. The animals were not fed for the two days before surgery was performed.Surgical proceduresBullfrogs were anaesthetized by immersion in a 1gl’ solution of tricainemethanesulfonate (Syndel) buffered to pH 7.0 with NaHCO3. To record buccal pressure, a pieceof polyethylene catheter (PE 50) was inserted into the frog’s buccal cavity via the tympanicmembrane according to the method of Jones (1970). Polyvinyl cannulae (Bolab BB3 15-B) wereinserted into both lungs at the apex via small incisions in the body wall of the flanks to permitunidirectional ventilation (UDV) of the lungs. The pulmonary end of each catheter was heat-flared, secured in place with a purse-string suture and the body wall and skin then suturedseparately.In three animals, the coeliac artery was cannulated occlusively in both directions (i.e.upstream and downstream) with polyvinyl cannulae (upstream: Bolab BB3 17-V/8 or BB3 17-V/4,depending on artery size; downstream: Bolab BB3 17-V/4) pre-treated with a 5% TDMACheparin coating solution (Polysciences, Inc.). The upstream cannula was positioned at thetrifurcation of the coeliac artery with the two systemic arteries. The body wall and the skinwere then sutured independently and both cannulae secured to the skin with sutures. All animalswere allowed to recover from surgery for at least 24h before the onset of experiments.Page 32Experimental proceduresThe system utilized for the continuous measurement of breathing and blood gases isillustrated in Fig. 2. 1. Following recovery, frogs were transposed to an experimental chamberconsisting of a 1.6 1 opaque plastic jar filled with 900 ml of water at ambient temperature (20-22°C). An airline was positioned at the bottom to permit bubbling with either air or a gasmixture (see below). The top compartment of the chamber was ventilated with the same gas.The animals were acclimated to the experimental chamber for at least 2h before the onset of theexperiment.Unidirectional ventilation of the frogs’ lungs commenced one hour before the beginningof the experiment at a rate of 50 mlmirr’, regulated by a flowmeter. Prior to delivery to thelungs, the inflowing gas was humidified by bubbling through an erlenmeyer flask half filled withwater. The lung outflow cannula was connected to a second erlenmeyer flask half filled withwater and connected to an aspiration line. The suction rate was regulated, by a valve, to matchair inflow. Continuous flow of air through the lungs, at normal lung volume, was ensured byvisual comparison of the bubbling rates of the two flasks and by monitoring lung pressure. Thelatter was achieved by installing a T-piece, connected to a pressure transducer, in the inflowcannula. The buccal cannula was then filled with water and connected to a second pressuretransducer.Evperimental protocol1. The effects of unidirectional ventilation on arterial blood gasesThis series of experiments was designed to determine whether oscillations in blood gasesand pH still occurred in animals on UDV. It compared the amplitudes of the respiratory relatedoscillations of blood gas variables and pH associated with spontaneous breathing in UDV (N =Page 333) and non-UDV frogs (N = 3). Once the calibration of the blood P02, Pco2 and pH electrodeswas completed (see below), the two arterial catheters were mounted in series with the electrodes.Blood flow through the arterial extracorporeal loop was set at 1 mlmin’ with a Gilson peristalticpump. Occasional blood samples and measurement of hematocrit revealed no sign of hemolysisor anaemia. The non-UDV group consisted of frogs that had their lung cannulae occluded so thatthey were responsible for their own air convection. Arterial blood gas levels and pH werecontinuously recorded from non-UDV animals breathing air and UDV animals breathing all thevarious gas mixtures described below. Due, in small part, to the technical difficulty of keepingblood flowing through the extracorporeal loop for the length of these experiments and, in largepart, to the consistency of the results, only three animals were used in this portion of the study.2. The effects of unidirectional ventilation on breathing patternThis series of experiments was designed to examine the effects of tonic vs phasicchemoreceptor input on the resting breathing pattern. It compared the breathing patterns ofbullfrogs unidirectionally ventilated with a 2% CO2 gas mixture (N = 6) to those of non-UDVanimals breathing air (N = 6). The latter group consisted of frogs that had their lung cannulaeoccluded so that they were responsible for their own air convection. Because of the constant lungconvection, the UDV frogs were administered a 2% CO2 gas mixture to produce arterial bloodrespiratory status comparable to that of resting frogs which were not artificially ventilated (Westet al. 1987; present study). The gas mixture was delivered to the water, the air space abovethe water and to the lungs of the frogs on UDV so that breathing activity could not alter lunggas composition. UDV frogs were ventilated with the 2% CO2 in air mixture and non-UDVfrogs were breathing air while their lung catheters were occluded for at least one hour beforebreathing pattern was recorded.Page 343. The effects of changing levels of tonic chemoreceptor input on ventilationIn these experiments, bullfrogs (N = 6) were unidirectionally ventilated with thefollowing gas mixtures: air (control); 2, 4, and 6% CO2 in air (hypercapnic series); 15, 10 and5%, 02 in N2 (hypoxic series); 50% 02 in N2 (hyperoxia); and 50% 02 + 6% CO2 balance N2(hyperoxic hypercapnia). The frogs were exposed to each gas mixture for at least one hourbefore breathing pattern was recorded. At the end of the one hour period on each gas, air wassubstituted for the experimental gas mixture in the chamber and UDV line. Recording ofrespiratory variables and blood gases began at least 10 mm before the 1 hour exposure to theexperimental gas mixture, and was continued through the first 30 mm of the recovery period.A recovery period of at least two hours was then imposed before the frog was exposed to thenext gas mixture. Measurement of blood gases and respiratory variables confirmed that thisperiod was sufficient to re-establish “resting” conditions for all variables before exposure to thenext gas mixture. Gases were presented to the animals in random order. Frogs were notsubjected to more than 3 treatments in any one day.Analytical proceduresArterial P02, Pc02 and pH were measured with miniature 02, CO2 and pH combinationelectrodes (Microelectrodes, Inc.) in conjunction with a Radiometer PHM-73 pH/blood gasmonitor. P02 and Pco2 electrodes were calibrated by pumping water equilibrated with preanalyzed gas mixtures (Medigas) at 1 mlmiir’ past the electrodes, while the pH electrode wascalibrated by pumping precision buffers (Radiometer) through the electrode at the same rate.The calibration procedures were done at room temperature before the beginning of eachexperimental run. In order to have an accurate match between the breathing recordings andcorresponding values of blood respiratory variables, the time delay of the extracorporeal loopPage 35was estimated. This was done by measuring the electrode’s response time after the introductionof 10% CO2 in N2 in the UDV line of breathing frogs. The delays were 10, 50 and 80 secondsrespectively for the 02, pH and CO2 electrodes. The mean blood gas and pH values for animalson UDV with different gas mixtures were obtained by averaging the levels of these variablesover the period during which breathing was analyzed. These were roughly constant. Becauseof the large amplitude of the oscillations in blood gases in the non-UDV animals, however,Pao2, Paco2 and pHa readings were taken frequently and then pooled to produce a mean value.The gas mixtures administered to the animals were obtained by mixing air with 02, N2or CO2 with flow meters. The gas mixture fed to the experimental chamber was sampledcontinuously from the inflowing airline and monitored using a Beckman OM- 11 oxygen analyzerand LB-2 CO2 analyzer calibrated with commercially purchased, analyzed gas mixtures. Tidalvolume (VT) was not quantified directly; instead, the time integral of buccal pressure wasmeasured (Sakakibara, 1984b; West et al., 1987). Although there is no doubt that integratedbuccal pressure is not a perfect reflection of tidal volume, there is a relatively tight correlation(although not necessarily a linear correlation) between peak buccal pressure and tidal volume(Wang, 1994). Consequently, multiplying the time integral of buccal pressure changesassociated with lung ventilation (VT index) by lung ventilation frequency (fL) yielded an indexof minute ventilation (V), reported here in arbitrary units. Blood respiratory variables (Pa02,Pac02 and pHa), lung pressure, buccal pressure and the time integral of buccal pressure weredisplayed on a six-channel chart recorder (Gould).It should be noted that episodes are designated on a somewhat subjective basis. Thus,the number of breaths within an episode was obtained by counting the number of large amplitudebuccal movements which occurred in succession with no pause longer than the length of twoventilations between them. Breathing frequency was quantified by analyzing the number ofPage 36respiratory events per unit time (termed the absolute frequency; fL,) as well as by calculatingthe inverse of the period between each two successive breaths in an episode with no pausebetween them and multiplying by 60 (termed the instantaneous frequency; fL1). Only fLabs wasused in the calculation of minute ventilation. The frequency of the small amplitude buccalmovements (fB), known as buccal oscillations, was quantified the same way (i.e. fBabs and fB1).The large difference in amplitude of the buccal movements associated with buccal oscillationsand lung ventilations made these events very distinct (see Fig. 1.1, 2.6).Data analysisValues for breathing and blood respiratory variables were obtained by analyzing a 5minute segment of data recording with a digitizing tablet (Jandel Scientific). In experiments inwhich ventilation rate was low (e.g., normoxia and hyperoxia), analysis of a longer segment wasnecessary to obtain representative values for some variables.All data are presented as means ± 1 standard error of the mean (S.E.M.). The resultswere statistically analyzed using analysis of variance followed by Fisher’s LSD multiplecomparison test. Comparisons involving only two means (e.g., Table 2.1) were done byunpaired t-test. Unless otherwise stated, 5% was taken as the fiducial limit of significance.Page 37RESULTS1. The effects of unidirectional ventilation on arterial blood gas oscillationsThe top three traces (pHa, Paco2 and Pao2 respectively) of the left panel of Fig. 2.2typify the fluctuations in blood respiratory variables observed during breathing episodes inspontaneously ventilating frogs. These traces show the pronounced effect of lung inflation ongas exchange, as pHa and Pao2 increased while Paco2 decreased immediately after a lunginflation cycle. It is noteworthy that buccal movements did not affect blood respiratory variablesand single lung ventilations had very little effect. The amplitudes of the respiratory relatedoscillations in Paco2 and pHa associated with breathing episodes were virtually abolished infrogs on UDV (Fig. 2.2 right panel, Table 2.1). Although some small Pao2 oscillationspersisted during UDV, they occurred randomly and were not correlated with the onset andtermination of breathing episodes. The mean amplitude of the Pao2 oscillation was greatest inanimals on UDV with air or 50% 02. In animals on UDV with air it was 1.5 ± 0.08 Torr,significantly less than that observed in spontaneously breathing animals (3.8 ± 0.5 Torr; Table2.1). In animals on UDV with 5% 02, the 02 oscillations were abolished.2. The effects of unidirectional ventilation on breathing patternBecause of the increased pulmonary convection during UDV compared to spontaneousventilation, it was only when 2% CO2 was added to the UDV gas mixture that arterial bloodgases and pH comparable to those of the spontaneously ventilating (non-UDV) frogs wereobserved (Table 2.1). This also produced comparable breathing patterns (Fig. 2.2, Table 2.2).The similarity was not only in terms of magnitude, but also in the temporal distribution ofbreathing events. Both groups had similar levels of fLabs, number of breaths per episode,mean VT index and V index. Finally, both groups had similar values of fBabs and fB1, (TablePage 382.2). In short, there were no significant differences between the mean levels of blood gases andpH and ventilatory variables between the two groups.3. The effects of changing levels of tonic chemoreceptor input on ventilation3.1. Carbon dioxide/nHThe addition of CO2 to the UDV gas mixture stimulated ventilation (Fig. 2.3). The meanvalues of Pao2 after 60 mm of UDV were not different regardless of whether the animals werebreathing 0%, 2%, 4% or 6% CO2 in air. Furthermore, statistical analysis revealed nodifference between mean values of ‘7 index for the 6% CO2 and 6% CO2 + 50% 02 groups eventhough Pao2 was significantly higher in the hyperoxic hypercapnic group (Table 2.3). Thus theincrease in the ‘J index was directly proportional to the increase in PacO2. Increases in the Vindex were due to increases in both fLabs and the VT index (Fig. 2.3). The increase in fL,5 wasdue to an augmentation of the number of breaths per episode and a reduction in the length ofthe pause between episodes; the fL5, remained constant (Fig. 2.3). At the higher levels ofinspired CO2 (4% and 6% C02), 1 frog out of 6 breathed continuously. UDV with C02-richgas mixtures resulted in a progressive decrease in the number of buccal oscillations per unit time(Fig. 2.4A) primarily due to a reduction in the non-ventilatory pauses during which buccaloscillations occur. The fB5, also showed a progressive decrease with elevation of PaCO2although this variable was difficult to quantify in animals breathing 6% CO2 as buccaloscillations were virtually absent in this group (Fig. 2.4A).Page 39Figure 2. 1: Schematic representation of the experimental system utilized for continuousmeasurement of breathing and blood gases.Page 40Page 41Figure 2.2: Continuous traces of blood respiratory variables and breathing patterns (depictedby changes in buccal and lung pressures) of a frog unidirectionally ventilated with 2%CO2 in air (right panel) and spontaneously ventilating air with no intervention (leftpanel).Page 42NonunidireCt10nallY ventilated Unidirectionally ventilated(am (2% C02)7.777.8885 [pHa]22.57.2 r %_.S%___b%% PaCO2(Torr)16.2 L I 21i80Pa02(ToIT)70 ]75Lung1Buccal(cmH2O)____________ __ ____ __ ____1’ I I I I I l UI I1 minutePage 43Figure 2.3: The effects of manipulating arterial Pco2 on A) minute ventilation index, B)instantaneous (open circles) and absolute (closed circles) breathing frequency, C) numberof breaths per breathing episode and D) tidal volume index. Mean values significantlydifferent from the air (UDV) group (PaCO2 = 5.2 ± 0.8 Torr) are either indicated by* (p < 0.05) or + (p < 0.10). In B, + indicates fL1 is significantly different fromthe corresponding fLabs (p < 0.05) and the numbers in parentheses indicate the numberof values used to determine that mean (only one animal still showed any ventilationon room air).Page 44A BVentilationindex (arb. units)10830 45(Torr)Breathingfrequency (mm )(4) (6)30 45(Torr)* * 706050403020100400300200100OLBreaths per episodeLFC**D***64200 15Pa0Tidal volume index(arb. units)302520151050_________0 15Pa0Page 45Figure 2.4: Instantaneous (open symbols) and absolute (closed symbols) buccal oscillationfrequency as a function of A) arterial Pco2 or B) P02. Mean values significantlydifferent from the air (UDV) group (PacO2 = 5.2 ± 0.8 Ton; Pa02 = 87 ± 21 Ton)are either indicated by * (p < 0.05) or + (p < 0.10). + indicates is significantlydifferent from the corresponding fBabs (p < 0.05) and the numbers in parentheses indicatethe number of fB1 values used to determine that mean.Page 46A B(3)(3)+ ÷ (2)E 70 f 70 (6) ++50 50+ *30_____-30- 20 20c)________ ___ _+10C—0_ __ _ _ __ _ _ _ _ _ __C)0 15 30 0 50 100 200250Pa0 (Torr) Pa0 (Torr)Page 47Figure 2.5: The effects of manipulating arterial P02 on A) minute ventilation index, B)instantaneous (open circles) and absolute (closed circles) breathing frequency, C) numberof breaths per breathing episode and D) tidal volume index. Mean values significantlydifferent from the air (UDV) group (Pao2 = 87 ± 21 Torr) are either indicated by * (p< 0.05) or + (p < 0.10). In B, + indicates is significantly different from thecorresponding fLabs (p < 0.05) and the numbers in parentheses indicate the number offL15 values used to determine that mean.Page 48Ventilationindex (arb. units)tA Breathingfrequency (min’)BBreaths per episode C Tidal volume index(arb. units)D0 50 100 200250Pa,- (Torr)‘-‘2__III__0 50 100 200250Pa0 (Torr)2400300200loot70 -60(5)505 +: (6)302010010 **8642302520151050*Page 49Figure 2.6: Buccal pressure recordings of adult bullfrogs illustrating the effects of differentventilatory stimulants on the instantaneous breathing (fL) and buccal oscillationfrequencies (fB).Page 50ttI___Buccal pressure(an H20)vi3Page 513.2. OxygenUnder normoxic conditions, breathing was practically abolished in 4 animals out of 6 onUDV. In these animals, only sporadic single breaths were observed. This pattern wascomparable to that of frogs unidirectionally ventilated with a 50% 02 in N2 gas mixture;hyperoxia did not suppress breathing further (Table 2.3). The mean Pao2 values of frogsartificially ventilated with air were not different from those of non-UDV animals. The bloodacid-base changed during UDV with air, however, as PacO2 was reduced in comparison to thenon-UDV group although the concomitant increase in pHa was not significant (Table 2.3).Hypoxemia provoked a modest hyperventilation in bullfrogs (Fig. 2.5). This wasprimarily due to an increase in the VT index as artificial ventilation of the lungs with hypoxicgas mixtures did not significantly affect fLabs nor It did, however, alter breathing patternonce Pao2 went below 20 Torr; there was a significant augmentation in the number of breathsper episode associated with a decrease in the pause between episodes. This increase in thenumber of breaths per episode was much less than the augmentation seen in frogs breathing 4%(p < 0.05) and 6% CO2 (p < 0.10). Altering Pao2 did not affect fBabs significantly (Fig. 2.4B),and only once Pao2 reached 15 Torr was a reduction in observed (Fig. 2.4B).Page 52Table 2.1: The effect of unidirectional ventilation on the amplitude of arterial blood gasfluctuations in spontaneously breathing bullfrogs. Values are means ± SE. *Significantly different from the corresponding non-UDV value (unpaired t-test; p <0.05). The number of parentheses indicates the total number of breathing cyclesquantified.Page 53Non-UDV UDV(breathing ambient air) (ventilated with 2% C02)mean value oscillation mean value oscillationamplitude amplitudePaco2 (Torr) 11 ± 2.7 0.31 ± 0.06 17.2 ± 0.9 0.0 ± 0.0*(N = 3) (n = 11) (N = 3) (n = 11)pHa 7.90 ± 0.02 ± 7.73 ± 0.04 0.001 ±0.04 0.003 (N = 3) 0.001*(N = 3) (n = 11) (n = 11)Pao2 (Torr) 50 ± 9.2 3.8 ± 0.5 61 ± 5.8 1.5 ± 0.08*(N = 3) (n = 11) (N = 3) (n = 11)Page 54Table 2.2: The effect of unidirectional ventilation on selected ventilatory variables inspontaneously breathing bullfrogs. Values are means ± SE. * Significantly differentfrom the corresponding non-UDV value (unpaired t-test; p < 0.05).Page 55Non-UDV UDV(breathing ambient air) (ventilated with 2% C02)(N=3) (N==3)MinuteVentilation 89 ± 25 98 ± 36(arb units)Absolutebreathing 6 ± 1.1 9 ± 3.6frequency (min’)Instantaneousbreathing 39 ± 8.0 43 ± 5.3frequency (min’)Tidal volumeindex (arb units) 16 ± 3.6 9 ± 2.1Breaths perepisode 2.8 ± 0.8 2.3 ± 0.6Absolute buccaloscillation 29 ± 8.7 27 ± 1.6frequency (miii’)Instantaneousbuccal oscillation 60 ± 3.5 50 ± 4.9frequency (miii’)Page 56Table 2.3: The effects of unidirectional ventilation of the lungs with different gas mixtures onventilatory variables and selected blood respiratory variables. Values are means ± SE.* Significantly different from the corresponding Air (UDV) value (p < 0.05); +significantly different from the corresponding Air (No UDV) value (p < 0.05). +significantly different from the corresponding Air (No UDV) value (p < 0.10). N =6 for respiratory variables, 3 for arterial blood gases and pH and 1 for breaths perepisode with UDV with 50% 02 (the other animals did not breathe while on this gasmixture).Page 57AIR50%02Air5%026%Co250%02+(noUDV)(UDV)(UDV)(UDV)(UDV)6%C02.(UDV)Minute89±252±1+5±5+66±6+317±48*+251±38*+ventilation(arbunits)Absolutebreathing6±1.10.1±0.10.3±0.24.2±0.823±7.6*+21±2.6*+frequency(min1)Tidalvolumeindex16±3.6*3±2.6+4±4.1+17±2.1*18±3•3*13±2.2*(arbunits)Breathsperepisode2.8±0.8*1.21±0—1—4.1±0.8*5.6±0.7*+5±0.7—f—Absolutebuccal29±8.710.±6.725±9.515±3.82±2.24±3.2oscillationfrequency(miii’)plla7.90±0.048.1±0.18.0±0.18.2±0.27.45±7.47±0.03*+0.01-J—Paco2(Ton)11±2.76±1.15.2±0.8+4.0±0.5+42±3.2*+36±1.3*+Pao2(Torr)50±9226±87±2115±8*83±13134±28+31*+00DISCUSSION1. The effects of unidirectional ventilation on arterial blood gas oscillationsMany recent studies have used UDV as a technique to manipulate arterial blood gasesin amphibians (West et al. 1987; Smatresk and Smits, 1991) and reptiles (Douse and Mitchell,1992). These workers assumed that this technique maintained arterial blood gases and pHconstant in spontaneously breathing animals. The present study confirms the efficiency of UDVfor manipulating gas exchange and controlling arterial blood gases even in an intermittentbreather. UDV completely eliminated the oscillations in Paco2 and pHa associated withventilation with any gas mixture. Admittedly, some small residual oscillations in Pao2 remainedduring UDV with normoxic or hyperoxic gas mixtures but not with hypoxic (5 % 02) gasmixtures. These oscillations in Pao2 did not correlate with the breathing episodes which is notsurprising given that lung gas content did not change as a result of the lung ventilations. Giventhe large partial pressure gradient for 02 between lung gas and arterial blood during UDV withnormoxic or hyperoxic gases, the small oscillations which remained could well reflect smallchanges in right to left or left to right blood shunting. Such changes in shunting pattern areusually associated with periods of ventilation (West and Burggren, 1984). It is still possible thatthis was the case in the present study but that due to circulation delays and patterns of bloodmixing, all correlation between the ventilatory events and the oscillations are lost.2. The effects of unidirectional ventilation on breathing patternThe present results demonstrate that removing phasic chemoreceptor input while retainingthe same mean level of chemoreceptor input, does not alter the breathing pattern of frogs. Thelack of any significant differences between the breathing patterns of animals with and withoutphasic arterial chemoreceptor feedback would suggest that such feedback plays very little rolePage 59in the control of the breathing pattern. This conclusion assumes that the small oscillations seenin Pao2 during UDV with normoxic or hyperoxic gases do not play a role in initiating orterminating breathing. Since large changes in Pao2 alone had relatively little effect onrespiratory drive (Fig. 2.5), it is difficult to imagine that the small oscillations in Pao2 seen inanimals on UDV, or even the larger respiratory related oscillations seen during spontaneousventilation, for that matter, affect ventilation. This is supported by the lack of correlationbetween the small oscillations and ventilatory activity in frogs unidirectionally ventilated with2% CO2. Also, similar oscillations remained during UDV with room air or 50% 02 yetbreathing was virtually eliminated in most animals. Finally, the oscillations in 02 werepractically eliminated in animals during UDV with hypoxic gas mixtures at which time breathingepisodes were significantly increased. It remains possible that oscillations in P02 and/orPco2/pH still occur at the local tissue level of the various chemoreceptors involved in regulatingventilation but given the apparent constancy of the arterial blood supplying these sites duringUDV, this does not seem likely.This conclusion is not consistent with the hypothesis that breathing episodes are dictatedsolely by oscillations of blood gases and pH (Shelton and Croghan, 1988). According to thishypothesis, one would predict that UDV with 2% CO2 would result in continuous breathing;despite ventilation, Paco2 would never fall sufficiently to turn the episodes off. This was neverobserved; the breathing pattern of the frogs on UDV with 2% CO2 remained episodic. Thisfinding suggests that, like turtles, the basic output of the central respiratory rhythm generatorof the frog is episodic, even under conditions during which all sensory feedback appears tonic(Douse and Mitchell, 1990b). West et al. (1987) also raised this possibility but despite dataidentical to ours, concluded that it was still “attractive to think” that the periodicity ofintermittent lung ventilation in undisturbed animals depended on blood gas changes duringPage 60periods of apnea and lung ventilation. It is possible that larger changes in blood gases whichoccur during periods of submergence play a role in regulating breathing pattern but we mustconclude that the smaller oscillations seen in animals resting with free access to air do not appearto do so.3. The effects of changing levels of tonic chemoreceptor input on ventilationThe effects on the breathing pattern of Rana catesbeiana of changing tome levels ofchemoreceptor input by UDV with hypoxic and hypercapnic gases were comparable to thosereported for Bufo marinus (West et al. 1987; Smatresk and Smits, 1991). UDV with normoxic,normocapnic or hyperoxic, normocapnic gas would eliminate spontaneous ventilation in mostanimals suggesting that chemoreceptor input of a certain level was required to initiate ventilation.Both decreases in Pao2 and increases in Paco2 (with concomitant decreases in pH) would servethis role although hypercapma was a much more potent stimulus than hypoxia. Increases inbreathing frequency were due entirely to decreases in the non-ventilatory periods; theinstantaneous breathing frequency remained constant at all levels of respiratory drive. There wasa notable change in breathing pattern as respiratory drive increased. At lower levels ofrespiratory drive the pattern consisted of many single breaths and a few small episodes. Asrespiratory drive increased, more of the breaths occurred in episodes until, at high levels ofrespiratory drive, almost all breaths occurred as part of lung inflation cycles. Thus, althoughthe chemoreceptor input was tonic at each level of drive, breathing was episodic. The buccalpressure amplitude and time integral of buccal pressure were both augmented with increases inrespiratory drive stemming from both hypoxia and hypercapnia. These increases were due inpart to a genuine increase in the pressure amplitude associated with individual breaths, whethertaken singly or as part of a lung inflation cycle, and in part due to the fact that more and morePage 61of the breaths were part of lung inflation cycles. Because of the increase in lung volume andlung pressure during a lung inflation cycle, the buccal pressure swings associated with eachsuccessive breath in a lung inflation cycle must increase whether VT increases or not. Themodest response to hypoxia was primarily due to the small increase in breathing frequency whichreflects the fact that neither the number of breaths per episode of breathing nor the length of thepause between breaths or breathing episodes changed much until Pao2 approaches the critical P02(30-37 Torr for Bufo marinus; POrtner et at. 1991). This suggests that intracellular acidosis maybe required to produce an increase in the number of breaths per episode of breathing. Overallthe data suggest that tonic chemoreceptor input not only plays a role in determining the totallevel of minute ventilation but also plays a role in determining breathing pattern. Thus theconversion from single breaths to episodes of breathing and finally to lung inflation cycles willall occur with an adequate rise in tonic afferent input; these changes in pattern do not requirechanges in phasic afferent modulation. These results are also very similar to those obtainedfrom animals breathing spontaneously without the intervention of UDV (Maclntyre and Toews,1976; Boutilier and Toews, 1977; West and Burggren, 1982; Kruhøffer et at. 1987; Boutilier,1988; Branco et al. 1992). Indeed, comparison with literature values support the conclusion thatphasic chemoreceptor input plays little role in the control of breathing pattern.4. The significance of instantaneous buccal oscillation and lung ventilation frequenciesBecause respiratory events occur intermittently in the bullfrog, the concept of a centralpattern generator for breathing in these animals has not received much thought. But, althoughno intrinsic rhythm is obvious in overall breathing events, there is an intrinsic rhythm to buccaloscillations as well as to the breaths occurring in a breathing episode (fL1). It has beensuggested that buccal oscillations may be homologous to gill ventilations in fish and their rhythmPage 62may reflect vestiges of the central rhythm generator for gill ventilation (Smatresk, 1990). Buccaloscillations and lung ventilations are produced by the same muscles. The primary differencebetween these two events is the force of the contraction and the positions of the glottis andnares. Lung ventilations are associated with more forceful contractions with the glottis open andnares closed; buccal oscillations are associated with less forceful contractions with the nares openand the glottis closed. It appears that in resting animals, buccal oscillations occur continuouslyduring periods of apnea and that lung ventilations normally occur at a time when another buccaloscillation would have been initiated (Figs. 1.1 and 2.6). Regardless of the level of respiratorydrive and fLabs, there appears to be an intrinsic rhythm to lung inflation events as revealed bythe Increasing respiratory drive simply appears to result in this rhythm being expresseda greater percentage of the time.These observations suggest at least two possible scenarios. The first is that there is asingle rhythm generator whose output is integrated with inputs from higher centers andperipheral feedback (mechano- and chemoreceptors) at two distinct pattern generators. At lowlevels of respiratory drive, only output from the pattern generator driving buccal oscillations isproduced but as respiratory drive increases, output is generated from the pattern generatordriving lung inflation which leads to the increase in the force of buccal contraction and theswitch in the state of the nares and glottis. The other possibility is that there are two distinctrhythm generators which are entrained under most circumstances with expression of the lungrhythm being conditional upon a higher level of central and/or peripheral receptor input.One set of data from the present study does argue for the latter case. Hypercapnia hadno effect on fLjflSL (Fig. 2.3) but did reduce both the occurrence of buccal oscillations and theirinstantaneous frequency when they did occur (Figs. 2.4 and 2.6). This might suggest that thereare separate rhythms for lung inflation and buccal oscillation which can be uncoupled. GivenPage 63the magnitude of the changes and the small number of observations of both fL1, and fB1 at lowlevels of Paco2 and of fB18 at high levels of Paco2, however, this conclusion should be treatedwith caution. Most animals did not breath when PaC02 was low and buccal oscillations werealmost eliminated in all animals when Paco2 was high. The disappearance of buccal oscillationsand the concomitant augmentation of lung inflation under hypercapnic and/or acidemic conditionshave been reported in many previous studies (De Marneffe-Foulon, 1962; Maclntyre and Toews,1976; Smith et al. 1991a).That resting and were slightly different (Figs. 2.3, 2.4, 2.5 and 2.6) might alsoargue favourably for two separate rhythm generators. It is also possible that the slower fL1,simply reflects the longer interval required to pump a larger volume of air during a lungventilation as opposed to a buccal oscillation and that during this time the intrinsic rhythm ismodified by pulmonary feedback.Extracellular recording from in vitro brain stem spinal cord preparations from Ranacatesbeiana tadpoles and adults have been obtained recently which reveal that it is possible tomanipulate the two types of buccal activities independently with pharmacological agents (Walkeret al., 1990; Smith et al. 199 la; McLean, 1992). The data, however, only argue that there areseparate pattern generators and sheds no light on the question of central rhythm generation.Whatever the case, the observation that breathing is completely suppressed whenconvective requirements are met by UDV (West et al. 1987; Smatresk and Smits, 1991; presentstudy) indicates that lung ventilation is conditional upon a minimal stimulatory input (West etal., 1987; Smatresk and Smits, 1991; Smatresk, 1990; Milsom, 1990). The fact that breathingwill still occur in episodes as respiratory drive increases even though the chemoreceptor inputis tonic raises questions as to what turns breathing episodes on or off. Clearly the tonicchemoreceptor input influences the length of the non-ventilatory period as well as the numberPage 64of breaths in a breathing episode but not through a simple system of “on” and “off” thresholds.The constancy of the observed once ventilation is initiated suggests there is a conditional,central rhythm generator involved. Our data does not allow us to determine whether buccaloscillations and lung inflations are generated by a single, or distinct, central rhythm generator(s).Finally, the fact that lung ventilations always occur at a time when a buccal oscillation wouldotherwise have occurred suggests that if there are separate rhythm generators, they are entrainedto a large degree.Page 65CHAPTER 3:THE ROLE OFC02-SENSITIVE OLFACTORY AND PULMONARY RECEPTORS INTHE CONTROL OF BREATHING PATTERNPage 66INTRODUCTIONMuch research has focused on the role of arterial and/or central chemoreceptors in thecontrol of episodic breathing in lower vertebrates (Douse and Mitchell, 1992a, b; Smatresk andSmits, 1991; West et a!., 1987; chapter 2). Recent studies on anuran amphibians have shownthat some minimal level of chemoreceptor drive is required for the expression of any respiratoryrhythm (Smatresk and Smits, 1991; chapter 2). Furthermore, they have demonstrated thatalthough exposure to increasing levels of hypoxia or hypercapnia promotes a more continuousbreathing pattern by increasing the number of breaths in each breathing episode and reducingthe duration of the non-ventilatory period between episodes, phasic changes in chemoreceptorafferent input are not required for the onset and cessation of breathing episodes in these animals(Smatresk and Smits, 1991: West et al., 1987; chapter 2).The work reported in the present study focuses on the potential contribution of two otherreceptor groups to the genesis of episodic breathing. The first group consists of the olfactoryreceptors situated in the nasal mucosa. These chemoreceptors have afferent projections in thetrigeminal and olfactory nerves (Sakakibara, 1978). In bullfrogs, it is known that tonicstimulation of these receptors with CO2 depresses ventilation transiently (Coates and Ballam,1990; Sakakibara, 1978). This contrasts with the overall stimulatory effect of hypercapnia onventilation (for review, see Milsom, 1990a; Shelton et al., 1986) and implies that the increasein episodic breathing (increases in the number of breaths in each episode, the incidence ofepisodes and the overall breathing frequency) during hypercarbia is a product of a depressantinput from upper airway receptors interacting with a stimulatory input from arterial and centralchemoreceptors. In lizards, transection of the olfactory peduncles results in a more continuousbreathing pattern during acute hypercarbia (Coates et al., 1990), suggesting that the interactionPage 67between olfactory and chemoreceptor inputs may play a role in initiating/terminating breathingepisodes.Pulmonary receptors are the second group ofC02-sensitive airway receptors examinedin this paper. The only receptors known to be CO2 sensitive which arise from the lungs ofanuran amphibians are the pulmonary stretch receptors (PSR). The mechanoreceptors ofamphibians report to the CNS via afferent fibres in the vagi (Kuhlmann and Fedde, 1979;Milsom and Jones, 1977). These receptors are mostly slowly adapting and their firing ratesdecrease when the intrapulmonary CO2 concentration is increased (Kuhlmann and Fedde, 1979;Milsom and Jones, 1977). Lung inflation inhibits (and deflation stimulates) breathing in allspecies of adult amphibians studied so far (for review see Milsom, 1 990b; Shelton and Boutilier,1982; Shelton et al., 1986). Furthermore, lung deflation increases the number of breaths in theongoing or subsequent breathing episode (for review, see Shelton and Boutilier, 1982). Theconsensus that emerges from these findings is that in amphibians, the initiationltermination ofbreathing episodes could be modified by the C02-modulated feedback from mechanoreceptorsin the lungs.In light of this, the present study examined the contribution of these two groups ofreceptors to the initiation/termination of breathing episodes by comparing the episodic breathingpattern of intact bullfrogs to that of frogs in which the olfactory or pulmonary receptor afferentinput was surgically eliminated. The comparisons were made under normo- and hypercarbicconditions.Page 68MATERIALS AND METHODSThe experiments were performed on 33 adult bullfrogs (Rana catesbeiana) of either sexweighing between 120 and 289g (mean mass = 209 ± 5g). The animals were obtained froma commercial supplier and maintained according to the protocol described in chapter 2.Surgical procedures.Bullfrogs were anaesthetized by immersion in a ig N solution of tricaine methanesulfonate (Syndel) buffered to pH 7.0 with NaHCO3. To record buccal pressure, a piece ofpolyethylene catheter (PE 50) was inserted into the buccal cavity via the tympanic membraneaccording to the method of Jones (1970). This study involved six groups of frogs. The controlgroup (N = 6) underwent no further surgical procedures; the frogs were returned to their tankfor recovery. For the olfactory denervated group (N = 6), a small hole was drilled with aDremel tool in the anterior part of the skull, above the olfactory lobes. Fine bone shears wereused to open the skull, and the dura above the olfactory nerves was removed with fine forceps.Both olfactory nerves were then lifted and sectioned with small scissors. The cranial openingwas repaired with bone wax (Ethicon, W-3 1) before the skin was sutured. The same surgicalprocedure was repeated on the sham olfactory denervated group (N = 6) except that theolfactory nerves were left intact. In the pulmonary vagotomy group (N = 6), a 2-cm incisionwas made behind each scapula. The underlying musculature was also cut to expose the rostralends of the lungs. The pulmonary branch of the vagus nerve was then identified on both sidesand sectioned with small scissors. Muscular and dermal tissue layers were sutured separately.For the sham vagotomy group (N = 6), the surgical procedure was identical to the one justdescribed except that the vagi were not sectioned. A sixth group of frogs (N = 3) was used toquantify the changes which occurred in airway CO2 profiles and arterial blood gases throughoutPage 69the experimental protocol. In this group, a second polyethylene catheter (PE 20; length = 15cm) was inserted into the buccal cavity. This catheter was positioned close to the glottis.Although the measurements obtained with this catheter will not correspond to the CO2 levels atthe receptor sites, they will reflect the nature and magnitude of CO2 changes in the buccal cavityand serve as a rough indication of the magnitude and direction of changes in CO2 concentrationwhich occurred in the airways (choanae and bronchi) during these experiments. The femoralartery was also cannulated occiusively with a polyethylene catheter (PE 50) in this group. Thiscatheter was used for arterial blood sampling. Each bullfrog was allowed to recover fromsurgery for at least 48h.Experimental procedures.Frogs were transferred from the recovery tank to a 1.6 1 opaque plastic jar filled withapproximately 900 ml of water at ambient temperature (20-22°C). The buccal cannula, forbuccal pressure measurements, was filled with water and connected to a pressure transducer(Narco Scientific). In the sixth group of animals the second buccal cannula, for CO2measurements, was connected to a CO2 analyzer (Beckman LB-2). A gas line was positionedat the bottom of the chamber and the water was bubbled with either air, 4.5 % CO2 in air or 6%CO2 in air. The top compartment of the chamber was ventilated with the same gas as the onebubbled into the water. The total flow rate of gas to the experimental chamber wasapproximately 11 per minute. Direct measurements of CO2 levels showed that, at this flow rate,following a step change in the CO2 level of gas supplied to the chamber, the gas phase of thechamber turned over in 60 seconds and the gas composition of the water phase equilibrated withthe new gas mixture in roughly 5 minutes. The animals acclimated to the experimental chamberfor at least 2h before the onset of the experiment.Page 70Experimental protocol.1. The breathing pattern of intact bullfrogs under normo- and hypercarbic conditions.In this series of experiments the breathing pattern of intact bullfrogs was quantified firstat rest, when the frogs were breathing air. Breathing was then assessed after lh of exposure to6% CO2 in air and again after air was substituted for the 6% CO2 in air mixture.2. The role of olfactory receptors in the control of episodic breathing.This series of experiments assessed the contribution of upper airway (olfactory) receptorsto the shaping of breathing pattern in bullfrogs. It compared the breathing pattern of olfactorydenervated frogs to that of sham operated and control frogs. Comparisons of ventilatoryvariables were made for each of the different gas regimes described in the previous section.3. The role of pulmonary receptors in the control of episodic breathing.This series of experiments assessed the contribution of pulmonary receptors to theproduction of breathing pattern in bullfrogs. It compared the breathing pattern of vagotomizedfrogs to that of sham operated and control frogs. Comparisons of ventilatory variables weremade for each of the different gas regimes described in section 1.4. Kinetics of changes in airway and arterial blood CO levels.Because a paradoxical increase in breathing was observed immediately following theswitch from hypercarbic to normocarbic conditions (see Results), an additional series ofexperiments was performed to quantify CO2 levels in the buccal cavity and arterial blood in anattempt to identify a possible underlying mechanism. Breathing, arterial blood gases and buccalcavity CO2 levels were assessed when the frogs were breathing air, after 30 mm of exposure toPage 714.5% CO2 and throughout the first 30 mm after air was substituted for the hypercarbic gasmixture. The restricted availability of commercial gas mixtures at the time the experiments wereperformed explains the use of 4.5 % CO2 in air in this series. To assess changes in respiratorydrive throughout the experiment, blood samples (450 jd each) were taken for measurements ofarterial blood gases. The samples were taken at rest, after 30 mm of hypercarbia and after 0.5,1.0, 1.5, 2.0, 15.0 and 30.0 miii of recovery. The blood was reinjected into the animal aftereach measurement. In this series, experiments were performed twice on each animal. Arecovery period of at least 2h was allowed between each run.Analytical procedures.The tidal volume index, breathing frequency, minute ventilation index, instantaneousfrequencies (lung and buccal oscillation) and breathing pattern were quantified according to themethods and criteria reported in chapter 2. Recording of respiratory variables began at least 10mm before the 60 mm exposure to the experimental gas mixture, and was continued through thefirst 30 mm of the recovery period.In the three animals in which buccal cavity CO2 profiles and arterial blood gases werequantified, arterial blood gases were measured at room temperature immediately after samplingwith Radiometer PCO2 (E 5036-0), Po2 (E 5046-0) and pH (G297/G2) electrodes in conjunctionwith a Radiometer PHM-73 pH/blood gas monitor. The PCO2 electrode was calibrated withhumidified gases provided by a Radiometer GMA2 precision gas pump, the electrode withair saturated water and deoxygenated water (5 mM sodium bisulfate) and the pH electrode withRadiometer precision buffer solutions. The CO2 level in the buccal cavity was measured witha Beckman LB-2 CO2 analyzer calibrated with room air and a pre-analyzed gas mixture. Thesampling rate of the analyzer was 40 ml. min’ and the time constant of the system was 1.1 sec.Page 72This time constant was long relative to the time during which gas was expelled from thebullfrog’s lungs (0.4 see). Consequently, although the end expired CO2 values measured in thisstudy were accurate, end inspired CO2 values were overestimated when the non-ventilatory pausebetween breaths was shorter than the response time of the CO2 analyzing system used here. Thiswas most evident when the frogs were breathing rapidly, such as during the post-hypercarbicperiod. The magnitude of the error was not large since the occurrence of non-ventilatory pausesduring the post-hypercarbic period, which allowed the end inspired values to achieve their realvalues, produced only small drops in the values measured (Fig. 3.5).Data analysis.Values for breathing variables under normocarbic and hypercarbic conditions wereobtained by analyzing a 5 mm segment of data recording with a digitizing tablet (JandelScientific). The segments quantified were taken immediately before hypercarbia and after 1 hof exposure to 6% CO2 in air. The post hypercarbia data is the mean data for the first 2 mmsegment of data recorded immediately at the onset of the post-hypercarbic period (recovery).All data are presented as means ± 1 standard error of the mean (S. E. M.). The results werestatistically analyzed using an analysis of variance followed by a Fisher’s LSD multiplecomparison test.Page 73Figure 3.1: A buccal pressure recording illustrating the breathing pattern of (A) intact, (B)olfactory denervated and (C) vagotomized bullfrogs while breathing air (left panel) andafter lh of breathing a 6% CO2 in air gas mixture (right panel). The arrow indicates theonset of recovery by reintroducing air into the experimental chamber. Note the smallbuccal oscillations during air breathing between the larger buccal pressure changesassociated with lung ventilation.Page 74UiA)(ONTROI.AIR6%CO1AIRH)OLFACTORYIWNERVATFI)4:t (L.ii___________________________________C)vA(;oToMIzEnIp :iillLI1LILUJJjwj.iJlliiLJuflLJ 11iLJn[LIl_P!!_InhCRESULTS1. The breathing pattern of intact bulifrogs under normo- and hypercarbic conditions.Figure 3. 1A illustrates the breathing pattern of an intact bullfrog breathing air, after 60mm of exposure to 6% C02, and upon return to breathing room air. While initially breathingair, breaths occurred mostly in doublets or single breaths. Small amplitude buccal oscillationsoccurred during the non-ventilatory pause between breathing episodes. Pre-hypercarbic valuesof fL1 and were not different (66 ± 5.6 min1 and 68 ± 6.3 min’, respectively). After60 mm of exposure to 6% C02, buccal oscillations had virtually disappeared; fB1, values couldonly be calculated in two animals. The mean values recorded were significantly greater thanthe corresponding two fL1 values at this time. Absolute breathing frequency (fLabs) increased74% (Fig. 3. 2A) despite a 47% reduction in (Fig. 3. 3A), because the number of breathsper episode increased (Fig. 3. 2B). The VindeX increased by 121 % of the normocarbic value (Fig.3.2D) due to increases in both fLthS and the VTlndex (160% increase; Fig. 3.2C).The transition from 6% CO2 to air caused a further increase in the fLabs (300% greaterthan values recorded during norinocarbia, 130% greater than values recorded duringhypercarbia; Fig. 3.2A). The post-hypercarbic level of fL5 was 41% greater than thehypercarbic one, but was still 26% lower than the normocarbic value (Fig. 3.3A). Sincebreathing was continuous (i.e. without any non-ventilatory pause) in most frogs during the posthypercarbic period, no value is reported for breaths per episode in the control animals at thistime (Fig. 3. 2B). The Vindex was 255% greater than values recorded during normocarbia (61 %greater than values recorded during hypercarbia; Fig. 3. 2D) due to increases in fLab, alone, asthe VTlndex returned to normocarbic levels (Fig. 3.2C).Page 76Figure 3.2: The effects of changing CO2 levels on A) absolute breathing frequency, B) thenumber of breaths per episode of breathing, C) the tidal volume index and D) theventilation index. Values were obtained from control (N = 6), vagotomized (N = 6)and olfactory denervated (N = 6) frogs. A value significantly different from thecorresponding control value is indicated by * (p < 0.05) or ** (p < 0.10). A valuesignificantly different from the corresponding normocarbic value is indicated by + (r< 0.05) or + (p < 0.10). A value significantly different from the correspondinghypercarbic value is indicated by + (p < 0.05) or + + (p < 0.10). Note that for somevalues of breaths per episode the number of measurements was not equal to 6 becausethe animals were breathing continuously during the post hypercapmc period. In thosecases, the N value is indicates in parentheses. This was the case for all animals in thecontrol group, for which no values of breaths per episodes were obtained.Page 77Control• Pulmonary vagus sectioned• Olfactory nerve sctionedAbsolute breathingfrequency (mi&’)•L.Tidal volumeindex (arb units)0Normocarbja Hypercarbia PostHypercarbjac Ventilation Dindex (arb units)**20L ,A Breathsper episode2520 715//10B• (4)+504o302010*600r-* •400k •2000 Normocarbia Hypercarbia PostHypercarbiaPage 782. The role of ofactoiy receptors in the control of episodic breathing.The breathing patterns of control frogs are compared with those of sham operated animalsin Table 3.1. Comparisons were performed on data collected during each of the threeexperimental regimes described previously. Statistical analysis failed to reveal any significantdifference between the sham and control groups for any of the breathing variables measured forall treatments. Therefore, the results obtained from the olfactory denervated frogs can beassumed to reflect the effect of removing afferent input from the olfactory receptors. Asillustrated by the physiograph recording in Fig. 3. 1B, the effect of the denervation procedureon the normocarbic breathing pattern was very subtle. The mean values of and fB1, of theolfactory denervated group were lower than those of the control group (Fig. 3.3). The meanvalues of fLabs, ‘Index VTlndex, and number of breaths per episode in the control and olfactorydenervated groups, however, were not statistically different while initially breathing air (Fig.3.2).Exposure to hypercarbia for 60 mm provoked a 565 % increase in fLabs which wassignificantly larger than the increase seen in control frogs (Fig. 3. 2A). Notice that, contrary tothe control group, did not change during hypercarbia in the olfactory denervated group(Fig. 3.3A). There was a 645% increase in the VIndex which, in absolute terms, was notsignificantly different from that seen in control animals (Fig. 3. 2D). Since the VTjfldex did notchange significantly (Fig. 3. 2C), this hyperventilation was produced by increases in fL alone.This was in part due to an increase in the number of breaths per episode recorded in this groupwhich was greater than that seen in the control group (Fig. 3. 2B).Unlike the control animals, the olfactory denervated frogs did not display a furtherincrease in ventilation during the post-hypercarbic period (Fig. 3.1 B). Breathing immediatelybegan to decrease towards control levels at the onset of recovery (Fig. 3. 1B, 2).Page 79Figure 3.3: The effects of changing CO2 levels on A) the instantaneous breathing frequency,and B) the instantaneous buccal oscillation frequency. See Fig. 3.2 legend for moredetails.Page 80Control• Pulmonary vagus sectioned• Olfactory nerve sectionedInstantaneous breathing Afrequency (min’)6o+t4O20o80Instantaneous buccal oscillationfrequency (min’)80BóOr*200Normocarbia Hypercarbia PostHypercarbiaNormocarbia Hypercarbia PostHypercarbiaPage 813. The role ofpulmonary receptors in the control of episodic breathing.The effects of the surgical procedure for exposing the pulmonary branches of the vagion breathing pattern was assessed for this series of experiments. The breathing pattern variablesmeasured under the three experimental regimes in the sham operated animals were similar tothose of the control group (Table 3.1). These data demonstrate that the breathing pattern ofvagotomized frogs was not affected by other aspects of the surgical procedure. Figure 3.1 Cillustrates the breathing pattern of a bullfrog after bilateral section of the pulmonary branch ofthe vagus nerve. The striking feature of this recording is the magnitude of the breaths. Thebuccal pump functioned more forcefully in these frogs than in any other group. Interestingly,removing afferent vagal feedback did not prevent the amplitude of the breaths from becomingprogressively larger during a breathing episode. In vagotomized frogs, the amplitude of twotypes of buccal movements, i.e. lung inflation and buccal oscillation, had a greater amplitudeand lung inflations occurred more frequently than in intact frogs. Data analysis confirmed that,while breathing air, the fLabs, VTjndex and Index of the vagotomized group were greater than thecorresponding values measured in the control frogs (Fig. 3.2). No other components ofbreathing pattern were affected by the bilateral vagotomy while breathing air.Bilateral vagotomy did not affect the changes in fL5 seen during hypercarbia in thecontrol group (38% decrease, Fig. 3.3A). Vagotomy also had no significant effect onunder normo- or hypercarbic conditions (Fig. 3.3B). Hypercarbia reduced fLabs by 52% butsimultaneously increased the VTIfldex by 78% (Figs. 3. 2A and 3. 2C respectively). These changesin the VTIfldex and fLabs counteracted each other and, thus, unlike the control and olfactorydenervated frogs, the vagotomized animals did not increase their Vifidex on exposure to 6% CO2(Fig. 3. 2D). The number of breaths per episode of breathing in vagotomized frogs wasPage 82increased significantly during hypercarbia (Fig. 3. 2B), and the reduction in breathing frequencywas due to longer non-ventilatory periods.The immediate transition back to breathing air did not affect the breathing pattern of thisgroup. This differs substantially from the response described for control frogs but was notsurprising given that the VIndex had not increased with hypercarbia. Thus, the fLabs, ttinst, VTIndexand 1Index did not change significantly during the immediate post-hypercarbic period invagotomized frogs (Figs. 3.2 and 3.3). None of the vagotomized frogs began to breathcontinuously at the transition and the number of breaths per episode returned directly tonormocarbic levels (Fig. 3. 2B).4. Kinetics of changes in airway and arterial blood CO2 levels.Figure 3 .4A shows that under normocarbic conditions each lung breath (indicated by thelarger increase in buccal pressure) was associated with a rise in CO2 in the buccal cavity.Hypercarbia initially reversed the direction of the CO2 oscillations associated with each lungventilation so that expiration now caused a reduction of buccal cavity CO2. while inspirationaugmented buccal cavity CO2 (Fig. 3 .4B). This trend lasted on average 15 mm; after whichbreathing no longer affected the CO2 level in the buccal cavity (Fig. 3. 4C).Figure 3.5 illustrates the recovery profile of selected variables for the first 2 mm of thepost-hypercarbic period. As the top panel illustrates, the milder hypercarbic stimulus used inthis series was equally capable of inducing the abrupt increase in breathing frequency observedin frogs exposed to 6% CO2 on return to breathing air. At the onset of the post-hypercarbicperiod, buccal cavity CO2 levels began to oscillate with each breath, decreasing duringinspiration and increasing during expiration. It took 60 sec for the gas phase of the chamber toreturn to pre-hypercarbic levels (Fig. 3.5) while roughly 5 mm were necessary for water PCO2Page 83(PwCO2)to return to normal levels. The CO2 measurements for the chamber and the buccalcavity could not be taken simultaneously and had to be obtained from separate runs. Thisexplains why the CO2 levels illustrated for the buccal cavity and chamber in Fig. 3.5 do notcorrelate perfectly during the irispiratory phase. After 30 mm of exposure to 4.5% CO2. frogswere acidotic; pHa and PaCO2 were significantly different from their respective normocarbicvalues (Table 2). Arterial pH and PaCO2 returned to normocarbic levels within 1.5 mm (Fig.3.5, Table 3.2). Arterial Po2 remained unchanged throughout the entire experiment (Table 3.2).Even though the in inspired air, the water in the chamber, as well as in arterial blood, hadall returned to pre-experimental levels within 5 mm, it took up to 30 mm for breathing to returnto pre-experimental levels and pattern.Page 84Figure 3.4: This figure illustrates the relationship between breathing activity (shown here bychanges in buccal pressure) and the resulting changes in CO2 levels in the buccal cavity(measured inside the buccal cavity near the glottis) during normocarbia (top two traces),after 5 mm (middle two traces) and 30 mm (lower two traces) of exposure to a 4.5 %CO2 in air gas mixture.Page 85Buccal pressure (cm H,O) Air5-4-3—2—Airway CO2 (%)1.00.8-0.6-0.4=0.2-0.0-Buccal pressure (cm H2O) 5 mm hypercarbia BA5-4-3—2-1—Airway CO2 (%)5.0 =4.8-Buccal pressure (cm H20 30 mm hypercarbia C5-.4 J u1IH s0Airway CO2 (%)5.0 :-4.8—4.6—-___4.4—4.2—4.0—0.0 0.5Time (mm)1.0 1.5 2.0Page 86Figure 3.5: A recording of the changes in breathing (buccal pressure; top panel), and thechanges in co in the buccal cavity (airways), the gas phase of the chamber (chamber),the water phase f the chamber (water), and in arterial blood (blood) that occurred at thetransition from hyper- to normocarbia. * indicates a value of Pa0 significantlydifferent (p < 0.05) from the pre-experimental (normocarbic) value whicl? are shown inTable 3.2. The buccal pressure, water and buccal cavity recordings were obtainedsimultaneously, while the chamber recording was2 taken from a separateexperimental run. Arterial values ar the means of 3 animals ± 1 SEM.Page 874.5% CO2 AirBuccal pressure(cm H20)54320Pco2 (Tori)35302520151050*WaterBloodAirwaysChamber-0.5 0.0 0.5 1.0 1.5Time (mm)2.0Page 88Table 3.1: The effects of sham surgery on the breathing pattern of bullfrogs breathingdifferent gases. Post-hypercarbia refers to the period immediately following the returnto breathing air.Page 89CONTROLSHAMOLFACTORYNERVESHAMVAGUSNERVESECTIONSECTIONAirHypercarbiaPostAirHypercarbiaPostAirHypercarbiaPostHypercarbiaHypercarbiaHypercarbiaVentilaijonindex(arbunits)141312+501++82314+405+++84251==598++±60±79±65±33±37±59±25±54±108Tidalvolumeindex915+11+8II+10+141616(arbunits)±1.9±1.7±1.9±1.6±1.0±0.8±±3.4±3.51.6Absolutebreathing1221=J=48++928+40++619+41++frequency(mm’)±3.4±4.6±3.7±2.5±2.2±3.2±4.4±4.4Instantaneousbreathing6635+49++6638+47+5936+46+++frequency(miii’)±6.1±4.6±3.4±6.3±2.0±1.0±±1.9±2.43.9Breathsperepisode1.58+41516+17+1.17+17±0.2±2.5±22.3±3.7±5.7±7.4±±2.3±7.8N=2N=40.1N=3Instantaneousbuccaloscillationfrequency(mm’)684349674548584144±7.1±5.6N=1±4.7±4.3±1.1±±7.3±4.64.2Table 3.2: Recovery profile of selected blood gas variables of bullfrogs after exposure to 4.5%CO2 for 30 mm. Values are expressed as means ± standard error. * indicates a valuesignificantly different from the corresponding air value at p < 0.05. N = 3, n = 6.Page 91AIR9±2.694±6.399±6.990±1.886±8.183±5.179±5.574±3.2Time0(mm)4.5%CO230POST-HYPERCARBIAPaCO2 (Torr)Pa02 (Torr)±7.54±7.85±7.92±7.89±7.97±7.97±*±1.1*14±1.7*15±2.8*12±0.510.9±0.810.9±0.4CDDISCUSSIONThe major focus of the present study was to examine the role of olfactory and pulmonaryC02-sensitive receptors in the initiationltermination of breathing episodes. Recent studies onanuran amphibians demonstrated that some minimal level of chemoreceptor drive was requiredfor the expression of any respiratory rhythm (Smatresk and Smits, 1991; chapter 2). However,they also demonstrated that phasic changes in chemoreceptor afferent input were not requiredfor the onset and cessation of breathing episodes in these animals (Smatresk and Smits, 1991,West et al., 1987; chapter 2). This raised the question “what turns breathing episodes on andoff?”. The present study confirms earlier studies showing that input from CO2-sensitiveolfactory and pulmonary airway receptors affect the incidence and length of breathing episodes,but leads to the conclusion that neither of these inputs is responsible for turning breathing on oroff per se either.1. The role of ofactoiy receptors in the control of episodic breathing.The breathing pattern and the response to hypercarbia of the control frogs were similarto those obtained in other studies (Jones and Chu, 1988, Maclntyre and Toews, 1976; Smatreskand Smits, 1991; West et al., 1987; chapter 2). What has not been reported in other studies onanuran amphibians is the increase in breathing frequency seen when the CO2 was removed fromthe inhaled gas mixture. This increase was extremely large in the present study; equal to thatof the initial response to hypercarbia. What was most surprising is that it occurred at a timewhen the respiratory drive would be expected to be decreasing and ventilation would be expectedto fall. This increase was similar to increases which have been reported in several species oflizard and snake (Glass and Johansen, 1976; Nielsen, 1961).Page 93There were no significant effects of removing olfactory receptor input on breathingpattern and overall ventilatory output in normocarbic frogs. In olfactory denervated animals,however, the peak ventilatory response was observed during the hypercarbic period; there wasno further increase in ventilation at the immediate transition from breathing 6% CO2 to breathingair. Furthermore, the ventilatory output and breathing pattern of the olfactory denervated frogsrecorded during hypercarbia were not significantly different from those of control animals duringthe post-hypercarbic period. These observations suggest that, as in lizards, stimulation ofolfactory receptors by high tonic airway CO2 levels restricts the increase in breaths per episodeand fLabs during hypercarbia in intact animals (Ballam and Coates, 1989; Coates et a!., 1990) andthat removal of this inhibition in intact animals at the transition from hypercarbic to normocarbicconditions somehow accounts for the post-hypercarbic hyperpnea.It is possible that the vomeronasal nerves were also sectioned by the olfactory transection,so that some of the effects observed after the denervation could have been due to cutting thevomeronasal nerves. However, because the vomeronasal organ detects compounds that areusually high-molecular-weight, nonvolatile and contain proteins, whereas the olfactory stimuliare primarily volatile, low-molecular-weight compounds (Coates and Ballam, 1989), it is veryunlikely that the vomeronasal organ was involved in the changes in breathing pattern producedby tonic CO2 levels in the upper airway.The absence of any increase in the VTTfldex during hypercarbia is another interestingpeculiarity of the olfactory denervated group. It suggests that stimulation of olfactory receptorsby elevated levels of CO2 mediates, at least partially, the increase in the VTJfld5X. Thus the VTIfldexreturned to normocarbic levels during the post-hypercarbic period in the control group despitethe post-hypercarbic hyperpnea.Page 94Although it is clear that olfactory receptors are responsible for the production of the-posthypercarbic hyperpnea illustrated here, it is also clear that modulation of the action of thesereceptors is not responsible for the production of intermittent lung ventilation. The fact thatbreathing was never continuous in olfactory denervated frogs during hypercarbia indicates that,although olfactory receptor afferent input affects breathing pattern significantly, it is not thereason that breathing remains episodic during hypercarbic exposure.2. The role ofpulmonary receptors in the control of intermittent breathing.Most investigations report that vagotomy in air breathing vertebrates produces a reductionin breathing frequency and an increase in VT stemming from a prolonged respiratory effort. Ofnote is the fact that this is quite different from the effect of reducing lung volume which isgenerally an increase in breathing frequency (for reviews, see Milsom, 1990b; Shelton et al.,1986). The literature documenting the effects of vagotomy on the breathing pattern ofamphibians, however, is rather scarce. In Xenopus laevis, vagotomy had no effect on VT butaugmented fLai,s (Evans and Shelton, 1984) while in the present study on Rana catesbeiana,vagotomy produced a significantly greater VTIfldex and fLabs. It would appear, therefore, thatunlike the situation in most other air breathing vertebrates, vagotomy in frogs does mimic theeffects of reducing lung volume.It is not clear why an increase was seen in the VTIfldex in Rana catesbeiana but notXenopus laevis, following vagotomy. The elevated VTlndex in Rana was not due to prolongedinspiration since the mean value of fL1 was unaffected by vagotomy suggesting that it was dueto a more powerful use of the buccal pump. In the present study, vagotomy also increasedbuccal oscillation amplitude, indicating that the inverse relationship that exists betweenPage 95pulmonary stretch receptor discharge and buccal pumping amplitude for gill ventilation intadpoles (West and Burggren, 1983) is maintained in adult frogs.There was one other difference between the effects of vagotomy on breathing in Xenopusand Rana; in Xenopus, vagotomy augmented fLabs by increasing the number of breaths perepisode (Evans and Shelton, 1984) while in Rana, vagotomy augmented fLthS by decreasing thenon-ventilatory pause; the number of breaths in each breathing episode was not affected by thebilateral vagotomy. The reasons for this difference are also not clear.Exposing the vagotomized frogs to hypercarbia led to further increases in the VTjfldex, adecrease in fLabs, and, consequently, no change in the Vj,ex. Despite the decrease in fL5, therewas a significant increase in the number of breaths in each breathing episode, but not as largean increase as that seen in control and sham animals. This would suggest that duringhypercarbia, pulmonary vagal input contributes to the increase in fLabs by both reducing the nonventilatory pause and further increasing the number of breaths per episode. The intriguingobservation here is that pulmonary vagal feedback appears to inhibit fLabs during normocarbia,by prolonging the non-ventilatory period, but stimulates fLabs under hypercarbic conditions, byshortening the-non-ventilatory period and augmenting the number of breaths per episode.The effect of vagotomy on the number of breaths in each breathing episode inhypercarbic frogs is similar to fmdings reported for other episodically breathing species (Naifehet al. 1971a, b; for review, see Milsom, 1990b; Shelton et al., 1986). In reptiles, vagotomytransforms the breathing pattern from one of episodes with many breaths to one of evenly spacedsingle breaths or small episodes with two or three breaths (Naifeh et al., 197 la, b; Vitalis andMilsom, 1986). The latter is the pattern normally observed in normocapnic bullfrogs and thiswas unaffected by vagotomy. Thus, although the afferent signal from the pulmonary receptorsin Rana catesbeiana exerts an inhibitory effect on the respiratory control centres that govern airPage 96flow rates, and influences the duration of both the ventilatory and non-ventilatory periods, thedata from the present study does not support the proposition that vagal feedback directly causesinitiation/termination of breathing episodes per Se. This conclusion is consistent with the factthat the respiratory related motor output of in vitro preparations of turtle brainstem-spinal cordconsists of small episodes of fictive breathing that occur in the absence of any peripheralfeedback (Douse and Mitchell, 1990). This suggests that episodic breathing is an endogenousproperty of the central nervous system. Unfortunately, very little is known about the neuralmechanisms underlying the spatio-temporal distribution of the breaths in episodic breathers. Itis conceivable that, as suggested by Jackson (1978), that the fundamental output unit of therespiratory controller is the episode, rather than the breath.3. Kinetics of changes in airway and arterial blood CO2 levels.As already pointed out, our observations suggested that stimulation of olfactory receptorsby high airway CO2 levels restricted the increase in breaths per episode and fLabs duringhypercarbia in intact animals and that removal of this inhibition at the transition fromhypercarbic to normocarbic conditions somehow accounted for the post-hypercarbic hyperpnea.To try to elucidate how this may have occurred we monitored buccal cavity and arterial bloodlevels of CO2 throughout the experimental protocol in three animals. We noted in the Methodssection, that the system used to quantify CO2 levels in the airways was not perfect. Because theairflow patterns and breathing mechanics of the bullfrog are very complex, the magnitude of thechanges in CO2 levels in the proximity of the CO2 sensitive receptors investigated here willundoubtedly have been somewhat different from those we obtained from our sampling site nearthe glottis. Our results do, however, indicate the rough magnitude and direction of changes inCO2 levels in the airways during the different phases of the experimental protocol. The resultsPage 97indicate that CO2 levels in the airways 1) change (oscillate) with the different phases of therespiratory cycle when frogs are breathing room air (normocarbia), 2) were constant duringhypercarbia, 3) began to oscillate again immediately when a normocarbic gas mixture wassubstituted for the hypercarbic gas mixture and 4) fell faster than PaCO2 during the posthypercarbic phase. They also indicate that PwCO2 in the chamber decreased slowly during thepost-hypercarbic phase, more slowly even than PaCO2 indicating that CO2 could not have beenlost from the body across the skin to the water during these experiments.These findings lead to at least two possible conclusions vis-à-vis the role of olfactoryreceptors in producing the post-hypercarbic hyperpnea. One is that the olfactory receptorsresponding to increased airway CO2 levels partially inhibited the full increase in whichwould otherwise have been associated with the rise in Paco2 alone. In this scenario, the posthypercarbic hyperpnea stemmed from the fact that this inhibition was removed faster than thestimulation of the arterial and central chemoreceptors during the recovery period. This isconsistent with the observation in lizards that increasing levels of tonic stimulation of theolfactory receptors lead to increasing levels of inhibition of the ventilatory response tohypercarbia. There is also evidence from more recent work on lizards that suggested that whiletonic stimulation of olfactory receptors with elevated levels of CO2 inhibited breathing, phasicstimulation of the olfactory receptors did not affect breathing frequency (Ballam and Coates,1989), regardless of which specific phase of the respiratory cycle the peak of the CO2 oscillationoccurred (Coates et a!., 1990). Thus, an alternative conclusion is that the post-hypercarbichyperpnea occurred simply because stimulation of the olfactory receptors became phasicallymodulated rather than tonically modulated. These possibilities are not mutually exclusive andboth may have contributed to the post-hypercarbic hyperpnea as well as other less obviousmechanisms.Page 98The measures of buccal cavity and arterial blood CO2 levels obtained in the present studygive rise to another interesting point. On the transition from hypercarbia to normocarbia, CO2levels had returned to normocarbic values in the gas phase of the chamber in 1 mm, in arterialblood in 1.5 mm and in the water phase of the chamber in 5 miii. Despite this rapid return tonormocarbia, breathing remained significantly elevated above normocarbic levels for up to 30mm. While it is possible that this reflects a longer term disturbance of central acid-base statein these animals, it is also possible that this represents a form of long term facilitation ofbreathing arising from the hypercarbic stimulation (Eldridge and Milthorn, 1986).4. The significance of buccal oscillation and lung ventilation frequencies.Although respiratory events occur intermittently in bullfrogs, there is an intrinsic rhythmto buccal oscillations as well as breaths occurring in a breathing episode. In the previouschapter, it was observed that these intrinsic rhythms were the same, both during normoxicnormocarbia (air) and under conditions of increased respiratory drive. This led to the suggestionof two possible scenarios. The first was that there is a single rhythm generator whose input isintegrated with other inputs at two distinct pattern generators. At low levels of respiratorydrive, only output from the pattern generator driving buccal oscillation is produced. Asrespiratory drive increases, the rhythm is more often expressed as output from the patterngenerator driving lung inflation. The other possibility is that there are two distinct rhythmgenerators which are entrained under most circumstances with expression of lung rhythm beingconditional upon a higher level of central and/or peripheral receptor input. Although the currentstudy can not resolve this issue further, it is important to note that and fB were virtuallyidentical under normocarbic conditions and changed in similar ways under all three regimes inthe control, olfactory denervated and pulmonary denervated groups. If the assumption thatPage 99instantaneous frequencies are reliable indicators of the endogenous respiratory rhythm inintermittent breathers is valid, then the data indicate a number of interesting things. Theyindicate that instantaneous frequency decreases during hypercarbia even though absolutefrequencies increase (at least for lung ventilation) in control animals. This indicates that eachbreath takes longer, which might be associated with the increased tidal volume index, but theperiod between breathing episodes is greatly shortened, more than making up for the prolongedbreaths. It also indicates that the olfactory receptors, but not the pulmonary receptors, exert astimulating action on respiratory rhythm in normocarbia and an inhibitory action on respiratoryrhythm during hypercarbia.5. Physiological significance.It has been suggested that reflex inhibition of respiration by high tonic airway CO2 inreptiles is a defense mechanism provided by olfactory receptors which is advantageous toamphibians and reptiles (Coates and Ballam, 1990). It can restrict futile breathing activity underhypercarbic conditions, when CO2 excretion is compromised. Whatever the physiologicalsignificance of the reflex, the present study demonstrates that a similar reflex is present in ananuran amphibian and further reinforces the observation that hypercarbia does not equalhypercapnia in studies on the control of breathing in lower vertebrates.It has now been shown that although arterial chemoreceptor, olfactory receptor andpulmonary receptor feedback all shape breathing pattern, none of these inputs is solelyresponsible for turning episodic breathing on or off. This conclusion is consistent with the factthat the respiratory related motor output of in vitro preparations of turtle and frog brainstemspinal cord consists of small episodes of fictive breathing that occur in the absence of anyperipheral feedback (Evans and Shelton, 1984; Maclntyre and Toews, 1976). This clearlyPage 100suggests that episodic breathing is an endogenous property of the central nervous system. Fromthis, it can be concluded that the breathing pattern observed during normocarbia either reflectsthe basic rhythm generated by the central nervous system or that there are redundant systemswhich somehow control the onset/termination of breathing episodes.Page 101CHAPTER 4:AN OPEN-LOOP ASSESSMENT OF THE ROLE OF VAGAL AFFERENT INPUT TOTHE CONTROL OF EPISODIC BREATHING.Page 102INTRODUCTIONIn the previous chapter, the role of vagal afferent feedback in the control of episodicbreathing was assessed by sectioning the pulmonary branch of the vagus nerve. Although thisprocedure lead to an increase in the amplitude and frequency of resting ventilation, indicatingthat pulmonary stretch receptor (PSR) feedback modulates breathing pattern, the data alsoindicates that it is not responsible for the onset/termination of the breathing episodes. Vagotomyalso abolished the increase in breathing frequency that anuran amphibians usually exhibit duringhypercarbia (chapters 2 and 3; Branco et al. 1992; Smatresk and Smits, 1991; West et al., 1987),suggesting that pulmonary vagal input is necessary for the production of normal respiratorychemoreflexes.Because vagal afferent traffic is very complex, however, it was impossible to determinewhether the ventilatory changes observed post-vagotomy actually reflected the absence of PSRfeedback specifically, or the removal of some other sensory feedback. A survey of the literaturereveals the importance, as well as the complexity of vagal feedback in respiratory control. Italso reveals that the data obtained from studies employing vagotomy must be interpretedcarefully. For instance, elimination of all vagal feedback by vagotomy mimics the classicHering-Breuer reflex in most vertebrates; that is a reduction in breathing frequency, an increasein inspiratory duration, and hence tidal volume, and a shortening of expiration. These changescontrast with the effects of reducing PSR input by deflating the lungs which increases breathingfrequency without affecting inspiratory duration or tidal volume; i.e. it decreases the expiratoryduration (Bartoli et al., 1973a, b; see review by Milsom, 1990b). Lung deflation removes bothtonic and phasic input arising from PSR. Studies on the timing of respiratory events have shownthat tonic PSR input primarily controls respiratory frequency through an effect on expiratoryduration (TE) (Bartoli et al., 1973a, b), but fails to modulate inspiratory duration (TI); whereasPage 103studies in which PSR feedback changed phasically with each breathing cycle showed an inverserelationship between TI and VT (Clark and von Euler, 1972; Cross et al., 1980; Feldman andGautier, 1976).The sum of this work argues that the nature of the PSR feedback (i.e. phasic vs tonicvagal feedback) must be considered in a proper investigation of its role in respiratory control.Furthermore, although deflation of the lungs reduces, but does not eliminate all, PSR and othervagal inputs, the breathing pattern observed upon lung deflation differs markedly from thebreathing pattern observed post-vagotomy. With this in mind, the present study furtherinvestigated the role of vagal afferent feedback in the control of breathing pattern using adecerebrate, paralysed, unidirectionally ventilated frog preparation similar to the one developedby Kogo and co-workers (1994). This approach made it possible to manipulate lung pressure,and PSR feedback, directly and relatively independent of other vagal afferent inputs that affectbreathing. It compared the fictive breathing pattern of bullfrogs exposed to phasic and tonicchanges in lung volume and compared the effects of vagotomy to that of deflating the lungs.This preparation provides great control over respiratory drive, so that it was also possible toopen the chemoreceptor feedback loop. Since chemoreceptor and pulmonary vagal afferents areknown to interact to modulate ventilatory control in mammals (see review by Mitchell et a!.,1990) these experiments were performed under different levels of respiratory drive so that it wasalso possible to determine whether PSR and chemoreceptor afferent inputs interact to modulatethe breathing pattern in bullfrogs.Page 104MATERIALS AND METHODSThe experiments were performed on 13 adult bullfrogs of either sex weighing between269 and 393g (mean weight 309 ± 10.6g). The animals were fed and maintained according tothe protocol described in chapter 2.Animal preparation.The reduced frog preparation used in this study was similar to the one developed byKogo et at. (1994). The frogs were anesthetized by immersion in a mixture of cold water andcrushed ice bubbled with 02 for at least an hour or until the toe pinch reflex was abolished(Stehouwer, 1987). The animal was then moved to the surgery table, and a small hole was madeon the dorsal surface of the skull with a dental drill, thus exposing the telencephalon. Fine boneshears were used to enlarge the hole caudally until the optic tectum was reached. The brain wastransected between the optic tectum and the rostral forebrain with a blunt spatula and the entireforebrain was aspirated with a suction device. The decerebration procedure took usually lessthan 5 mm. Bleeding was controlled by cautery, absorbable haemostat, and by filling the cranialcavity with cotton pellets. The frogs were then paralysed by a pancuronium bromide injectionin the dorsal lymph sacs (Pavulon, 0.50mg100g’, 2mgm1’).The left femoral artery was cannulated occlusively with polyethylene tubing (PE 50) tomonitor blood pressure. Both lungs were then cannulated for unidirectional ventilation (UDV)according to the method described in chapter 2. The frog was then moved on to its back andthe skin covering the mandible was cut and retracted. The mandibular branch of one trigeminalnerve was located and isolated. A piece of 5-0 silk was placed around the nerve before it wascut distally. The glottis was sealed with two wound clips to prevent gas leaking from theartificially ventilated lungs. The tympanic membranes were cut on both sides before the animalPage 105was placed on its stomach and its head was positioned (dorsal surface facing upward) in astereotaxic holder. The vagus nerve was located on one side from a dorsal approach. Anincision was made above the scapula, the muscles were cut and the scapula retracted to isolatethe laryngeal branch of the vagus nerve. A piece of 5-0 silk was placed around the nerve beforeit was cut. A custom made water bath bubbled with gas was positioned under the animal, andwet towelling was draped over the frog with both ends in the bath acting as a wick to keep thefrog’s skin moist.In one group of frogs (N = 7), the vagus nerves were located bilaterally at the cervicallevel and isolated by placing a loose suture around each nerve for subsequent vagotomy. Thenerves were kept moist by placing a small cotton pellet soaked with Ringer solution next to eachnerve before closing the skin openings.Efferent nerve activity to the respiratory muscles was recorded from the transectedmandibular branch of the trigeminal (Vm) nerve and the laryngeal branch of the vagus nerve(Xl) (Sakakibara, 1 984a; Kogo et al. 1994). The mandibular branch of the trigeminal nerveinnervates the elevator muscles of the buccal cavity, whereas the laryngeal branch of the vagusnerve innervates the opening and closing muscles of the glottis. As described in chapter 1, frogsdo not use an aspiration pump but instead use a buccal force pump to ventilate their lungs. Thispump produces two different events. Small pumping motions produced with a closed glottistidally ventilate only the buccal cavity (buccal oscillations). Larger pumping motions producedwith an opened glottis ventilate the lungs (lung ventilations). By recording from both Vm andXl, fictive buccal oscillations could be discriminated from fictive lung ventilations. Smallamplitude activity in Vm without activity in Xl is interpreted to represent attempts at expulsionof air from the buccal cavity (buccal elevation) to the atmosphere. Activity in Vm with activityPage 106in Xl is interpreted to represent attempts at pumping air from the buccal cavity through an openglottis into the lungs (lung inflation).&perimental procedures.Once the surgery was completed, the frogs were unidirectionally ventilated with air ata rate of 250 mlmin’ and 75 mlmin’ in series I and II, respectively, regulated by a flow meter.The high flow rate in series I was selected to ensure rapid refilling of the lungs after they weredeflated in the phasic feedback experiments (see below). Prior to delivery to the lungs, theinflowing gas was humidified by bubbling through an Erlenmeyer flask half filled with water.The lung pressure was monitored by installing a T-piece, connected to a pressure transducer,in the inflow cannula. The isolated branches of the trigeminal and vagus nerves were eachpositioned on bipolar platinum hook electrodes and covered with a 1:1 mixture of vaseline andmineral oil. Electrical nerve activity recorded from the bipolar electrodes was amplified [filtersettings: 200 Hz (high pass) and 10 kHz (low pass)], full wave rectified and integrated (Gould)in 67-ms intervals. The raw and integrated nerve signals were viewed on an oscilloscope andstored on a polygraph recorder and on a computer disk with a data acquisition system (ACCodas). The sampling rate of the analog-to-digital conversion was 2500 Hz. The animal wasallowed to recover from cold anaesthesia for at least 30 mm before the onset of experiments.Henderson and Burtsaert physiological saline (in mM: 118 NaC1, 2.5 CaC12, 4.7 KC1, 1.1KH2PO4,24 NaHCO3, 1.2 MgSO4, 4.5 glucose) or adrenaline (0.05 mgml’ in physiologicalsaline) was injected into the arterial cannula when mean arterial blood pressure dropped below25cm of H20.Page 107Experimental protocol.:Series I: The effects of yhasic vs tonic PSR feedback.This series of experiments was performed on 6 bullfrogs, and compared the effects ofphasic vagal volume-related feedback on “fictive breathing pattern”, as indicated by theelectroneurograms (ENG), to that of a tonic level of lung pressure equal to the peak of thephasic pressure swing. The phasic changes in vagal feedback were produced by deflating andinflating the lungs with each attempted breath (fictive). The changes in lung pressure wereproduced by connecting the UDV outflow catheter to a solenoid valve triggered by a peakdetector. The resting lung pressure was set to equal 4 cm of H20 by submerging the UDVoutflow catheter to the appropriate level in a water colunm to control the outflow resistance.When the integrated vagus nerve discharge rose 4.5 tVs above its baseline, indicating increasedactivity to the muscles opening the glottis, the solenoid valve opened, and suction was appliedto the UDV outflow to rapidly deflate the lungs. The solenoid valve was deactivated whenintegrated Xl discharge fell below the threshold value, and the lungs reinflated to 4 cm of H20.Typically, the lung pressure at the end of the deflation was 2 cm of H20. The fictive breathingpattern recorded was then compared to that recorded previously for frogs in which lung pressurewas maintained constant at 5 cm H20. To determine whether the pattern of vagal feedbackaffected the frog’s responsiveness to hypercarbia, the animals were exposed to three levels ofhypercarbia which were set by changing the composition of the gas mixture in the UDV line.These mixtures were air, 1.7 ± 0.1% CO2 in air and 3.3 ± 0.2% CO2 in air. They wereadministered randomly, and a 25 mm equilibration period was allowed between each run witha new mixture. Once a level of respiratory drive was established, the animals were exposed tothe different patterns of vagal feedback.Page 108Series II: The effects of changing the level of tonic PSR feedback.This series of experiments assessed the effects of changes in tonic PSR feedback underdifferent levels of respiratory drive. The experiments involved two groups of bullfrogs; in thehypoxic group (N = 7), the animals were UDV with a hypoxic gas mixtures (6.0 ± 0.3% 02in air), while in the second group (N = 6), the animals were UDV with gas mixture containingthree different levels of CO2 (air, 1.7 ± 0.1%, and 3.3 ± 0.2% CO2 in air). The gas mixtureswere administered at random, and according to the protocol described in the previous series.Once respiratory drive was established, PSR feedback was set at different levels by immersingthe UDV outflow catheter in a water cylinder to change the outflow resistance of the UDV line,and thus, increase or decrease lung pressure accordingly. Pulmonary pressure was set to either0, 2 or 5 cm of H20 at random.Series III: Comparison between vagotomy and tonic vagal feedback.This series of experiments was performed on 7 frogs to determine if vagotomy and toniclung deflation (PL = 0 cm 1120) had similar effects on the fictive pattern of breathing. Bullfrogswere exposed to a moderate respiratory drive (1.7 ± 0.04% C02) and the fictive breathingpattern was recorded at the different levels of tonic lung pressure described previously beforeand after the vagus nerve was cut bilaterally at the cervical level. The fact that changing lungpressure after the bilateral vagotomy had no effect on any of the variables confirmed that thevagotomy was successful at removing mechano-receptor feedback.Analytical procedures.The gas mixtures administered to the animals were produced by mixing air with CO2 orN2 with flow meters. The gas mixture fed to the UDV line and water bath was sampledPage 109continuously from the outfiowing gas line and monitored with a Beckman LB-2 CO2 analyzerand OM-1 1 oxygen analyzer calibrated with gas mixtures produced by a Radiometer GMA2precision gas supplier. Both pressure transducers were calibrated against a static water column.Because paralysed frogs could not spontaneously generate air flow to their lungs, manyrespiratory variables could not be measured directly. Thus, the duration of the breathing cycle(TT0T) and inspiratory duration (Ti) were estimated by measuring burst duration from the Xland Vm, respectively. These assessments are based on the fact that a) pulmonary ventilation infrogs can only occur while the glottis is opened and b) lung inflation (inspiratory phase) isproduced by rapid elevation of the floor of the buccal cavity when the glottis is opened (Westand Jones, 1975; Sakakibara, 1984a), as previously demonstrated by others (Kogo et at., 1994).For similar reasons, tidal volume could not be measured directly. Instead, the peak ofthe full wave rectified and integrated trigeminal ENG was used as an index of tidal volume.Sakakibara (1984a) has correlated the pattern of respiratory nerve activity with the changes inbuccal and lung pressure that occur during the breathing cycle in the bullfrog. His studydemonstrated that the compound trigeminal ENG recorded just proximal to the site of the nervebranching off to the respiratory muscles (the recording site of the present study) had activitycorresponding to simultaneous elevation of buccal and lung pressures. This author subsequentlydemonstrated that, in the glottis occluded frog, the integrated peak trigeminal activity waslinearly related with peak buccal pressure (Sakakibara, 1984b), validating the use of peaktrigeminal nerve activity as an index of total inspiratory activity. This measure compares withthe use of peak integrated phrenic nerve activity as a correlate of tidal volume in mammals(Eldridge, 1971; 1975). Consequently, multiplying the peak integrated trigeminal activity(VTindex) by breathing frequency (fL) yielded an index of total ventilation (inciex) reported herein arbitrary units.Page 110Breathing episodes are designated on a subjective basis. The criteria and methods usedto quantify the number of breaths within an episode, breathing frequencies (absolute andinstantaneous) and instantaneous buccal oscillation frequency were described in chapter 2. Notethat when breathing was continuous, no values were obtained for breaths per episode or episodesper minute. Recordings of respiratory variables were obtained only when the fictive breathingpattern was stable; this usually occurred within the first 1 minute following a static change inlung pressure.Data analysis.Values for fictive breathing variables were obtained by analyzing 2-mm of data. All dataare presented as means ± SE. The results were statistically analyzed using a two-way analysisof variance followed by a Student-Newman-Keuls test (P <0.05).Page 111Figure 4.1: Recordings of integrated trigeminal nerve activity ( V) illustrating the changes inthe fictive breathing pattern associated with increasing lung pressure from 0 to 2 cm H20(bottom trace) at each of four levels of respiratory drive, indicated to the right hand side.Page 112fvAmphibianinsilupreparalion-----kiI-IIiIIIt._-——-Ilill...alfiftIIliL......ili...jlfllihIh—fthlililluilliIIIiImill111thES[IIti111iiFco=0.0003Fco=0.017Fco=0.033F0=0.06ct101ccRESULTSThe fictive breathing pattern recorded from the decerebrate, paralysed andunidirectionally ventilated frog preparation was very similar to the one recorded from intactbullfrogs (see chapters 1, 2 and 3). The pattern of activity was episodic, and the number offictive breaths in each bout of activity varied according to the level of respiratory drive, as itis described below (Fig. 4.1). In some preparations, smaller amplitude bursts of activity wererecorded from the trigeminal nerve between larger amplitude bursts of activity (Fig. 4.1). Theseoccurred very rhythmically, and comparable activity could rarely be recorded from the vagusnerve. These small rhythmic bursts appear to represent the small-amplitude buccal oscillationsdescribed previously for intact frogs (see chapters 2 and 3).1: The effects ofphasic vs tonic vagal feedback.Figure 4.2 compares the fictive breathing pattern of a bullfrog with phasic vagal feedback(top panel) to the pattern recorded when PL was maintained constant at a level comparable tothe peak phasic pressure (bottom panel; tonic vagal feedback). The onset of a fictive breathtriggered a sharp decrease in PL from 4 to 2 cm 1120. Once the suction valve was shut at theend of the vagal discharge, PL returned to its initial level before the onset of the next breath.For both patterns of vagal feedback, increasing hypercarbic respiratory drive stimulatedbreathing (p < 0.001) (‘index; Fig. 4.3A), owing mainly to increases in fLabs (p < 0.001) (Fig.4.3B) since the VTjfldex was not affected by hypercarbia (p 0.837) (Fig. 4.3C). Hypercarbiacaused substantial changes in breathing pattern which accounted for the increase in fLabsIncreasing the F2 of the UDV prolonged the duration of each fictive bout of breathing sinceeach of them contained more breaths (p < 0.001) (Fig. 4. 3D). Increasing hypercarbicrespiratory drive also increased the number of episodes per minute (p < 0.001), thus shorteningPage 114the duration of the non-ventilatory pause (Fig. 4.3E). At the highest level of however,breathing became continuous. For both, phasic and tonic vagal feedback, there was an inverserelationship betweenF2 and trigeminal and vagal burst duration (p = 0.009 and p = 0.042,respectively), such that hypercarbia shortened the duration of each fictive breath (Fig. 4.4).Since hypercarbia shortened trigeminal and vagal burst durations equally, the relative timing ofthe inspiration and expiration was unaltered by hypercarbia (p = 0.263). From the ENGrecordings (Fig. 4.2) and the summarized data (Fig. 4.3), it is clear that phasic and tonic vagalfeedback had identical effects on the fictive breathing pattern in this preparation, as there wereno significant differences between either group (phasic vs tonic vagal feedback) for any of thebreathing pattern variables measured (Figs. 4.3, 4.4 and 4.5). The rest of the data presentationwill therefore focus on the effects tonic lung inflation on breathing pattern.2. The effects of changing tonic vagal feedback on breathing pattern.The ENG recording shown in Fig. 4.1 illustrates the changes in the fictive breathingpattern that were typically observed when pulmonary pressure (PL) was tonically increased from0 cm H20 (left hand side) to 2 cm H20 (right hand side) under different levels of respiratorydrive. Overall, raising PL stimulated breathing (p < 0.001) (‘‘index; Fig. 4. 5A), owing toincreases in fLabs (p < 0.001) (Fig. 4.5B); raising PL decreased the VTjfldex (p < 0.001) (Fig.4.5C). The increase in fLabs observed upon static lung inflation was due a greater number ofbreathing episodes per minute (p = 0.033) (Fig. 4.5E). The effects of PL on fLabs, and thusVindex, were directly proportional to the respiratory drive (p < 0.001). This explains why theeffects of increasing lung pressure on Index, fLabs, breaths per episode and episodes per minutehad no significant on animals on UDV with air or hypoxia (P > 0.05 for all) (Fig. 4.6). Thenumber of breaths in a breathing episode was not significantly affected by tonic changes in PLPage 115(p = 0.383) (Fig. 4.5D). At the higher levels of hypercarbia, increasing PL made the frogsbreathe continuously (1.7% CO2 in air in 11 frogs, 3.3% CO2 in air in all 13 frogs), so that thebreathing pattern consisted of a single continuous episode (Fig. 4.1, 4.5D and E). A two-wayANOVA confirmed that hypercarbic respiratory drive and PL did interact to modulate thenumber of episodes per minute (p = 0.002), fL and ‘ex (p < 0.001 for both) (fig. 4.5).Figure 4.7A shows that the effects of changing PL on breathing described previouslywere reversible. In this particular recording, the frog was unidirectionally ventilated with a3.3% CO2 in air gas mixture, and reducing PL from 2 to 0 cm H20 changed the fictive breathingpattern from continuous to evenly-spaced single breaths; fictive breathing frequency wassignificantly reduced (p < 0.001). As predicted from the previous description of therelationship between static PL and burst amplitude (Fig 4.5C; 4.6C), decreasing tonic PSRfeedback augmented the burst amplitude of both, trigeminal and vagal ENG. The effects ofreducing PL were, however, more pronounced in the trigeminal than in the vagal ENG (Fig.4. 3B).The two fictive breaths presented in the lower panel of Fig 4. 7B demonstrate that vagalafferent feedback is also an important modulator of the timing of the fictive breaths. Thechanges in timing are caused by shortening trigeminal burst duration only (p = 0.023), sincechanging vagal feedback had no effect on vagal burst duration (p = 0.099) (Fig. 4.8). Asdepicted in Fig. 4. 7B, when lung pressure was greater than 0 cm H2O, in a typical fictive breaththe vagal discharge increased first, while the trigeminal ENG remained at baseline levels.Trigeminal activity then increased some time later so that, for a significant portion of the cycle,both nerves fired simultaneously. The fictive breath ended when both nerves became silent atroughly the same time. In non-paralysed frogs, these events would correspond to glottal openingto allow lung deflation, followed by activation of the buccal levators to reinflate the lungs byPage 116reversing the direction of gas flow at the opened glottis. Reduction of vagal feedback (Fig.4. 6B) prolonged the duration of trigeminal activity, to the point that the increase in dischargebegan at the same time, and in some cases before, any increase in vagal activity was recorded.As predicted from the inverse relationship between the TI/TT0T ratio and PL (p = 0.001; datanot shown), in intact frogs, this effect of reducing vagal afferent traffic would translate into areduction or complete elimination of the expiratory phase and a concomitant prolongation of theinspiratory phase as the animal attempted to reestablish a normal lung volume.Both vagal and trigeminal burst durations were shortened by increasing F2 (p = 0.009and p = 0.042, respectively), but were prolonged by reducing F02 (p = 0.006 and p = 0.002,respectively) (Figs. 4 .4B; 4.8). The relative changes in burst duration (TI/TT0T) that favourlung inflation with decreasing PL (lung deflation with increasing PL) were not significantlyaffected by changes in respiratory drive (p = 0.077).Neither changes in respiratory drive nor PL had a significant effect on either buccal orlung instantaneous frequencies (Fig. 4.9).Page 117Figure 4.2: A comparison between the effects of phasic (panel A) and tonic (panel B) vagalafferent feedback on the fictive breathing pattern of bullfrogs unidirectionally ventilatedwith a 1.7% CO2 in air gas mixture.Page 118AI V - -—_____fx rLMJ______Lung pressure(cm H.O)w wBI V ___L..b_.k____k_.%..%.b,l..4___?_ -fxLung pressure(cm H20)a —10 secPage 119Figure 4.3: A comparison of the effects of phasic (4-2-4 cm H20; open symbols) and tonic (5cm H20; closed symbols) vagal feedback on A) ventilatory index (‘Index), B) absolutebreathing frequency (fLabs), C) peak integrated trigeminal nerve activity (VTIfldSX), D)breaths per episode and E) episodes per minute at different levels of hypercarbicrespiratory drive.Page 1202PeakIrigeminalactivity(MV0IL*sec)Cl.00L....__-_,=_..._C__-.--.- 20 I0 00.8:zz1t0. 40 30 20 10 6 5 4 3 2 0—..1.j_.-.-__..0.000010020. 8 6 4F0Figure 4.4: A comparison of the effects of phasic (4-2-4 cm H20; open symbols) and tonic (5cm H20; closed symbols) changes in lung pressure on A) trigemmal and B) vagal burstduration at different levels of hypercarbic respiratory drive.Page 122Trigeminal burst Aduration (sec)1.00.8 —0.00 0.01 0.02 0.03 0.04Vagal burstduration (see) Ba.0F -0.4 L0.2-0.00 0.01 0.02 0.03 0.04Fc0Page 123Figure 4.5: The effects of tonic changes in lung pressure on A) ventilatory index (Vindex), B)absolute breathing frequency (fL), C) peak integrated trigeminal nerve activity (VTIfldex),D) breaths per episode and E) episodes per minute. Each curve represent a differentlevel of hypercarbic respiratory drive (• air; D 1.7% C02; • 3.3% CO).Page 124Absolutebreathingfrequency(mm1)A40 30 20I2345Lungpressure(cm1120)Ventilatoryindex(arbunits)30 20BCPeakintegratedtrigeminalburstactivity(VoIt*sec)2.0 1.5 10 0.50.00I2345I0 00I2340I2345BreathsperepisodeDEpisodesperminuteE81061846Lz42_____________2-H0..0_______________I.__.00I234Lungpressure(cmH20)5Figure 4.6: The effects of tonic changes in lung pressure on A) ventilatory index (‘Index), B)absolute breathing frequency (fLabs), C) peak integrated trigeminal nerve activity (VTIfldej,D) breaths per episode and E) episodes per minute. Each curve represent a differentlevel of hypoxic respiratory drive (• air; 0 6.0% 02).Page 126Ventilatoryindex(arbunits)SrAbsolutebreathingfrequency(mm1)A8Peakintegratedtrigeniinalburst activityQzVolt*sec)aot_C 0I2345Lungpressure(cm1120)DEpisodesperminuteE[t-.-4 3 2 0’B60I23452,0C1.5 to 0.5 0.0II.I0I2345Breathsperepisode4 3 2I’-)Lungpressure(cm1120)Figure 4.7: Panel A shows the changes in electroneurogram activity recorded from the vagus(X; top trace) and trigeminal nerve (V; middle trace) when lung pressure (bottom trace)was reduced from 2 cm H20 to 0 cm H20. Panel B is an enlargement of two fictivebreaths from panel A to illustrate the effects of changing lung pressure on the timing andcoordination of the trigeminal and vagal bursts. The bursts shown on the left hand sideof panel B were obtained at a lung pressure of 2 cm H20 while the bursts on the righthand side were recorded at a lung pressure of 0 cm H20. These recordings wereobtained from a bullfrog unidirectionally ventilated with a 3.3% CO2 in air gas mixture.Dotted lines indicate 1) the start of the vagal burst, 2) the start of the trigeminal burstand 3) the approximate termination of both vagal and trigeminal bursts.Page 128Ax Vlungpressure(cmtlO)510sec0BI23123IIIIIIVIIIIIIIIIiIIILIJ.h..I.IIiIJIjUJIlI...I-IIV:0.5secCt I’.)II1rIIIIjHjIIUIffiIHItjItfIiIlflIIflh1fjt(j--j)rIIWflIILFigure 4.8: The effects of tonic changes in lung pressure on vagal (open symbols) and trigeminal(closed symbols) burst duration. Each panel depicts the responses that were recorded atone of the four levels of respiratory drive. Note that when vagal burst duration equalsor is less than trigemmal burst duration (stippled area), no lung deflation would occur.Conversely, the greater the extent to which vagal burst duration exceeds trigeminal burstduration, the more time spent in expiration (hatched area).Page 130ABBurstduration(sec) C3.3%Co2Burstduration(sec)1.7%CO26.0%02L_..J_LII012345Lungpressure(cm1120) 4.9: The effects of tonic changes in lung pressure on the instantaneous frequencies of thetwo types of fictive buccal movements recorded the bullfrogs, lung or breathingfrequency (Panel A) and buccal oscillation frequency (Panel B). Each curve representsa different level of respiratory drive (• air; C 1.7% C02; + 3.3 % C02; 0 6.0% 02).Page 132Instantaneous br1eathingfrequency (mm ) A70-60r50:::10 L0E, I0 1 2 3 4 5Instantaneous buccaloscillation frequency (miii B703O—2010 -0 1 2 3 4 5Lung pressure (cm H20) Page 1333. Comparison between vagotomy and tonic Vagal feedback.The data obtained in this series of experiments showed that in frogs on UDV with 1.7%CO2 in air, increasing PL prior to vagotomy stimulated fictive ventilation (p < 0.001) (Fig.4.10). Although the absolute values for index, fLabs, peak trigeminal activity, trigeminal andvagal burst duration obtained for this series were not identical those reported in series I(compare Figs. 4.10 and 4.11 with Figs 4.5 and 4.8), the trends were identical. That is, staticincreases in vagal afferent feedback increased fLabs but reduced VTjndex causing an overallincrease in the index (p < 0.001 for all) (Fig. 4.10), while it also reduced trigeminal burstduration without affecting vagal burst duration (p < 0.00 1 and p = 0.092, respectively) (Fig.4.11).For unknown reasons, the frogs in this series of experiments were more sensitive to thehypercarbic stimulus than in series I. In the present series of experiments, raising PL above 0cm H2O transformed the breathing pattern from episodic to continuous (Fig. 4.10) in frogs OnUDV with 1.7% CO2 in air. In the previous series, however, this transformation was observedat 3.3% CO2 in air (Fig. 4.5).The summarized data presented in Figs. 4.10 and 4.11 show that, in vagotomizedbullfrogs, changing PL did not have any effect on breathing pattern (p > 0.05 for all). Thesefigures also show that a PL of 0 cm H20 and bilateral vagotomy, at any lung pressure, hadsimilar effects on the fictive breathing pattern on all of the breathing pattern variables that werequantified (p > 0.05 for all). That is, vagotomy had the same effect as deflation to 0 pressure.Fig. 4.11 demonstrates that elimination of PSR feedback had no significant effect on thevagal burst duration, but prolonged the duration of the trigeminal bursts, such that invagotomized frogs, vagal duration was always shorter than trigeminal duration. In spontaneouslyPage 134breathing frogs, vagotomy would result in substantial prolongation of the inspiratory phase,which is represented by the stippled area of Fig. 4.1 lB.Page 135Figure 4.10: The effects of changing lung pressure on A) ventilatory index (Index), B) absolutebreathing frequency (fLabs), C) peak integrated trigeminal nerve activity (VTJfldSX), D)breaths per episode and E) episodes per minute in vagotomized bullfrogs (open symbols),and bullfrogs with both vagi intact (closed symbols). All frogs were unidirectionallyventilated with a 1.7% CO2 in air gas mixture.Page 13615 I0 5 0Absolutebreathingfrequency(mmt)0I2345Episodes/mm0I2345Ventilatoryindex(arhunits)I0 5 0BPeaktrigenlinalactivityC(iVoItssec)2.0 1.5 1.0I0.50.00I234540 30 20 I0 00I2345Breaths/episodeI)8 6(4)4 2 00I2345(4)Lungpressure(cm1120)Lungpressure(cnt 1120)Figure 4.11: The effects of changing lung pressure on trigeminal (closed symbols) and vagal(open symbols) burst duration in bullfrogs with intact vagi (panel A), and after bilateralcervical vagotomy (panel B). All frogs were unidirectionally ventilated with a 1.7% CO2in air gas mixture. Note that after vagotomy, vagal burst duration was less thantrigeminal burst duration, so no lung deflation (stippled area) would occur. Conversely,in frogs with intact vagi, the extent to which vagal burst duration exceeds trigeminalburst duration indicates that more time was spent in expiration (hatched area).Page 138Burstduration(see)1.4 cJII12345Lungpressure(cm1120)DISCUSSION1. Critique of methods:The “reduced amphibian preparation used in the present study made it possible toinvestigate the ventilatory reflexes elicited by changing respiratory drive independent of changesin pulmonary pressure, and vice versa, because both feedback ioops could be openedindependently. Before considering the data, however, it is useful to compare the breathingpattern obtained with this preparation with that produced by intact frogs. Table 4.1 comparesthe breathing pattern and ventilatory responses that were observed in intact frogs on UDV,capable to spontaneously of producing changes in lung pressure (chapter 2) to those observedin the present study from decerebrate, paralysed frogs also on UDV, but with lung pressure setat 2 cm 1120. The reduced effect of the different stimuli on VTIfldex and the overall reduction infB1,, were the only notable differences between the two preparations. In the preparation usedin the present chapter, afferent input from upper airway (olfactory) receptors was alsoeliminated. In the previous chapter, this input was shown to modulate breathing pattern byexerting a depressant effect on ventilation when tonically elevated by CO2. Olfactory denervatedfrogs on the other hand showed a much larger increase in fLabs but no increase in VTjfldex whenexposed to hypercarbia, and had a significantly lower fB1 than intact frogs. It is not surprising,therefore, to observe these same differences in the in situ preparation compared to intact animalssince this reduced preparation was devoid of olfactory afferent input also.Page 140Table 4.1: A comparison of the breathing pattern variables between intact and decerebrate,paralysed bullfrogs. In intact bullfrogs, lung pressure oscillated spontaneously with eachbreath, whereas in the case of the paralysed frogs, lung pressure was set at 2 cm H20.The variables were measured at different levels of respiratory drive which was set bychanging the composition of the gas mixture used to unidirectionally ventilate theanimals. Note that because two different indicators were used as the VTjfldex for eachpreparation (integrated buccal pressure in the intact frogs, and peak integrated trigeminalactivity in the paralysed frogs), the absolute values cannot be compared directly and arenot reported. Given that VTindex is used in the calculation of the jndex, these absolutevalues are not reported either. For these two variables, the values are expressed as a %change from the 2% CO2 value. This level of respiratory drive was chosen as a pointof reference because frogs on UDV with this gas mixture have blood gases close to theones measured in spontaneously ventilating frogs not on UDV (chapter 2). * indicatesa statistically significant difference (two-way ANOVA, p < 0.05) between the twopreparations for that specific breathing pattern variable after allowing for the effects ofdifferences in CO2.Page 141Intact(chapter2)Decerebrate,paralysed(presentstudy)Air2%CO24%CO25%02Air1.7%CO23.3%026%02‘1index94±4.30±37213±20-33±-90±110±30182±25-65±38(%)(7)(6)(6)6.4(6)(6)(6)(7)(6)fLabs0.3±0.29±3.624±6.84.2±0.9±0.710±3.032±5.74±1.8(mirr1)(7)(6)(6)0.8(6)(6)(6)(7)(6)*IVTindex-64±440±1545±1354±128±40±1815±14-2±18(%)(2)(5)(6)(6)(2)(6)(6)(7)Breathsper1±02.8±0.68±1.14.1±1.4±0.12.8±0.94.1±1.7±0.3episode(2)(5)(6)0.7(2)(6)0.7(7)(6)(2)fL1,3443±5.344±2.837±3.240.6±0.835±1.738±1.729±3.7(min1)(1)(5)(6)(6)(2)(6)(6)(2)*fB62±8.750±4.944±2.749±3.439±0.632±2.134±2.640±1.8(min1)(3)(4)(5)(6)(3)(6)(4)(6)CDFigure 4.12: Recordings of integrated vagal activity (top trace; X) and trigeminal activity(bottom trace; V) that illustrate a fictive lung inflation cycle (type II episode) in whichthe motor output driving each fictive breath within an episode gets progressively larger.This particular recording was obtained from a vagotomized bullfrog unidirectionallyventilated with a 1.7% CO2 in air gas mixture.Page 143I7ITs01IVAfXfIn chapter 2, two different types of breathing episodes were commonly observed in intactbullfrogs. Each type of episode was characterized by its breath amplitude and PL profile. Forthe first type of episode (type I episode), the breath amplitude was relatively constant, and theonset of each breath was associated with a sharp decrease in lung pressure, which was restoredat the end of each breath. Thus, the PL profile of a type I episode revealed a rapid successionof relatively uniform PL oscillations, each of which was caused by deflation/inflation of thelungs with each breath. Since type I episodes were most frequently observed in the presentstudy, it was this type of PL profile that was reproduced in the phasic feedback experiments.This contrasts with lung inflation cycles (type II episode) in which the buccal pump wasactivated in a “ramp-like” fashion so that each breath within an episode produced a progressivelylarger increase in PL. Thus, the first breath in a type 11 episode was associated with a largereduction in PL, which was restored partially at the end of the breath as in a type I episode.With each subsequent breath, however, PL fell less at the start of the breath and was elevatedmore by the breath. In intact bullfrogs, type II episodes were usually observed when respiratorydrive was high. Although this type of episode was rarely observed in the present study, theENG recording shown in Fig. 4.12 demonstrates that the in situ preparation could producefictive lung inflation cycles. Given that this recording was obtained from a vagotomized animal,it would appear that the progressive increase of the bursting amplitude from one breath to thenext could be an endogenous property of the CNS.Finally, unlike vagotomy which removes afferent feedback from a variety of differentreceptors, it is assumed that changing PL in the present study altered only PSR feedback.Recording of impulses from pulmonary receptors conducted by the vagus nerve has led to theconclusion that the only lung receptors in the frog responding to lung inflation are stretchreceptors (Carleson and Luckhardt, 1920; Kuhlmann and Fedde, 1979; Milsom and Jones, 1977;Page 145Taglietti and Casella, 1966). Thus, the subsequent discussion will address the role of PSRfeedback in the control of episodic breathing.2. The effects ofphasic vs tonic PSR feedback.2.1. The role of phasic PSR feedback.By comparing results obtained under conditions where only static changes in PSRfeedback were imposed on the respiratory system to those obtained where the “natural”oscillations normally associated with the breathing cycle also occur, it is possible to assess theeffects of the pattern of the feedback signal on the breath by breath regulation of ventilation.In the present study, phasic changes in PSR feedback had no significant effect on trigeminalburst duration (TI) or any other variable that was measured. This contrasts with the results ofstudies performed on mammals in which phasic changes in PSR feedback during the breathingcycle were shown to be responsible for the decreases in TI associated with increases in VT(Clark and von Euler, 1972; Cross et al., 1980; Feldman and Gautier, 1976).Studies that have investigated the dynamic component of PSR reflexes have shown thatchanges in PSR feedback during the first 30% of the breathing cycle do not influence TI (Crosset at., 1980; for review, see Milsom, 1990b). The fact that lung inflation is rapid and, hence,changes in PSR feedback occurred early in the inflation cycle in bullfrogs may explain whyphasic PSR feedback had no additional effect on fictive breathing than tonic PSR feedback did.It is possible that manipulating the timing at which PSR feedback began to change during thebreathing cycle could have modulated breathing pattern differently; but the main goal of thisseries of experiments was to compare the effects of tonic feedback to a phasic signal similar tothe one produced by spontaneously breathing bullfrogs. For this reason, the effect of changingPSR feedback during different parts of the breathing cycle was not tested. Based on the dataPage 146obtained in this series of experiments, it must be concluded that if phasic feedback plays a rolein the control of breathing in bullfrogs, its contribution is likely to be limited.2.2. The role of tonic PSR feedback.2.2.1. Effects on breathing frequency.The present study has shown that increasing PSR feedback caused an increase in fLabs,owing mainly to a shortening of the duration of the non-ventilatory pause. Increasinghypercarbic respiratory drive also stimulated fL,5, but by increasing both the number of breathsper episode and the number of episodes per minute. The results have also shown a significantinteraction between PSR and chemoreceptor feedback, so that it is now possible to conclude thatPSR feedback exerts a permissive effect on (or disinhibits) chemoreceptor afferent input whichresults in an increase in breathing frequency.Kogo et al. (1994) also reported increases in fL,, upon lung inflation in bullfrogs. Theseauthors, however, observed this phenomenon in only 25% of the bullfrogs investigated. Thepresent study has shown that the slope of the relationship between PL and fLabs was directlyproportional to respiratory drive; it was very low in animals on UDV with air. It is possible,therefore, that Kogo and co-workers (1994) only saw an increase in fL5 on lung inflation in afew animals because they performed the lung inflation test only on frogs ventilated with air.There are few other studies that have addressed the role of PSR feedback on the controlof breathing pattern in ectotherms. One of the challenge of such studies is to distinguish thereflex effects of pulmonary mechanoreceptor input from those associated with other respiratoryrelated events. In turtles, chronic changes in resting lung volume (and thus, PSR feedback) havebeen shown to change breathing pattern under resting conditions but, unlike the present study,they had no significant effect on breathing frequency (Milsom and Chan, 1986). One majorPage 147difference between that study and the experiments presented in this chapter was that the timebetween the change in lung volume the onset of data recording was much longer in the turtlestudy (24 to 48h). It is conceivable, therefore, that habituation or adaptation of PSR dischargeaffected the responses observed in the turtles. In both turtles and frogs, slowly adapting stretchreceptors constitute the majority of the stretch receptor population, and are the main class ofreceptors involved in respiratory control (for review, see Milsom, 1990b). Following anincrease in PL, slowly adapting receptors undergo a brief period of adaptation over roughly 3mm in turtles (McLean et al., 1989) and 30 sec in frogs (Kuhlmann and Fedde, 1979). Inturtles, however, the adaptation is not complete and SAR discharge remains elevated indefinitely(McLean et al., 1989). Although a similar study on the effects of prolonged lung inflation onpulmonary stretch receptor discharge has not been performed in frogs, the fact that PSRdischarge is still elevated after a prolonged increase in lung volume in turtles indicates that thedifferent responses observed after chronic inflation in turtles and acute lung inflation in frogscan not be related to PSR adaptation. Habituation at the level of the CNS still remains apossible explanation for the discrepancies between the two studies.2.2.2. Effects on breathing pattern.Analysis of the breathing pattern recorded in the present study showed that the increaseof fLabs upon lung inflation was due to an increase in the number of breathing episodes perminute (or shortening of the non-ventilatory pause); the average number of breaths in eachepisode remained constant. The number of breaths in each episode was determined byrespiratory drive. The stimulating effect of hypercarbia on the number of breaths per episodewas very clear, and consistent with other reports on episodic breathers (for review, see Milsom,1991).Page 148This positive relationship between the level of PSR feedback and the number of breathingepisodes per minute was valid only at lower levels of CO2 (air and 1.7% CO2 in air) since inthe frogs on UDV with 3.3% CO2 in air, increasing PL above 0 cm H20 immediatelytransformed the breathing pattern from episodic to continuous breathing. It appears that the levelof respiratory drive at which changes in PSR feedback produced this dramatic transformationof the breathing pattern was close to 1.7% C02, since in the third series of experiments, thetransformation occurred at 1.7% CO2. This transformation of the breathing pattern fromepisodic to continuous breathing always occurred eventually, if PSR feedback was present.Continuous breathing was never observed in vagotomized bullfrogs, even at high levels ofrespiratory drive (6% CO2 in air; chapter 3). Unlike hypercarbia, hypoxia had no significanteffect on fL,, or on any component of the breathing pattern that was measured. This result wasnot too surprising, however, since the relative weakness of hypoxia as a ventilatory stimulantin frogs (chapter 2) and toads (Wang et al., 1994) has been documented previously.In turtles, chronic increases in lung volume reduced the number of episodes of breathingper minute but also increased the number of breaths in each episode; the net result being no netchange in minute ventilation. These effects of changing PSR afferent input on breathing patternwere eliminated by increasing respiratory drive (Milsom and Chan, 1986). In the Africanlungfish (Protopterus annectens) the role of lung volume, and thus PSR feedback, was studiedusing a reduced preparation similar to that of the present study (Pack et al., 1990). The resultsshowed that, much like the turtle (Milsom and Chan, 1986), increasing intra-pulmonary pressureprolonged the interval between lung breaths under normoxic condition. In lungfish, however,the effect of lung inflation on inter-breath interval was still maintained under conditions ofincreasing hypoxic respiratory drive. The question arises then as to why increasing vagalfeedback in frogs shortened, rather than prolonged the duration of the non-ventilatory pause (orPage 149increased rather than decreased the number of episodes per minute)? Although there weresignificant differences in experimental protocol between the other studies and the present work,none suggest a plausible explanation for the observed differences.The changes in breathing pattern produced by manipulating PSR and chemoreceptorafferent inputs were always accompanied by a change in fLab,. This suggests, therefore, that themechanisms responsible for the clustering of the breaths and the onset/termination of breathingepisodes are related to (or could not be dissociated from) the “total” respiratory drive, which isinfluenced by PSR and chemoreceptor afferent inputs.2.2.3. Effects on timing of the fictive breathing cycle.The effects of tonic increases in PSR feedback on the timing of the breathing cycle wasanother component of breathing pattern that was investigated. The scope of the analysis wassomewhat limited since it was based on measurement of the durations of the trigeminal and vagalENG bursts, which are estimations of TI and TT0T, respectively. Based on the EMG recordsfrom muscles involved in lung ventilation and the breathing mechanics model subsequentlyproposed by De Jongh and Gans (1969) for Rana catesbeiana of a similar size range,representative TB, TI and TT0T values in intact, freely moving bullfrogs breathing air would be0.21 sec, 0.41 and 0.62 sec, respectively2. From these data, the values of the different phasesof the breathing cycle obtained from trigeminal and vagal burst duration in the present studyappear accurate. These values made it possible to determine how chemo- and mechanoreceptorafferent inputs affected the relative duration of the expiratory and inspiratory phases of thebreathing cycle in bullfrogs.These values were obtained by digitizing the different EMG and buccal pressure tracessince no values were reported in this paper.Page 150Tonic afferent feedback from PSR’s had a significant effect on the timing of the breathingcycle in bullfrogs. Vagal feedback shortened trigeminal burst duration, without any significanteffect on vagal burst duration. In spontaneously breathing bullfrogs, such changes in burstduration would shorten inspiration and prolong expiration without any change in the totalduration of the breathing cycle; the net result would be a reduction in tidal volume. Thesereflexive changes in the timing of the bursts correspond to the expiratory-excitatory and theinspiratory-inhibitory reflexes that are well characterized in mammals (for review, see Milsom,1990b).The data also demonstrated that hypoxia and hypercarbia affect the timing of the burstsdifferently. Hypoxia had no effect on TI, but prolonged TT0T (thus, increasing TE), whilehypercarbia shortened both TI and TT0T (and also TE). Hypoxia and hypercarbia also havedifferent effects on the timing of the phases of the respiratory cycle in mammals (for review,see Mitchell et al., 1990). The specific effects of each stimulus on the timing of the breathswas, however, quite different from what has been reported for mammals. In mammals, hypoxiashortens TI and TE, and hypercarbia only shortens TE. The reasons why such differences existbetween bullfrogs and mammals are unknown. In amphibians as well as in mammals, hypoxiais detected by peripheral chemoreceptors, whereas increases inP02IH are detected primarilyby central chemoreceptors. This would suggest, therefore, that central and peripheralchemoreceptor afferent inputs have different projections and/or are integrated differently in thetwo groups.From these findings, it can be concluded that in bullfrogs, tonic PSR feedback is essentialfor proper and efficient coordination of the buccal pump and the glottal valve, so that the animalcan control tidal volume, and thus lung volume, on a breath by breath basis. ChangingPage 151chemoreceptor feedback affects the duration of trigeminal and vagal bursts differently,suggesting that each motoneuron pooi is controlled independently.3. Comparison between vagotomy and reducing tonic PSR feedback.Because the afferent information carried by the vagus nerve is not restricted solely tomechanoreceptors, it was quite conceivable that breathing pattern would be affected differentlyby vagotomy than by reducing or eliminating mechanoreceptor feedback. However, when thebreathing pattern of vagotomized frogs at any lung volume or pressure and non-vagotomizedfrogs at PL = 0 cm 1120 were compared, it became clear that mechanoreceptor feedback wasthe main afferent modulator of breathing carried by the vagus nerve; there were no significantdifferences between the breathing pattern of vagotomized and non-vagotomized bullfrogs. Thedata does, however, suggest that there was some residual tonic discharge from themechanoreceptors when PL = 0 cm 1120, and that vagotomy removed this residual tone.4. Interaction between PSR feedback and respiratoly drive.A two-way ANOVA confirmed that respiratory drive and PL interact in a positivefashion to modulate fL and jndex (p < 0.001). Interestingly, in mammals, mechanoreceptorinputs attenuate, rather than potentiate, ventilatory chemoreflexes (Cherniack et at., 1973;Kelsen et a!., 1977; Mitchell et at., 1980; 1982; Mitchell and Vidruk, 1987; Woldring, 1965).Until the neural pathways associated with PSR and central and peripheral chemoreceptor inputsare well described in amphibians, it is difficult to speculate on the mechanisms that allow PSRfeedback to facilitate (or disinhibit) chemoreceptor input. Such interaction is likely to occur atthe level of the central nervous system (CNS), and based upon analysis of the different domainsof receptor interactions proposed by Mitchell and co-workers (1990), this type of interactionPage 152would be a modulatory neural interaction. The interaction between PSR feedback andrespiratory drive could be significant during submergence, where deflation of the lungs wouldminimize the effects of increasing respiratory drive and perhaps prolong the duration ofsubmergence.5. Significance of buccal oscillation and lung ventilation frequencies.The instantaneous buccal oscillation and lung ventilation frequencies have been suggestedto be reliable indicators of endogenous respiratory rhythm in intermittent breathers (chapters 2and 3). The data obtained in the present study showed that instantaneous buccal oscillation andlung ventilation frequencies were not significantly different, and that they were not modulatedby chemoreceptor or PSR feedback. The fact that both events are produced by the same musclesleads to the supposition that they should somehow be related. Previous discussions on the topic(see chapters 2 and 3) indicate that both types of buccal movements could be driven by a singlerhythm generator, or, that they could be driven by two separate rhythm generators that werehighly entrained. The present data does not resolve this issue.6. Physiological significance.From the data presented in this chapter, it becomes increasingly evident that many basicrespiratory control mechanisms are shared by amphibians and mammals. For instance, allvertebrate groups investigated so far have a Hering-Breuer reflex which controls breathing ona breath by breath basis by modulating the timing of the ventilatory muscles. On a longer timescale, an interaction between chemo- and mechanoreceptors modulates breathing pattern andventilatory reflexes. It is obvious, however, that many details of these mechanisms are differentfor each group which reflect nuances that adapt the controller to the needs of the anina1 (e.g.Page 153continuous vs episodic breathing pattern; aspiration vs buccal force pump). Finally, it is alsoclear that while changes in PSR and chemoreceptor feedback, and their interaction, do haveprofound effects on the length of the interval between breathing episodes as well as the numberof breaths in each breathing episode, as long as respiratory drive is sufficiently high to stimulatebreathing, breathing occurs episodically.Page 154CHAPTER 5:INTRINSIC BRAINSTEM GENERATION OF BREATHING PATTERNPage 155INTRODUCTIONIn vitro brainstem-spinal cord preparations from neonatal rats (Issa and Remmers, 1992;Okada et al., 1993a, b; Smith and Feldman, 1987; Suzue, 1984) and lower vertebrates (turtle:Douse and Mitchell, 1990; bullfrogs: McLean, 1992; Walker et at. ,1990; carp: Adrian andBuytendijk, 1931; lamprey: Rovainen, 1985) have been used to study the neurophysiology ofbreathing. The recordings of respiratory related motor output reported in those studies haveshown that the distinctive features of the breathing patterns of endotherms and ectotherms aresometimes preserved in vitro, despite the fact that these preparations are essentially devoid ofany afferent feedback and descending neural influences. For instance, the “breathing” patternsof the neonatal rat in vitro and in vivo are continuous and rhythmic (Okada et al., 1993a, b;Smith and Feldman, 1987; Suzue, 1984), whereas that of turtle is mostly episodic and oftenappears arrhythmic (Douse and Mitchell, 1990). Thus, although differences clearly existbetween the fictive breathing pattern of in vitro preparations and intact animals, in vitropreparations appear to be an excellent experimental model to investigate certain components ofthe neural substrate of breathing patterns.The Doctoral dissertation of McLean (1992) established the brainstem-spinal cordpreparation from adult frogs as a valid experimental model for neurophysiological investigationsof respiratory control. Briefly, she has shown that 1) rhythmic motor activity can be recordedfrom specific nerve branches that innervate the respiratory musculature, and that the isolatedbrainstem of frogs demonstrates the expected central chemoreceptor induced burst frequencyresponse to changes in superfusate [W] (McLean, 1992). Unfortunately, her research did notaddress the nature of the breathing pattern produced by this preparation, although the data didindicate that the respiratory-related motor output produced by this preparations consisted mainlyPage 156of evenly spaced single “breaths”. Besides McLean’s work, descriptions of the breathing patternproduced by amphibian in vitro preparations consist of a few anecdotal reports of data obtainedfrom tadpole preparations (Walker et a!., 1990; Galante et al., 1992; Liao et a!., 1994) whichalso show a pattern of infrequent, evenly-spaced single breaths.The first objective of this chapter, therefore, was to determine the nature of the basicrespiratory-related motor output produced by this preparation (i.e. is it episodic?); and secondly,to determine if the respiratory-related motor output, or fictive breathing pattern, could bemodulated by 1) stimulation of the central chemoreceptors (produced by changing the pH/P02of the superfusate), 2) fictive vagal feedback (produced by stimulation of the pulmonary branchof the vagus nerve) and 3) whether an interaction between chemoreceptor and vagal feedback,similar to the one demonstrated in the previous chapter, also existed in this preparation. Thedata obtained from this work could substantiate the conclusion that emerged from the previousthree chapters; that while the onset/termination of breathing episodes is modulated by afferentfeedback, episodic patterns are an intrinsic property of the central respiratory network.Page 157MATERIALS AND METHODSThe study was performed on 11 adult bullfrogs (Rana catesbeiana) of either sex weighingbetween 153 and 262 g (mean weight = 192 ± 10.3 g, SEM). The animals were obtained andmaintained according to the methods described in chapter 2.Surgical procedures.The animals were cooled to 0°C in oxygenated water for at least lh before making asmall opening in the skull with a dental drill. The brain was then transected between the optictectum and the rostral forebrain. Following this, the frogs were paralysed by an injection ofpancuronium bromide (Pavulon; 2mgml’, 0.2 mglOOg’) into the dorsal lymph sacs. Thepulmonary branch of the right vagus nerve was located at the rostral end of the lungs anddissected back to its origin at the vagal ganglion, close to the cranium. A larger opening of thecranium was made with bone shears to allow dissection of the cranial nerves. Throughout theprocedure, the brain was irrigated with ice cold (0 - 5°C) mock CSF equilibrated with a 2% CO2balance 02 gas mixture. The mock CSF consisted of (in mM): NaC1, 75; KC1, 4.5; CaC12,2.5;MgCl2, 1.0; NaH2PO4,1.0; NaHCO3,40 and glucose, 7.5. The NaHCO3 concentration waschosen because preliminary experiments indicated that the preparation was more stable whenperfused with a mock CSF containing a high [HCO3-]. Okada and co-workers have suggestedthat high [HCO3-] may facilitate CO2 diffusion out of the tissue and, thus maintain tissue pH atthe depth of the respiratory-related neurons at a more physiological level (Okada et al., 1993).Thus, the “control” pH of the mock CSF (8.0) was slightly above the normal pH range (— 7.9)for CSF in frogs at 22°C. The brainstem was transected caudal to the hypoglossal (XII) nerveand transferred to a small petri dish coated with Sylgard (Dow Corning) where it wasimmobilized with insect pins; the arachnoid and pia membranes were removed. The brain wasPage 158then moved to the recording chamber where it was pinned ventral side up. The left cranialrootlets of nerves V and X were fitted in suction electrodes while the pulmonary branch of theright vagus was positioned on a bipolar silver hook electrode and covered with a 1:1 mixtureof vaseline and mineral oil. The recording chamber was then perfused with mock CSF bubbledwith a 2% CO2 in 02 gas mixture (pH 8.0) at a rate of — 5ml/min. A diagram of theexperimental system is illustrated in Fig. 5.1. Experiments were conducted at room temperature(22°C).Experimental procedures.The fictive breathing recorded shortly after the dissection was often rapid and unsteady.The preparation was therefore allowed to stabilize for at least lh before experiments began.Preliminary experiments demonstrated that most preparations were stable for at least 8 hours.Fifteen percent of the preparations did not remain stable for the entire duration of the experimentas judged by a significant deviation from the fictive breathing pattern initially produced at pH8.0; these data were not included in the data set. Electrical nerve activity recorded from thesuction electrodes was amplified (filter settings: 50 Hz (high pass) and 10 kHz (low pass)), fullwave rectified and integrated (Gould) in 67 ms intervals. The raw and integrated nerve signalswere viewed on an oscilloscope and stored on a polygraph recorder and on computer disk witha data acquisition system (Codas). The sampling rate of the AID conversion was 2500 Hz.Page 159Figure 5.1: Schematic representation of the experimental system utilized for the amphibian invitro brainstem-spinal cord preparation.Page 160C————MockCSFAMPHIBIANBRAINSTEM-SPINALCORDPREPARATIONSuctionrVagusnerve(pulmonarybranch)iiLbII‘TIILi11‘1IExperimental protocol.The experiments consisted of recording extracellular nerve activity from cranial nervesV and X at 3 different pH levels (pH = 7.7, 8.0 and 8.3) with and without stimulation of thepulmonary branch of the vagus nerve. As mentioned earlier, pH 8.0 was chosen to approximatenormal CSF pH in frogs at 22°C and pH 7.7 and 8.3 were chosen to represent an acidotic andailcalotic stimulus respectively. The pH was determined by changing the CO2 composition ofa gas mixture bubbling through the mock CSF (Fo2 = 0.045, 0.020 and 0.013; calculatedCO = 32.5 ± 1 Torr, 17.1 ± 0.2 Ton and 9.3 ± 0.3 Ton respectively). The pH of themock CSF was measured with a pH meter (Corning) with its electrode positioned in theperfusate reservoir. The CO2 composition of the gas mixture delivered to the perfusate wasmonitored continuously using a CO2 analyzer (Beckman LB-2). To ensure that there were noresidual effects between each test, two experimental groups were used; each of which receivedthe acidotic or alkalotic test in a different order. For group 1 (N = 6), the order was: control(8.0), acidotic (7.7), control, alkalotie (8.3), control. For group 2 (N = 5) the order was:control, alkalotic, control, acidotic, control. A 15 mm equilibration period was allowed betweeneach pH change. Since statistical analysis failed to demonstrate any difference between thegroups, the data for each pH and vagal stimulation condition were pooled. Each test began witha recording of the ‘resting” activity of the preparation under control conditions (pH = 8.0) for5 mm. This was followed by a s mm period of phasic stimulation of the pulmonary branch ofthe right vagus. The stimulus (20 Hz, 0.2 ms) was delivered each time a burst of respiratoryrelated discharge (fictive breath) occurred in the contralateral vagus. The stimulation frequencywas based on literature values previously reported for resting pulmonary stretch receptordischarge in frogs at room temperature (Kuhlmann and Fedde, 1979; Milsom and Jones, 1980;Taglietti and Casella, 1966; 1968). Before the onset of the experiment, the lowest voltagePage 162capable of affecting bursting frequency was determined (typically, this was 2 Volts) and thestimulus intensity was set just above this level. To ensure the synchrony between the fictivebreaths and vagal feedback, the stimulator (Grass S88) was triggered by a peak detector that wasactivated when integrated trigeminal nerve discharge rose 4.5 V sec above its baseline, and deactivated when it fell below this value. The bout of phasic stimulation was followed by a 5 mmrecovery period. This procedure was then repeated at the next pH for the specific protocol.Data analysisValues for fictive breathing frequency were obtained by analyzing 5 mm segments ofdata. Any increase in integrated nerve discharge 4.5 jV . see above baseline was considered tobe a fictive lung ventilation, while the smaller discharges were considered to be buccaloscillations. The breathing pattern variables (fLthS, VTifldex, Index, breaths per episodes, episodesper minute, trigeminal and vagal burst durations, fL and fB1) were quantified the same wayas in chapters 2, 3 and 4. All data are presented as means ± 1 standard error of the mean(S.E.M.), and are plotted with mock CSF rather than pH as the independent variable tofacilitate comparison with the data reported in the previous chapters. The results werestatistically analyzed using a two-way analysis of variance followed by a Student-Newman-Keulstest (p < 0.05).Page 163Figure 5.2: Comparison between the breathing pattern recorded from an intact bullfrogbreathing air (top panel) and motor output recorded from an in vitro brainstem-spinalcord preparation (other panels). The middle pair of recordings were obtained from thetrigeminal nerve; the first trace is the raw signal while the one immediately below is thesame signal after full wave rectification and integration. The bottom pair of recordingsare the raw and full wave rectified/integrated signals simultaneously obtained from thevagus nerve.Page 164Buccal presswe Whole animal breathing air(cm H20)L• In vitro brainstem-spinal cord preparationTrigeminal nerveactivity (pH 7.7; no vagal stimulation)IHwr1III14rIIsr4IIh i L1 r:ri )iIi rii zjiVagal nerveactivity.•j. i-- Lk. -__ _I I30 secPage 165RESULTS1. “Fictive breathing pattern” of the bulfrog in vitro brainstem-spinal cord preparation.The recordings presented in Fig. 5.2 compare the breathing pattern of an intact bullfrogto the pattern recorded from an in vitro preparation perfused with mock CSF at pH 7.7 withoutvagal stimulation. The buccal pressure recording from the whole animal shows that breathingoccurred mostly in doublets or as evenly spaced single breaths. Small amplitude buccaloscillations occurred during the non-ventilatory pause between episodes. A similar pattern offictive lung ventilation was also seen in the motor output recorded from the cranial nerve rootletsV and X in most in vitro preparations. Furthermore, in some preparations, bursts of smalleramplitude were recorded from the trigeminal nerve rootlet between bursts of larger amplitude.These occurred in a very rhythmic fashion and comparable activity could rarely be recorded inthe vagal nerve rootlet of any preparation (Fig. 5.2). These small rhythmic bursts resemble thesmall amplitude buccal oscillations described in intact frogs which involve activation of cranialnerve V to the buccal pump muscles, but not cranial nerve X to the glottis (De Marneffe-Foulon,1962; chapters 2 and 3).It should be noted that although breathing patterns were obtained from the in vitropreparation which were similar to those of intact animals, the overall level of motor outputproduced by the in vitro preparation was reduced. The comparison shown in Fig. 5.2 is betweenan intact animal breathing air and an in vitro preparation with no vagal feedback but with anacidic pH.Page 166Figure 5.3: Panel A shows neurograms of fictive breaths recorded simultaneously from thevagal (top trace) and trigeminal (bottom trace) cranial nerve rootlets in the brainstemspinal cord preparation. Panel B shows that adding vagal feedback did not affect theduration or the relative timing of the bursts.Page 167pH 8.0, no stimulation AxVxVpH 8.0, vagal stimulation(2.0 V, 20 Hz, 0.2 ms)I0.5 secBPage 168Figure 5.4: The effects of changing mock CSFIP0 on the trigeminal (panel A) and vagal(panel B) burst duration with and without vaga feedback.Page 169Trigeminal burstduration (sec)0.7- A0.6 —0.4 H0.30.25 10 15 20 25 30 35Vagal burstduration (see)0.7E BLLL0.6 F0.50.3r0.25 10 15 20 25 30 35Mock CSF P (Torr)C2Page 170Figure 5.3 illustrates ENG recordings obtained from the trigeminal and vagal nerverootlets simultaneously. Typically, activity in the trigeminal nerve rootlet preceded that of thevagus, and ended at the same time or slightly after the vagal burst. The net result was that theduration of the trigeminal burst was always longer than that of the vagus. The durations of thebursts recorded from the trigeminal and vagal nerve rootlets were constant; neither changes inmock CSF nor addition of vagal feedback affected the duration of the fictive breaths (Fig.5.4).2. Respiratory reflexes in vitro.Figure 5.5 illustrates trigeminal nerve activity recorded at different values of mock CSFpH with and without phasic stimulation of the pulmonary branch of one vagus nerve. Thisfigure was obtained from a preparation that responded particularly well to vagal stimulation.This response was much larger than that obtained on average (Fig. 5. 6B) and is shown toindicate the range of responsiveness which can be obtained with this preparation. In the absenceof vagal stimulation, changing mock CSF pH had a very modest, yet significant (p < 0.05),effect on fictive breathing frequency (Fig. 5.5; Fig 5. 6B, closed circles). The addition of vagalfeedback had a significant effect on the relationship between bursting frequency and pH,indicating that the responsiveness of the preparation to changes in pH/CO2was increased (p <0.001) (Fig. 5.5; Fig. 5.6B, open circles). Neither changes in mock CSF pH/PC02nor vagalstimulation affected VTifldex,(p = 0.167 and p = 0.994, respectively) (Fig. 5.6C), thus, thechanges in Vlndex were caused solely by the frequency response (Fig. 5. 6B).Breathing remained episodic at all levels of chemoreceptor stimulation with or withoutvagal feedback. Without vagal feedback, the average breathing episode rarely had more thanone breath, even at the highest level of chemoreceptor stimulation. The effect of increasingPage 171superfusate on breathing pattern was significantly augmented by the addition of vagalfeedback such that, at the highest level, the average breathing episode now consisted of3 or 4 fictive breaths (Fig 5.6D) (p < 0.001). In addition, the number of breathing episodesper minute was also related to the of the superfusate, but was not significantly affectedby the addition of vagal feedback (p = 0.292) (Fig. 5.6 E). Figure 5.7 shows the summarizeddata for both instantaneous buccal and lung frequencies. Overall, was significantly less (p<0.05) than and neither were significantly affected by chemoreceptor or vagus nervestimulation (p < 0.05).Page 172Figure 5.5: The effect of changing mock CSF pH/P02 on respiratory related activity recordedfrom cranial nerve rootlet V of the bullfrog before and during phasic stimulation of thepulmonary branch of the contralateral vagus nerve. This experiment was performed at20°C. Each trace is the full wave rectified, integrated signal. The dark lines under theintegrated signal indicate the times during which stimuli were delivered. This figureillustrates one of the most vigorous responses obtained with vagal stimulation.Page 173Iiivitrobrainstem-spinal cordpreparationMock(‘SFIvp11=8.325p[-:—---iIll---iI111111-Il-IlIl—iIp11=8.()25pV•sec[———HhIU———-——1111111Illi[—III1111111liii—II——2SpV.sec[ jjjjjg7.7--1111111111 umwHillIft UIUNWWMIUUMliHi NMMMIIIHUNILSiiii,ulaioi(Hit)SCUFigure 5.6: The effect of changing mock CSF pH/Pco2 via changes in on A) ventilatoryindex (‘Index), B) fictive breathing frequency (fictive breaths; fLabs), C) peak integratedtrigeminal activity (VTIfldex), D) breaths per episode and E) episodes per minute. The datawere obtained with (filled circles) and without (open circles) stimulation of the pulmonarybranch of the vagus nerve. Values are from 11 preparations and are presented as means± 1 standard error of the mean (S.E.M.).Page 175Ventilatoryindex(arbunits)152535BPeakintegratedtrigeminalburstactivity(Vo1t*sec) 3 2 1D5101520253035Episodes*mill110 8 6 4 2 0Eonostimulus•stimulusonMockCSFP02(Torr)30 20 10 0AAbsolutebreathingfrequency(mm)30 20 10C05101520253035L 50.0I..I...20253035H5101520253035Figure 5.7: The effects of changing mock CSF pH/PO on the instantaneous breathing (panelA) and buccal oscillation (panel B) frequencies with (filled circles) and without (opencircles) vagal feedback.Page 17780- A70 n60k-i..50 H40Instantaneous bijeathingfrequency (min)5 10 15 20 25 30 35Instantaneous buccaloscillation frequency (min1)8070B5 10 15 20 25 30 35Mock CSF P02 (Torr)Page 178DISCUSSION1. Critique of method.Because in vitro brainstem-spinal cord preparations are not perfused naturally, thedelivery of 02 and removal of metabolic end products are limited since they rely on diffusionalone. This potential degradation of the internal milieu in the vicinity of the respiratoryneurones must be prevented; otherwise, the preparation would become unstable, and therespiratory-related motor output would reflect a pathological condition that would yield spuriousinformation. To circumvent these problems, researchers interested specifically in mammalianrespiratory control have used the brainstem-spinal cord of neonatal rats. Its small size reducesthe surface to volume ratio, and minimizes the diffusion distance between the respiratoryneurones and the superfusate. In addition, the metabolism of the preparation is reduced bymaintaining it at 27°C (Smith and Feldman, 1987; Suzue, 1984). These precautions helpmaintain the quality of the microenvironment of the respiratory neurones (Brockhaus et al.,1993; Okada et a!., 1993a, b), and have produced a reliable experimental model which hascontributed to the production of significant advances to our understanding of respiratoryrhythmogenesis (Smith and Feldman, 1987; Smith et a!., 1991b; Suzue, 1984). The fact thatthe experiments are performed on an immature system at subphysiological temperatures,however, may limit the applicability of the information derived from this preparation to theintact, behaving animal.These latter criticisms which have been levelled at the mammalian preparation are notrelevant to the in vitro preparation from ectothermic animals. Many authors have outlined theadvantages of these preparations over the mammalian model (See reviews by Berger, 1990;Mitchell, 1993). Briefly, the brains of most ectothermic animals are relatively small, theirPage 179metabolic rate is usually lower than that of mammalian tissue at the same temperature, they canfunction normally over a wide range of temperatures, and they can tolerate hypoxia. For thesereasons, it is possible to perform in vitro studies on adult brains of lower vertebrates atphysiologically relevant temperatures. Moreover, a comparative approach to the study ofneurophysiology of breathing can help broaden the scope of our understanding of respiratorycontrol in vertebrates and is preferable to becoming overly focused on a single experimentalmodel.One must always be careful when interpreting the data obtained from a reduced systemsuch as the one described in this chapter, since removing elements of the respiratory controlsystem may cause complex transfonnations rather than simple reductions (Feldman et al., 1990;Mitchell, 1992). As discussed below, some of the respiratory reflexes affecting burst patternformation were not observed in the present study. Nonetheless, the sum of the data presentedin this report confirms the conclusion of McLean (1992), which stated that the in vitro brainstemspinal cord preparation from the bullfrog is a useful experimental tool for studying theneurophysiology of respiratory control. The data also strongly suggest that the differences thatexist between the motor output recorded from this experimental model and that of the intactbullfrog reflect reductional, rather than transformational changes.2. “Fictive breathing pattern” of the bullfrog in vitro brainstem-spinal cord preparation.The recordings presented in Fig 5.2 have shown that it is possible to record a rhythmicmotor output from the in vitro brainstem-spinal cord preparation which appears virtually identicalto the breathing pattern of intact bullfrogs. As it was described in intact frogs (De Jongh andGans, 1969; West and Jones, 1975; for review, see chapter 1), two types of bursting activitywere recorded from the trigeminal cranial nerve rootlet; bursts of small amplitude, whichPage 180represent the buccal oscillations, and the larger bursts which represent lung inflation in the intactfrog. Most often, bursts of large amplitude which represented activation of the glottis, wererecorded in the vagus; small-amplitude bursts were rarely recorded from this nerve rootlet.Although the duration of the bursts recorded in vitro were shorter than those recorded in situ,the discharge profile and timing of the bursts recorded from the vagal and trigeminal nerverootlets were very similar to those reported for the in situ frog preparation (chapter 4) afterbilateral vagotomy; i.e. the trigeminal burst was longer than the vagal burst. This is notsurprising given that the in vitro preparation was vagotomized. The fictive breathing frequencyof the preparation under “control” conditions (pH 8.0, no stimulation) was slower than that ofintact frogs post-vagotomy breathing air (chapter 3), but was similar to the frequency recordedin situ after bilateral vagotomy (Chapter 4). Because removal of olfactory receptor input hadno significant effect on breathing pattern and overall ventilatory output in normocarbic frogs(chapter 3), these differences which were present in both the in situ and in vitro preparationsmay reflect the effects of decerebration.In the context of the present investigation, the important finding from this in vitro studyis that the spatio-temporal distribution of the breaths was episodic, much like that of the intactanimals, despite the fact that this preparation was devoid of afferent feedback.3. Respiratoiy reflexes in vitro.3.1. The effects of central chemoreceptor stimulation.The relationship between the number of breaths in a breathing episode and hypercapnicrespiratory drive is well established in intermittent breathers with intact PSR feedback (frogs:chapters 2, 3, and 4; turtles: Milsom and Chan, 1986; hibernating ground squirrels: McArthurPage 181and Milsom, 1991a, b; For reviews, see Milsom, 1990a; 1991). The data indicate that thosereflexive change in breathing pattern were also present in vitro, but were more noticeable in thepresence of vagal feedback (discussed below). This may explain the notable absence of distinctbreathing episodes in the work of McLean (1992) using a similar preparation to the one used inthe present study. Based on the description of the surgical procedures, it would appear that theneural elements present in McLean’s preparations (1992) were the same as those described inthis chapter. In the former study, however, the brainstem-spinal cord preparation wassuperfused with a solution that contained no CO2 (hepes buffer aerated with 100% 02). It isvery likely, therefore, that absence of breathing episodes of more than a single breath inMcLean’s (1992) preparation reflected the lack of central chemoreceptor stimulation and possiblythe lack of phasic vagal feedback.It is noteworthy that, in neonatal rat preparations, substitution of hepes buffered saline(equilibrated with a 100% 02 gas mixture) with bicarbonate buffered solution (equilibrated witha 95% 02, 5% CO2 gas mixture) has been reported to transiently induce episodic breathing 50%of the time (Brockhaus et at., 1993). Since episodic breathing has also been reported inmammalian species, such as the golden mantled ground squirrel (Milsom, 1991) and the northernelephant seal (Castellini et at., 1994), under conditions of reduced metabolism (hibernation andsleep, respectively), and in dogs in which the hypercapnic respiratory drive was reducedartificially (Bartoli et al. 1974), it is conceivable that the transient observation of episodicbreathing in the neonatal rat at 27°C was not an artefact, but perhaps a reflection of the “true’motor output of this preparation. The demonstration that mammalian in vitro brainstem-spinalcord preparations can reliably produce a fictive breathing pattern which are episodic would bringstrong support to the argument that respiratory control is a system highly conserved inPage 182vertebrates. Unfortunately, until this question is addressed specifically, this attractive idea willremain speculative.Page 183Figure 5.8: The effect of tonic vagal stimulation on the fictive breathing pattern. In thisrecording, the top trace is full-wave rectified and integrated signal recorded from thetrigeminal nerve rootlet. The dark line under the integrated signal indicate the time atwhich the stimulus was delivered.Page 184CD 00 L1pH7.7,tonicvagalstimulation(2.0V,20Hz,0.2ms)N30sec3.2. The effects of vagal stimulation.Although vagal feedback produced dramatic changes to the breathing pattern and CO2sensitivity of the in vitro preparation, the pattern of the vagal stimulus was important since onlyphasic stimulation caused changes in breathing pattern. Tonic stimulation resulted in randomdischarge or, as depicted by the ENG recording shown in Fig. 5.8, reversibly silenced the motoroutput. These results differed from those obtained with the in situ preparation, whichdemonstrated that phasic and tonic PSR feedback had identical effects on fictive breathing. Thereasons underlying this discrepancy are unknown, but may indicate that the stimulus parametersdid not perfectly mimic PSR feedback. A second possibility is that vagal stimulation elicitedmore reflex effects than those arising from PSR, and that this was particularly true with tonicstimulation. The effects of tonic lung inflation on respiratory-related motor output have beenstudied in an isolated brainstem-lung preparation of the newborn rat (Murakoshi and Otsuka,1985). That investigation of mammalian respiratory reflexes in vitro revealed that inflation ofan isolated lung resulted in a transient inhibition of respiratory-related activity which wasbelieved to correspond to the Hering-Breuer inflation reflex. These results were consistent withthe effects of lung inflation reported in mammals, and suggest that tonic stimulation of the PSR,rather than the pulmonary branch of the vagus might have produced a more predictable resultsin the present study. These suggestions would also explain the fact that the addition of vagalfeedback in vitro did not modulate the amplitude or duration of the trigeminal burst as it did inchapter 4. However, changes in mock CSF also failed to reduce trigeminal and vagalburst durations in vitro, as reported in the in situ preparation in the previous chapter. Clearly,the CO2 stimulus was being perceived by central chemoreceptors in vitro, but the effects of thestimulus, just like the vagal stimulus, were primarily confined to the respiratory rhythm, and hadPage 186little effect on burst pattern formation. Again, the reasons underlying this shortcoming of thepreparation are unknown, but may reflect the absence of either some descending influence ofperipheral input. Even though the in vitro preparation used in the present study is a goodexperimental model to investigate respiratory rhythm (frequency), it cannot currently be usedreliably to study the mechanisms underlying burst pattern formation.Although it is clear that vagal feedback affects breathing frequency and breathing patternin bullfrogs, the different responses observed with each preparation make it is difficult todetermine precisely which of the two main components of the pattern it modulates. Forinstance, “fictive vagal feedback was shown to affect only the number of breaths per episodein vitro (present study), whereas in situ, lung inflation influenced only the number of episodeper minute (chapter 4), while both components of breathing pattern were affected by bilateralvagotomy in intact frogs (chapter 3). Although some discrepancy is not surprising given thehigh variability inherent in the quantification of breathing pattern, the important differencesbetween the different experimental approaches may also contribute to these discrepancies.3.3. Interaction between central chemoreceptor input and vagal feedback.Notwithstanding the intricate differences associated with the changes in breathing patternupon vagal stimulation in the different experimental systems, the results have demonstrated thatvagal feedback exerts a facilitating (or disinhibiting) effect on the chemoreceptor input in vitro;the responsiveness of the preparation to the same changes in increased nearly threefoldwith the addition of vagal feedback. The data obtained from the in situ experiments (chapter 4)led to the suggestion that this modulatory interaction occurred at the level of the central nervoussystem. Since these results were corroborated in vitro, it is now clear that this interaction doesPage 187occur at the level of the CNS. Until the neural pathways associated with PSR, central andperipheral chemoreceptor inputs are well described in amphibians, it is difficult to speculatefurther on the mechanisms that allow vagal feedback to facilitate (or disinhibit) chemoreceptorinput. This effect of vagal input on CO2 sensitivity in bullfrogs differs from the attenuation thatPSR feedback exerts on ventilatory chemoreflexes in mammals (for review, see Mitchell et a!.,1990). These differences have been discussed in detail in the previous chapter, and theargument can be summarized here by reiterating that the facilitation of the hypercarbicventilatory chemoreflex by vagal feedback is, perhaps, specific to episodic breathers.4. Significance of buccal oscillation and lung ventilation frequencies.The data obtained in the present study showed significant differences between buccal andlung instantaneous frequencies. Since instantaneous frequencies have been used as indicators ofintrinsic respiratory rhythms (see discussion of chapter 2) the data reported here support thehypothesis that two separate rhythm generators drive each type of event independently.Regardless of the case, because fL and fL1 were little affected by vagal feedback or centralchemoreceptor stimulation, it seems that these central rhythms were relatively constant.5. Physiological significance.The present investigation confirmed that the onset/termination of breathing episodes isnot dictated solely by afferent feedback because intermittent breathing was observed in apreparation that was almost completely devoid of afferent feedback. It is now clear that episodicbreathing is an endogenous property of the central nervous system. This conclusion supportsthe hypothesis proposed by Jackson (1978), that in intermittent breathers, the episode is thefundamental output unit of the respiratory control system rather than the individual breath withinPage 188the episode. This then raises the question: “How is this control system organized to producethis type of motor output?”Page 189CHAPTER 6:IS THE NUCLEUS ISTHMI RESPONSIBLE FOR THE ONSET/TERMINATION OFBREATHING EPISODES?Page 190INTRODUCTIONThe preceding chapters in this thesis have led to the conclusion that the production ofepisodic breathing is an endogenous property of the respiratory controller at the level of thecentral nervous system. The role of the central nervous system in the control of respiration hasintrigued researchers for many years, and a significant portion of work on this topic hasconsisted of studies employing progressive brain transections in the rostral-caudal axis designedto find the “noeud vital”, the centre responsible for providing the continuous drive to breathe.Of particular interest to the present research are a few studies performed on lower vertebrateswhich also documented the changes in breathing pattern associated with the transections, as theyconstitute the empirical basis of the investigations in the present chapter on the central neuralsubstrate for episodic breathing. These studies by Oka (1958a, b) and Naifeh and co-workers(197 la, b) showed in bullfrogs and caiman, respectively, that transection of the brainstembetween the optic lobes and the cerebellum eliminated breathing episodes such that the breathingpattern appeared to consist of evenly spaced single breaths. These findings suggest that thegrouping of breaths into episodes is under the influence of a site independent of that responsiblefor producing the respiratory rhythm; i.e. there is a site responsible for turning the breathingepisodes on and off.From their descriptions, the area in which transections elicited these intriguing changesin breathing pattern in bullfrogs appeared to correspond to the nucleus isthmi (NI), a nucleuslocated between the midbrain roof and the base of the cerebellum (Fig. 6.1). This nucleus is oneof the most conspicuous structures of the anuran brain, and although it has been suggested thatthe NI may be a special sensory relay centre, its functional significance is entirely obscure(Nieuwenhuys and Opdam, 1976; Opdam et a!, 1976). Interestingly, this nucleus goes throughsubstantial cellular arrangement and differentiation during metamorphosis (Senn, 1972); a periodPage 191which is also associated with the onset of episodic breathing in the bullfrog. Together, thiscircumstantial evidence leads to the hypothesis that the NI plays a role in the control of episodicbreathing in bullfrogs.The objectives of this study were to 1) determine if functional elimination of the NIaffects breathing pattern, and if so, 2) how this affects the other respiratory control mechanismsand reflexes that have been described in the previous chapters. This study was performed onthe in situ preparation that was used in chapter 4, because the fictive breathing pattern andrespiratory reflexes produced by this preparation were very similar to those of intact bullfrogs.The study compared the breathing pattern and ventilatory responses of bullfrogs before (intact)and after bilateral lesions of the NI produced by microinjection of the local anesthetic lidocaine,and the neurotoxin kaimc acid (KA) into this area.Page 192Figure 6.1: Lateral view of the midbrain, medulla oblongata and part of the spinal cord of Ranatemporaria. Topographic reconstruction of the cranial nerves and nuclei. Cer.cerebellum; coil, post = colliculus posterior; VIII do. = dorsal V1IIth nucleus; VIII ye.= ventral VIlith nucleus; VIII a. = nucleus vestibularis anterior; CN = cerebeliarnucleus; se. = sensory; mo. = motor; N. spin. = spinal nerves. The arrow indicatethe location of the nucleus isthmi (isth) (Adapted from Senn, 1972).Page 193VIPage 194MATERIALS AND METHODSExperiments were performed on 17 adult bullfrogs (Rana catesbeiana) of either sexweighing between 165 and 469g (mean mass = 295 ± 20g), obtained from a commercialsupplier. The animals were maintained according to the protocol described in chapter 2.Surgical procedures.The frogs were decerebrated, paralysed and unidirectionally ventilated (UDV) accordingto the protocol described in chapter 4. In these present experiments, however, the UDV flowrate was set at 70 mlmin’, the craniotomy reached more caudally to expose the rostral portionof the choroid plexus, and the dura covering the area between the cerebellum and the optictectum was removed to facilitate the penetration of the microinjection pipette. Small cottonpellets soaked with physiological saline were then placed over the area to prevent desiccation.Experimental procedures.The basic procedures for nerve recording, data acquisition and blood pressure monitoringwere identical those described in chapter 4.Microinjections of kainic acid (KA, 4.7mM in saline), lidocaine hydrochloride (1 % inNaCI), saline (0.7% NaC1) and fast green FCF (10% solution in saline) into the area of the NIwere made with multibarrell pipettes built from microfilament capillary glass (A-M Systems,Inc.). Each barrel had an inner diameter of 0.6 mm and an outer diameter of 1.2 mm. Theglass was pulled with a Narashige pipette puller and broken back until the total tip diametermeasured roughly 40 xm. Each barrel was 2/3 filled with its respective solution so that eachmeniscus was approximately the same height. A piece of polyethylene tubing (PE 10, ClayAdams) was placed inside each barrel of the pipette, and then secured with 5-minute epoxyPage 195resin. A 27 gauge needle was inserted in each piece of PE tubing before being connected to thepressure ejection system (Picospritzer II, General valve corp.).The injection volumes were calculated from the following equation:Volume = height• ir radius 2where the changes in meniscus height were measured with the reticule of a dissection microscopeduring the injections.Experimental protocol.Series I: The “fictive” breathing pattern of bullfrogs under conditions of normo- and hynercarbicrespiratory drive at different tonic levels of lung pressure.This series of experiments was designed to measure the changes in total “fictive”ventilation and breathing pattern associated with hypercarbia and tonic changes in lung pressurein animals with their NI intact (control group). Although this is the same protocol as that usedin chapter 4, by comparing the changes in breathing pattern and the overall responses tohypercarbia with data in chapter 4, it was possible to determine whether the proceduresassociated with the more invasive craniotomy had a significant effect on ventilatory output.Breathing was assessed in frogs unidirectionally ventilated with three gas mixtures: 1 %, 2% and3.5 % CO2 in air. For each CO2 level, fictive breathing was monitored at three different toniclevels of lung pressure: 0, 2, and 5 cm H20. The details of the experimental protocol weregiven in chapter 4.Page 196Series II: The effects of bilateral lesions of the nucleus isthmi on breathing pattern andventilatory reflexes.This series of experiments compared the breathing pattern and reflexive changes inventilatory output observed after changes in PSR feedback and chemoreceptor drive of frogsafter bilateral lesions of the NI to that of intact animals (control group; series I). Themicropipette was positioned from a dorsal approach into the region of the NI. This region lies1.5 - 2.0 mm below the level of the cerebellum and 0.8 mm from the midline at the level of thejunction between the optic tectum and the cerebellum. Each experiment began with 175 nl shaminjection of physiological saline (0.7% NaC1) using a series of 1 to 5 rapid pressure injections(80 PSI, 400-800 ms duration) until the meniscus moved the appropriate distance. This volumeof KA was shown in preliminary experiments to consistently produce the changes in breathingpattern that will be described below. The sham microinjection was followed approximately 5mm later by an injection of a similar volume of KA; this glutamate analogue has been shownpreviously to selectively lesion neuronal cell bodies while sparing axons of passage (Coyle et al.,1978; Denavit-Saubié et al., 1980). The injection of KA in this area initially provoked anincrease in bursting frequency and amplitude from both nerves, as well as the level of tonicdischarge mainly in the vagus nerve. This transient period of increased activity was followedby a period of quiescence that ranged between 15 mm to 1 h (Fig. 6.2). Once fictive breathingreappeared, the breathing pattern became stable in, on average, 25 mm. At this point, theexperimental protocol described in the previous series was repeated. The pipette was then movedto the contralateral side and the same volume of vehicle or KA was microinjected into theequivalent site on that side of the brain and the experimental protocol followed once again.Page 197Figure 6.2: An ENG recording illustrating the short term effects of kainic acid injection on thebursting pattern recorded from the vagal (X) and trigeminal (V) nerves.Page 198Kainicacid(175 nI,4.7mM)1x...L1ihLliIIndillaililitia111111]L1U.L1lThFJ__V1.•‘‘.Ii1lLAi.Iivi,vmwww..y...-1VVWW.1.IY!lT!rThfJirrtiri....—-—-—-•_______“_I11,11II’,‘,PT‘I‘!IIThIJ,,IJII.IITII10sec(t ‘0In four animals, a similar volume of lidocaine hydrochloride was microinjected into thisregion following the sham microinjection to demonstrate the effects of ‘progressive’ andreversible lesion of the NI. Kainic acid was then injected into this site after the effects of thelidocaine were completely dissipated (—30 mm). In all experiments, after each KA injection,an equal volume of fast green was injected for subsequent localization of the injection site.Series III: the effects of vagotomy after bilateral lesions of the nucleus isthmi.Because it is well established that the interaction between vagal and CNS inputs plays alarge role in ventilatory pattern formation in mammals (for review, see Milsom, 1991), thisseries of experiments compared the breathing patterns of 4 frogs in which the NI had beensevered before and after the vagus nerves were cut bilaterally at the cervical level. Sincebilateral microinjections of KA significantly reduced breathing (see results section below), thecomparison could only be made at the highest level of CO2 used in this experimental protocol.Analytical procedures.The procedures for administering and measuring gas mixtures, and for quantifyingtrigeminal and vagal burst duration and all other fictive breathing pattern variables have beendescribed in the previous chapter.Page 200Brain tissue histology.At the end of the experiment, the brain was removed from the bullfrog, and fixedovernight in a cold 4% paraformaldehyde solution in phosphate buffer neutralized to pH = 7.9.The tissue was then transferred to a 30% sucrose, 0.1% sodium azide solution for 24 h beforesectioning (50 m) on a cryostat at -23°C. The sections were then stained either with cresylviolet or eosine-hematoxylin according to standard procedures, and examined for fast greenstaining to determine the sites of the micropipette placements. Histological data was obtainedfrom 6 bullfrogs.Data analysis.Values for fictive breathing variables were obtained by analyzing a 2-mm segment ofdata. All data are presented as means ± SE. The results were statistically analyzed using atwo-way analysis of variance followed by a Student-Newman-Keuls test (p < 0.05).Page 201RESULTSThe fictive breathing pattern recorded in control bullfrogs and in bullfrogs aftermicroinjection of saline into the NI was similar to that reported in chapter 4. Briefly, thisbreathing pattern was episodic, and the number of breaths in each fictive episode variedaccording to the respiratory drive, while the incidence of the episodes was directly proportionalto the level of PSR feedback. In some preparations, fictive buccal oscillations were recordedfrom the trigeminal nerve between the fictive episodes of lung ventilation (Fig. 6.3A).1: The ‘fictive” breathing pattern of bulifrogs under conditions of normo-. and hypercarbicrespiratory drive at different tonic levels of PSR feedback.Because the changes in fictive breathing pattern observed after changing PSR andchemoreceptor afferent feedback were very similar to those described in chapter 4, thedescription of these observations will be concise, and simply outline the major effects observedalong with the similarities/differences between both studies.1.1. The effects of increasing hypercarbic respiratory driveIn sham microinjected bullfrogs, the effects of changing hypercarbic respiratory drivewere the same as those reported in chapter 4. Thus, increasing the FO2 of the UDV gasmixture stimulated fictive breathing (p < 0.001) (Fig. 6.4). The increase in ventilatory motoroutput was mediated by an increase in breathing frequency alone (p = 0.021), since the VTjfldeXdid not change significantly with the increase in respiratory drive (p = 0.87). The increase inbreathing frequency observed under hypercarbic conditions was caused only by an increase inthe number of breaths in each bout of fictive ventilatory activity (p < 0.033), since the numberof episodes of breathing per minute was unchanged (p = 0.187). The latter result differed fromPage 202the response obtained in chapter 4, where increasing hypercarbic respiratory drive shortened theduration of the non-ventilatory period and increased the number of breathing episodes perminute. This difference is only relative, however, since there was a trend for the number ofepisodes per minute to increase in animals with PL = 0 cm H20 in this study and episodes fusedand breathing became continuous as FCO2 increased in animals with PL = 2 or 5 cm H20.Increasing hypercarbic respiratory drive had no effect on trigeminal or vagal burstduration (p = 0.397 and p= 0.593, respectively) (Fig. 6.6), or fL1 (p = 0.05) (data notshown), but reduced (p = 0.009) (Fig. 6.8). These results differ from the data reportedin chapter 4 which showed that FCO2 shortened both, thgeminal and vagal burst duration andalso reduced1.2. The effects of tonic changes in PSR feedback.Tonic increases in PL stimulated fictive breathing. For each level of respiratory drive,increasing PL increased breathing frequency, reduced the peak integrated trigeminal burst activity(VTifldex) and the net result was an increase in the Vlndex (p < 0.001 for all) (Fig. 6.7A,B and C,respectively). The increase in fictive breathing frequency produced by increasing PL wasmediated by a change in breathing pattern. In bullfrogs on UDV with 2% and 3.5% CO2 in air,increasing PL above 0 cm H20 transformed the breathing pattern from episodic to continuousin most animals. In bullfrogs that did not breath continuously, however, increasing PLaugmented the number of breathing episodes per minute (p = 0.003), but had no effect on thenumber of breaths in a breathing episode (P = 0.187) (Fig. 6.7E and D, respectively). Theeffects of PL on fLabs were directly proportional to the magnitude of the hypercarbic respiratorydrive, as there was a significant interaction (p = 0.036) between both variables.Page 203Figure 6.3: A comparison of the effects of bilateral microinjections of saline (panel A),lidocaine hydrochloride (panel B) and kainic acid (panel C) into the NI on the fictivebreathing pattern of decerebrate, paralysed, unidirectionally ventilated bullfrogs, asshown by a recording of integrated trigeminal nerve activity ( V).Page 204lv.i11..L1...111I.i.ii..1!ik..EliiLL..1Lii1IikiuiiIU,.i .1.11.11Lidocaine(175n1,1%)BJrJJliiJi1hiiij•‘M-‘f.Sla1LrrLJwrL_Saline(175nI)AlvKainicacid(175 nl,4.7mM)lvJr20mmfJQ (p CC10secFigure 6.4: The relationship between hypercarbic respiratory drive (FCO2) and A) ventilatoryindex, B) absolute breathing frequency, C) peak integrated trigeminal nerve activity, D)breaths per episode and E) episodes per minute in bullfrogs with (closed symbols) andwithout their nucleus isthmi intact (open symbols). Each curve represents a differentlevel of tonic PSR feedback (circles: PL = 0 cm H20; squares: PL = 5 cm H20).Page 206Breaths/episode jDEpisodes/minuteE---.—-I0. activity(Volt*sec)20 15 I0 5 0-0.6F00.00.01I0.—0.01.-..__j—.——-—--——.—-—.----..—I0.020.030.048 6 415 I0 5 00.0l0.02_._.,__.I____0.030.04FcoTonic increases in PL had no significant effect on the duration of the vagal bursts (p =0.191). This means that when PL was equal to 0 cm H20, the trigeminal burst began at thesame time or before the onset of the vagal burst, and both nerves became silent at approximatelythe same time. With increases in PL, however, trigeminal burst duration was shortened (p <0.001) such that the fraction of the gas exchange cycle during which gas would be expelled fromthe lungs was prolonged and the lung inflation phase was shortened as in chapter 4 (Figs 6.5,6.6 A and C). Changes in PL had no effect on either (data not shown) or (Fig. 6.8).2. The effects of bilateral lesions of the nucleus isthmi on breathing pattern and ventilatoiyreflexes.Three types of ventilatory responses were observed after bilateral injection of KA intothe area of the NI; in none of the cases, however, was a change in blood pressure associatedwith the microinjection of KA. In 6 animals, the bilateral injection had very little or nosignificant effect on the breathing output. The anatomical location of the fast green markers inone of these cases is shown by the open circles in Fig. 6.9, which represents a cross section— 4mm rostral to the obex. In this particular frog, only one of the markers was located in theNI, whereas the other was situated —200 m below the NI. In 2 experiments, the respiratory-related motor output remained silent 1.5 h after the KA injections. Histology was done on thebrain of one of these animals, and it was found that the markers were located —50 tm abovethe NI. None of these 8 experiments were considered “successful” and they were not includedin the data analysis.In 53% of the trials (9 bullfrogs), both fast green markers were located inside or in theimmediate vicinity (< 40 tim) of the NI bilaterally (filled circles, Fig. 6.9). These “successful’injections caused the important changes in breathing pattern which are shown in Fig. 6.3C. ThePage 208ENG recordings shown in Fig. 6. 3B demonstrate that bilateral NI lesions caused by lidocaineprovoked important changes in breathing pattern and an overall reduction in the rate of fictivebreathing (Fig. 6. 3B and C; 6.4, 6.7). In saline microinjected frogs, the fictive breathingpattern was either episodic or continuous as described previously, depending on the level ofrespiratory drive. After lidocaine injection, however, the breathing pattern consisted mainly ofevenly-spaced single breaths, and there were fewer fictive lung breaths per minute. Fictivebuccal oscillations were often observed during the non-ventilatory period. The effects oflidocaine usually lasted for approximately 20 miii, after which breathing began to accelerate, andeach episode of fictive breathing had progressively more breaths.As illustrated by Fig. 6. 3B, bilateral KA injections had similar effects to those justdescribed for lidocaine injections. An important distinction, however, was that the effects ofKA were not reversible, and after the initial cycle of stimulation/inhibition, breathing was stablefor the entire duration of the experiment. As indicated by the trigeminal and vagal burstingpattern, bilateral lesion of the NI prolonged the duration of the ventilatory cycle (p = 0.004),but had no significant effect on the relative duration of the inspiratory phase (p = 0.217) (Fig.6.5, 6.6B and D).When comparing the fictive breathing recorded from bullfrogs before and after NIlesions, the most striking difference was the overall reduction in ventilatory output (p < 0.001)(Figs. 6.3, 6.4A, 6.7A). The latter was mediated only by a reduction in fLabs (p < 0.001), sincebilateral injection of KA had no effect on the VTjfldex (p = 0.073) (Fig. 6.4C). When comparingthe breathing pattern data of both groups, the reduced fL21, after bilateral NI lesions was causedby a reduced number of fictive breaths per episode (longer non-ventilatory pause; p < 0.001)and less episodes of breathing per minute (p < 0.001). These data reflect the evenly-spacedPage 209single breath pattern of fictive ventilation typically observed after bilateral NI lesions, andexplains why there were not sufficient data to quantify fL1 in this group.Page 210Figure 6.5: Simultaneous ENG recordings from the vagal (top trace) and trigeminal (bottomtrace) nerves which illustrate and contrast the bursting activity of a “fictive breath”typically recorded before (panel A) and after bilateral KA microinjections (panel B).These recordings were obtained from a single bullfrog on UDV with 3.5% CO2 in airat PL = 5 cm H20. Dotted lines indicate 1) the start of the vagal burst, 2) the start ofthe trigeminal burst and 3) the approximate termination of both vagal and trigeminalbursts. This particular case also illustrates the reduced definition of the vagal burst afterNI lesions.Page 211Intact Ax II0!’ JIrmIVr” ‘-•Nucleus isthmi lesion BI - :1xV*aInm1h I1JrthaflFir.1ITI0.5 secPage 212Figure 6.6: The effects of changing lung pressure on the duration of the vagal (open circles)and trigeminal (closed circles) burst duration at two levels of hypercarbic respiratorydrive in bullfrogs with (panels A and C) and without their nucleus isthmi intact (panelsB and D). Note that when vagal burst duration equals or is less than trigeminal burstduration (stippled area), no lung deflation would occur in spontaneously breathingbullfrogs. Conversely, the greater the extent to which vagal burst duration exceedstrigeminal burst duration, the more time spent in expiration (hatched area).Page 213A1.4 1.2 1.0 1.2 1.0 0.8- 1.4 1.2 1.0 I—IL_I__-01234Lungpressure(cm1120)I 5L 012345Lungpressure(cm1120)Figure 6.7: The relationship between tonic changes in lung pressure and A) ventilatory index,B) absolute breathing frequency, C) peak integrated trigeminal nerve activity, D) breathsper episode and E) episodes per minute in bullfrogs with (closed symbols) and withouttheir nucleus isthmi intact (open symbols). Each curve represents a different level ofhypercarbic respiratory drive (circles = 1 % CO2 in air; squares = 3.5% CO2 in air).Page 215Ventilatoryindex(arb units)20Absolutebreathingfrequency(min)45 30 15 0(LBreaths/episodeDEpisodes/minuteEAB15 10 5 0//CPeakintegratedtrigeminalburst activity(Volt*sec)—0123458 6 4 2 015 10F50I234Lungpressure(cmH20)501234Lungpressure(cmH20)-J 5Figure 6.8: The effects of changing lung pressure on instantaneous buccal oscillation frequencyat two levels of hypercarbic respiratory drive (circles = 1 % CO2 in air; squares = 3.5%CO2 in air) in bullfrogs with (closed symbols) and without (open symbols) their nucleusisthmi intact.Page 217Instantaneous buccal1oscillation frequency (mi& )30150I I I I0 1 2 3 4 5Lung pressure (cm H20)Page 218Figure 6.9: Drawing of a cross section of the bullfrog brain 4mm rostra! of the obex showingsites in the nucleus isthmi (NI) area where bilateral microinjections of kainic acid wereperformed. In four bullfrogs, the injections induced important changes in breathingpattern (closed symbols). In two other cases, however, kainic acid microinjections eitherhad no effect (open circles) or completely eliminated bursting activity (open triangles).NI: nucleus isthmi; vm: ventriculus mesencephali.Page 219I- I400 jtmPage 2202.1. The effects of hypercarbic respiratory drive after bilateral NI lesions.Fictive breathing was not recorded in 63 % (5 out of 8) of the bullfrogs on UDV with a1 % CO2 in air gas mixture at any level of PL following bilateral NI lesions, whereas all intactfrogs produced respiratory related motor output under these conditions. Fictive breathing wasrecorded in all bullfrogs on UDV with 3.5% CO2 in air after bilateral NI lesions. However, theventilatory index was still much lower (p < 0.001) than that of intact bullfrogs (Fig. 6.4A).The lack of fictive breathing at lower CO2 levels reflects the overall reduction in responsivenessto chemostimulation that was observed in KA microinjected bullfrogs.Increasing hypercarbic respiratory drive following KA microinjection had no significanteffect on the breathing pattern except at PL = 5 cm H2O and FCO = 0.035, since the numberof breaths per episode and the number of episodes per minute remained unchanged at all othercombinations of PL and FCO (i’ = 0.269 and p = 0.098, respectively) (Fig. 6.4D and E).Table 6.1 reports the number of breaths in a breathing episode that were observed after bilateralNI lesions in bullfrogs on UDV with 3.5 % CO2 in air at PL = 5 cm H20. These data showthat, in two animals (E and I), no breathing was recorded under these experimental conditions,while in bullfrogs F and G, breathing episodes could still occur after the bilateral NI lesions; thenumber of breaths in each episode was a function of FCO2. All other bullfrogs took singlebreaths or, on average, pairs of breaths under these conditions. The breathing pattern that wasrecorded for animal G is shown in Fig. 6.10. This recording also demonstrates that, in thiscase, bilateral vagotomy reduced the breathing frequency and the number of breaths in anepisode to one. It is noteworthy that the amplitude of buccal oscillations was high immediatelyafter a lung breath, and progressively decayed before the onset of the next one.Page 221Changing the FCO2 of the UDV gas mixture also had no effect on the vagal andtrigeminal burst durations (p = 0.532 and p = 0.326, respectively) (Fig. 6.6B and D). Thisprocedure did, however, significantly reduce the fB1 (p = 0.009) (Fig. 6.8).2.2. The effects of changes in PSR feedback after bilateral NI lesions.An important feature of the fictive breathing recorded after bilateral NI lesions was thattonic increases in PL no longer had any significant effect on fLabs (p = 0.054), Vlndex (P =0.472), the number of breath per episode (p 0.242) or the incidence of fictive episodes ofbreathing (p = 0.559). It should be mentioned, however, that in 4 bullfrogs on UDV with 3.5%CO2 in air, increasing PSR feedback increased the average number of breaths in a breathingepisode above 1. The VTjfldex was the only variable consistently affected by changes in PSRfeedback (p = 0.046) (Fig. 6.7C).Changes in tonic levels of PSR feedback could still affect the timing of the fictive breathsin KA microinjected bullfrogs, since progressive increases in PL shortened trigeminal burstduration (p = 0.002) but, contrary to bullfrogs with their NI intact, it also prolonged theduration of the vagal bursts (p = 0.002). This means that at higher PL’s, frogs without theirNI would have a greater proportion of their breathing cycle devoted to expiration, andconversely, a shorter inspiratory phase than frogs with their NI intact. It is noteworthy that, insome instances, the vagal bursts were not well defined after NI lesions. In these cases, it wasdifficult to accurately determine the onset and termination of a burst (Fig. 6.5). Increasing PLhad no effect on fBIL (p = 0.235).Page 2223. The effects of bilateral vagotomy after bilateral lesions of the nucleus isthmi.The results of this experiment indicate that the effects of complete elimination of vagalfeedback on breathing pattern were not different than those observed upon lung deflation. Forall fictive breathing variables recorded, the data for vagotomized bullfrogs were neversignificantly different from the values recorded at PL = 0 cm H20 (Fig 6.11).Page 223Table 6.1: Individual values for the number of fictive breaths in a breathing episode that wererecorded after NI lesions from bullfrogs on UDV with 3.5% CO2 in air at PL = 5 cmH20.Page 224Bullfrog # Average number ofbreaths per episodeA 2B 1C 1.1D 1EF 18G 11H 2I-Note that no breathing was recorded for bullfrogs E and I.Page 225Figure 6.10: Simultaneous ENG recordings of full wave-rectified, integrated vagal (S X) andtrigeminal ( 5 V) nerve activity obtained from bullfrog ‘G” illustrating the occurrence ofbreathing episodes after bilateral lesions of the NI. These recordings were obtained atthree levels of respiratory drive: 1 % CO2 in air (Panel A), 2% CO2 in air (Panel B) and3.5% CO2 in air (Panel C). Panel D shows the effects of bilateral vagotomy in thisparticular animal when on UDV with 3.5 % CO2 in air (Panel D). Note the decay patternof the small-amplitude buccal oscillations between each fictive breath post-vagotomy.Page 2261% COJX A—-- -b, 1f — -. — — -t -Sv_.__a_ —ut_ — .—— ———— — ——2% CO,Ix B-3.5% CO2fx CA IL MAIL‘V--.---- ,.r-t-- - - -‘-• - -— —n- •d-’3.5% CO2 post vagotomy.Ix Dfv______________________ ____________________________-- . .10 secPage 227Figure 6.11: A comparison of the relationship between tonic changes in lung pressure and A)ventilatory index, B) absolute breathing frequency, C) peak integrated trigeminal nerveactivity, D) breaths per episode and E) episodes per minute before (closed symbols) andafter (open symbols) bilateral vagotomy in bullfrogs without their nucleus isthmi. In thisstudy, the bullfrogs were on UDV with 3.5% CO2 in air.Page 228ABCVentilatoryindex(arbuniLs)5r 4 3 2012345Absolutebreathiigfrequency(mm)15r5Peakintegratedtrigeminalburstactivity(tVoIt*sec)’0123450123458 6 4 2 06 S 4 3 201234Lungpressure(em1120)5DISCUSSION1. Critique of methodsDiscrepancies between the data obtained for bullfrogs with their NI intact in the presentchapter and data obtained from animals in chapter 4 were all related to the effects of hypercarbicrespiratory drive (see section 1.1.) on breathing pattern. Overall, the sham injected frogs inseries I of the present study behaved and responded to changes in PL and/or respiratory drivein a fashion that closely resembled the responses described in chapter 4. However, at lowlevels, the effects of hypercarbia in the present study did not always elicit changes in breathingpattern that were statistically significant while they did in the experiments in chapter 4. At thehigher levels of respiratory drive, breathing was often continuous in the present study, and thus,it was not possible to obtain values for the number of breaths per episode or episodes per minutefrom many animals to compare with the data in chapter 4. Hence, the difference seems to beless sensitivity at low levels of hypercarbia, but more sensitivity at high levels of hypercarbiain the present study.As described by Larsell (1923) and Senn (1972), the NI has a particular pattern oforganization. The cells are not equally distributed within the nuclear field, but are denselyarranged at the periphery of the nucleus. The centre of the NI contains but a few scattered cells;it consists mainly of fibers. Larsell (1923), therefore, divided the nucleus into a “cortex”(peripherally) and a “medulla” (internally). Given the large size of this structure (radius 200tm) and its arrangement, a relatively large volume of KA or lidocaine had to be injected inorder to reach a significant portion of the cell bodies situated at the periphery. This particulararrangement of the NI made it difficult to focus the effects of the pharmacological agents, andmay explain why, in some cases, microinjections of similar volumes in the same area did notalways affect the respiratory-related motor output. Injection volumes could not be too large,Page 230however, since in preliminary experiments, bilateral injections of KA caudal to the NI sometimesresulted in loss of ENG activity from the trigeminal nerve, while vagal activity remainedunaffected; suggesting that the KA was diffusing to the trigeminal motor nucleus. Estimationof diffusive spread of the volumes of KA injected here are well within prediction based onNicholson’s model (1985) for diffusion in brain tissue. Briefly, this model predicts thatfollowing a 10 nI injection, the concentration at 200 .tm from the injection site will be— 35%of the initial value after 10 secs and 5% of that value after 2 to 3 mm. The use of a largervolume, such as the one used in the present study, means that more cells will be reached morequickly, and thus produce a faster response. Based on this argument, it will be assumed thatthe effects of KA microinjection on breathing pattern that were observed reflect bilateral lesionsto the NI.2. Effects of bilateral lesions of the nucleus isthmi.Following bilateral microinjections of lidocaine into the NI, fictive breathing frequencywas immediately reduced and the breathing pattern consisted of evenly-spaced single breaths fora period of roughly 20 mm, without affecting the VTlndex. Lidocaine is a local anesthetic whichaffects both axons of passage and cell bodies. The fact that bilateral microinjections of kainicacid, a potent glutamate analogue which only affects cell bodies, also resulted in a severereduction in fictive breathing frequency, and, for the most part, changed the breathing patternto one of evenly-spaced single breaths, indicates that these changes in ventilatory output werethe result of bilateral lesions to the NI, rather than disruption or impairment of neuronalpathways in this brainstem area. These effects of bilateral lesions to the NI are similar to theeffects of vagotomy or lung deflation, and/or reduction of respiratory drive reported previously(chapters 3 and 4, present study). They are also similar to the results of Naifeh et al. (197 la,Page 231b) which have shown that, in alligators and caimans, both mild anesthesia (chloroform orpentobarbital) and brainstem section could convert the episodic breathing pattern normally seenin these animals into a single breath pattern. In their study, the typical breathing episodesdisappeared early during the induction into anesthesia while brainstem transection also reducedthe number of breaths in each breathing episode to a single breath.2.1. Effects on hypercarbic resnonse.One of the most striking feature of the breathing pattern following lesions to the NI wasthe reduced responsiveness of the preparation to the hypercarbic stimuli. This reduction insensitivity was roughly proportional to the overall reduction in total ventilation. This leads tothe conclusion that the frog NI is an endogenous source of glutamate that maintains eucapnicmotor output and allows full expression of the CO2 response.This reduction of eucapnic ventilation and CO2 chemosensitivity is similar to the effectof retrotrapezoid nucleus (RTN) lesions in cats (Nattie and Li, 1990; Nattie et at., 1991). TheRTN is a region of the rostral ventrolateral medulla which may represent the most rostralextension of the column of respiratory neurons in the ventral medulla in mammals (Nattie andLi, 1995). In these studies, however, unilateral lesions alone were effective and they affectedonly the phreriic amplitude response, and had no effect on the fictive breathing frequency. Sincethe ventilatory response of cats to hypercarbia was predominantly mediated by an increase intidal volume, however, with breathing frequency remaining relatively constant under thesecircumstances, the net effect was a large reduction in sensitivity to hypercarbia (Schlaefke et at.,1979). Based on their data, these workers concluded that the role of the retrotrapezoid area wasin the regulation of tidal volume rather than breathing frequency. They also concluded that thisarea in mammals facilitated eucapnic ventilation and allowed full expression of CO2-Page 232chemosensitivity. Although the RTN of mammals and the NI of frogs are not homologousstructures, the analogous effects of lesions in these areas is striking.2.2. Effects on lung pressure response.The effect of increasing PSR feedback on fictive breathing frequency was greatlyreduced; hence, the shortening of the non-ventilatory pause usually provoked by increasing PSRfeedback did not occur. This was somewhat predictable given that PSR feedback andchemoreceptor input were shown previously to interact to modulate fLabs and breathing pattern(chapters 4 and 5), and that the most important effect of the kainate lesions in the present studywas the reduced responsiveness of the frogs to C02/W. Thus, the only effect of changing PLon breathing frequency occurred at the highest PL (5 cm H20) and the highest level ofchemoreceptor drive (FCO2 = 0.035). Bilateral NI lesions had no significant effect on thechanges in VTIndex elicited by PSR feedback, thus indicating that this reflex does not require NIinput, and that the role of the NI is limited to integrating inputs pertaining to respiratory rhythmrather than bursting pattern.3. Significance of buccal oscillation and lung ventilation frequencies.The present data has shown that the NI plays a role in chemoreceptor reflexes modulatinglung breathing frequency. The fact that fB1 remained the same and was still affected by CO2after NI lesions suggests that the role of the NI in chemodetection is restricted to lungventilation, and that mechanisms modulating the frequency of lung ventilations and breathingfrequency are distinct. Since, in the preceding chapter, changing mock CSF had nosignificant effect on fB in vitro, it is possible that this type of buccal movement is modulatedexclusively by peripheral chemoreceptor input.Page 2334. Physiological signficance. the role of the nucleus isthmi.The breathing pattern observed after NI lesions was very similar to that ofpremetamorphic tadpoles, which consists of single air breaths interspersed with gill ventilations,which are thought to be the precursors of buccal oscillations. As for obligate air-breathingpremetamorphic tadpoles (Taylor-Kollors stage XVI-XIX), increasing CO2 caused an increasein the frequency of air breaths, without any episodes, and reduced the buccal oscillationfrequency (Infantino; 1989). The similarity of the breathing pattern of tadpoles and adultbullfrogs post NI lesions is consistent with the correlation between the onset of episodicbreathing in bullfrogs and the developmental changes in the NI (Senn, 1972) that occur atmetamorphosis, and suggests that the emergence of CO2 chemosensitivity near metamorphosisis an important factor in the production of breathing episodes.The transformation of the breathing pattern from episodic to evenly spaced single breathswas always associated with a reduction in breathing frequency rather than a change in thetemporal distribution of breaths. This raises the possibility that the NI is responsible forproducing clusters of breaths or episodes or that the NI simply provides a tonic input and thatepisodes only occur when respiratory drive is high. Since breathing episodes were still observedthe two animals following microinjections of KA in the area of the NI (Table 6.1; Fig. 6.11),it would appear that the NI only provides tonic respiratory drive, and in its absence, the levelsof hypercarbia and PSR feedback utilized to stimulate breathing in this study were not sufficientto produce episodes except in two frogs at the maximum level of stimulation (F2 = 0.035;PL = 5 cm H20). Thus, the NI does not appear to be directly responsible for turning breathingepisodes on and off.In mammals, it has been shown that either vagotomy or lesions to the pontine respiratorygroup (PRG) slow breathing frequency and increase tidal volume, and that bilateral vagotomyPage 234after PRG lesions produces apneusis or very slow deep breaths, depending on the level of thelesion (Feldman and Gautier, 1976). These results suggest that both vagal feedback and thePRG provide a tonic input to eupneic breathing, and therefore, play an important role in patternformation. Bilateral lesions to the NI in bullfrogs had a similar effect on breathing frequencyto PRG lesions in mammals, and in some cases, subsequent vagotomy slowed breathing further.Although apneusis was not observed in this frog preparation; i.e. there were no prolongedinspiratory efforts, there is an interesting parallel between the influence of the NI and the PRGon breathing pattern.In summary, the present data demonstrate that bilateral lesions to the NI cause areduction in respiration and in the sensitivity to hypercapnia and changes in PL, roughlyproportional to the overall reduction in total ventilation. These findings lead to the conclusionthat, in the bullfrog, the NI is an endogenous source of glutamate that maintains eucapnic motoroutput and allows full expression of the CO2 response.Page 235CHAPTER 7:GENERAL CONCLUSIONPage 236Given that the data, and the methods used to obtain it, have already been discussed indetail in each of the previous chapters, I will conclude by discussing the data in the broadercontext of concepts pertaining to the control of breathing, with a strong bias towards patternformation. This will be an amalgamation of the knowledge that I have gathered from mylaboratory work, data reported by other workers, and “intuitive inference”. Owing to the lackof neuroanatomical and neurophysiological data on respiratory control in amphibians, thisdiscussion will be primarily conceptual. Nonetheless, I hope it will help generate new ideas andresearch directions.1. The role ofperipheral feedback in the control of breathing in bulifrogs.As it has been pointed out throughout the thesis, although feedback from peripheralreceptors did contribute to the shaping of the breathing pattern, none of the receptor groupsinvestigated could account for either respiratory rhythmogenesis or the clustering of the breathsinto distinct episodes. This was firmly established in chapter 5 where it was shown that therespiratory-related motor output produced by an in vitro brainstem-spinal cord preparation wasepisodic, despite the fact that this preparation was essentially devoid of peripheral feedback. Itshould be stated that in all studies, increasing respiratory drive increased the probability ofbreathing occurring. Hence, one could argue that the drive to breathe turns breathing on; thisis undeniably true. Because a tonic respiratory drive still produces episodes of breathing,however, we still have no explanation for what turns breathing off, and consequently, what isresponsible for producing this pattern of breathing. This indicates that the mechanismsunderlying episodic breathing are an intrinsic property of the central respiratory control system.Page 2372. Central rhythm generation in bullfrogs.Current models of respiratory control in lower vertebrates have put emphasis on therelationship/interaction between buccal oscillations and lung ventilations. The buccal oscillationsin adults are believed to be residual manifestations of gill ventilations in pre-metamorphictadpoles. Because both types of events share at least one muscle group (buccal levators), it hasalways been assumed that they must also share control mechanisms. An understanding of therelationship between these two types of buccal movement may offer insight into the pivotal stepsin the evolution of air breathing (Smatresk, 1990). The first question that arises in this regardis whether buccal oscillations and lung ventilations are driven by one or two rhythm generators?Both types of events are produced in a very rhythmic fashion, and instantaneousfrequency measurements have been presented in the different chapters as indicators of theintrinsic rhythms underlying buccal oscillations and lung ventilation. In chapter 2 and 3, it wasshown that in “resting” animals, buccal oscillations occur continuously during the non-ventilatoryperiods, and that lung ventilations normally occur at a time when another buccal oscillationwould have been initiated. Since increasing respiratory drive increased fLabs without affectingit would appear as though the stimuli simply “facilitated” the expression of the underlyingrhythm. Under most experimental conditions, both instantaneous frequencies were similar,suggesting that either the two types of events were produced by a single CR0, or that they wereclosely entrained. For instance, under normoxic normocapnic conditions, both instantaneousfrequencies were similar, whether the animals were intact (chapters 2 and 3) or decerebrated(chapters 6 and 7), with or without intact vagi (chapter 3), with or without olfactory receptorfeedback (chapter 3), whether PSR feedback was tonic or phasic (chapter 4), and whether tonicPSR feedback was high or low (chapter 4). Making bullfrogs hypoxic also had little effect onthe and values, which were very similar (chapter 2). There were circumstances,Page 238however, which dissociated the rhythms of the two breathing events. In some studies, wasmarginally slower than fB1, (e.g. in vitro preparation; chapter 5); however, in all preparationshypercarbia consistently suppressed buccal oscillations without having much effect onWhen could be measured under hypercarbic conditions, however, the mean values weretypically less than the normocarbic values, while fLIflSL little affected. This data suggests,therefore, that under some circumstances, the two rhythms are different and can be modulateddifferently.Finally, it has recently been shown that in tadpoles, gill ventilation appears to be drivenby a network oscillator while lung movements appear to depend on a system involving someform of pacemaker cells or inhibitory interactions that are not dependent on chloride ions (Packet al., 1992). This data further supports the idea that the two events are generated by differentmechanisms which are often entrained, and are not produced by a single CRG.3. Central pattern generation in bulfrogs.In many instances, the data reported in the thesis has shown that if the convectiverequirements of the bullfrog are met artificially, breathing is absent. Consequently, it wasargued that if a CRG was part of the respiratory control system of the bullfrog, it’s activity wasconditional upon a minimum tonic level of respiratory drive (chapter 2). Once that minimumlevel of drive is provided to the control system, breathing occurs, and the output may or maynot occur in episodes, even in the absence of peripheral feedback (chapter 5). The fact that thefictive breathing pattern produced by the in vitro brainstem-spinal cord preparation was virtuallyidentical to that of intact bullfrogs certainly questions the need for a central pattern generatorin our modelling of respiratory control in the bullfrog. It is important at this juncture toreiterate the distinction between the concepts of respiratory rhythm and pattern. The effects ofPage 239hypercarbia on the fictive breathing pattern recorded in situ (chapter 4) will help make thispoint.At high levels of hypercarbia, bullfrogs breathe continuously, and the motor output ofthe preparation occurs in a very “rhythmic” fashion. When the hypercarbic respiratory driveand/or the level of PSR feedback was reduced, breathing became episodic. Even then, theperiod between each fictive breath within an episode was fairly constant, as were the small-amplitude fictive buccal oscillations produced between the episodes. There is more to breathingpattern than respiratory rhythm, however. For instance, increasing PSR feedback reduced theamplitude of the fictive breaths, prolonged the duration of the expiratory phase and shortenedthe inspiratory phase without affecting the duration of the entire cycle (TT0T). Increasing PSRfeedback also reduced the duration of the non-ventilatory period. Conversely, increasing FCO2shortened the total duration of the breathing cycle (TT0T) without affecting the relative durationof the expiratory and inspiratory phases, and increased the number of fictive breaths in eachepisode. Under different circumstances, there are many different components of the motoroutput that can be changed relatively independently from the respiratory rhythm. Thesecomponents of the motor output also constitute part of the breathing pattern. In spontaneouslybreathing bullfrogs, processing of all afferent inputs pertaining to breathing is an essential steptowards the production of a pattern of motor output that meets the metabolic demands. This isthe basis for the concept of having a central pattern generator as a “functional” entity, distinctfrom the respiratory rhythm generator, in the respiratory control system.Tadpoles primarily modulate gill ventilation to accommodate the demands imposed onthe respiratory system, while in adult bullfrogs, both buccal oscillations and lung ventilationsoccur, but are modulated differently by hypercapnia. These observations have raised thequestion of whether these independent changes are the product of separate central patternPage 240generators. The results of the experiments reported in chapter 6 have shown that, in adultbullfrogs, bilateral kainic acid lesions in the area of the NI had profound effects on CO2sensitivity, and consequently, the breathing frequency was significantly reduced. After lesionsin the NI area, the breathing pattern consisted mainly of evenly-spaced single breaths. Thissuggested that input from the nucleus isthmi provided a tonic level of respiratory drive, andinfluenced the integration of afferent inputs affecting ventilation. These lesions, however, hadno effect on the occurrence and the intrinsic frequency of the buccal oscillations, nor on theinhibitory effects of CO2 on This indicates, therefore, that the signal from chemoreceptorsis processed differently for both types of events, and that input from the NI does not modulatethe reflexive reduction of fB1 induced by CO2. Because was not altered when centralchemoreceptors alone were stimulated by increases in CO (see in vitro data, chapter 5), itwould appear that the inhibitory input to the buccal oscillation network is provided by peripheralchemoreceptors which respond to increases in arterial levels of CO2/H and/or decreases in Pa02(Van Vliet and West, 1992). The absence of this inhibitory input may explain why fB1 wasmuch larger than in the in vitro preparation (chapter 5). This inhibition of the buccaloscillation network would also explain the gating” between the two types of events whichoccurs when the level of drive becomes elevated.Buccal pumping in tadpoles and buccal oscillations in adults appear to be produced bythe reciprocal inhibition of two antagonistic muscle groups: the buccal levators and depressors.In the case of their inhibition by hypercarbia, it is conceivable that the afferent chemoreceptorsignal impinges exclusively at the level of the premotor or motoneurone pool, where asintegration of hypercarbic stimuli appears to occur higher in the brainstem for lung breathingin the adults. Based on the defmition of a CPG that was put forward originally, the fact thatPage 241both buccal oscillations and lung ventilations can be modulated differently by different afferentinputs, suggests that there are two CPGs as well as two CRGs.4. What turns breathing on and off?The data gathered throughout this thesis have eliminated numerous mechanisms whichhave been postulated over the years to produce breathing episodes. The results have alsoindicated that the answer to this question more than likely lies in understanding some of the basicneurophysiological properties of the central respiratory neurones. That breathing occursintermittently, even when all afferent inputs are tonic, suggests that the membrane potentials ofcertain groups of respiratory-related neurones must oscillate. Tonic inputs, from the NI perhapsas well as peripheral receptors, could raise the membrane potentials of these respiratory neuronesallowing an intrinsic rhythm to be expressed until these cells became quiescent or refractory.The cause(s) of this latter phenomenon underlie the termination of a breathing episode. To beconsistent with the empirical findings, increasing different tonic drives must also be capable ofaltering the period of “repetitive bursting”, and reducing the duration of the refractory periodof the respiratory neurones between episodes differently. This would correlate with differencesin the alternation between breathing episodes and non-ventilatory pauses which occur in the faceof different tonic, elevated respiratory drives.This type of neuronal behaviour can be produced in a number of ways, but the recentfindings of Liao et al. (1994) suggest that this process may be related to afterhyperpolarizationcurrents produced by Ca2-activ ted K channels. Their study has shown that adding apamine(a pharmacological agent which specifically blocks these K channels) to the solutionsuperfusing the tadpole brainstem-spinal cord preparation does not disrupt the rhythmicbehaviour of the fictive gill ventilation or the lung ventilation. What is interesting, however,Page 242is that application of this drug prolongs the duration of the lung bursts. Since this researchgroup does not distinguish between breathing episodes of one or many breaths but considers allbreathing episodes as a single lung event (Filmyer, personal communication), it is possible thattheir results indicate an increase in the number of breaths per episode. If this is the case, andif this occurred at a constant respiratory drive, it could very well be that increasingrecruitment/derecruitment of these channels on key respiratory neurones is responsible for thegenesis of breathing episodes.Lower vertebrates, such as the bullfrog, have often been considered as being simpler ormore “primitive” in terms of their respiratory control system. The data obtained throughout thethesis strongly suggests that, more often than not, the respiratory control system is highlyconserved amongst vertebrates. For instance, basic ventilatory reflexes, such as chemoresponsesand the Hering-Breuer inspiratory-inhibiting reflex are present in the frog as well as inmammals. Even at the level of the central control mechanisms, there appears to be a highdegree of similarity with respect to the anatomical projection of afferent nerve inputs amongstvertebrates, and this final discussion has outlined the arguments supporting the existence of aCRG and a CPG in bullfrogs much like those described for mammals. The differences inbreathing mechanics are the most obvious distinction between amphibian and mammalianbreathing. It is clear that the emergence of the diaphragm and the aspiration pump wereimportant steps in the evolution of air breathing, but the data thus far do not indicate that thisprocess required major reorganization at the level of the control system, given that the breathingcycle of the bullfrog also appears to consist of three distinct phases. At first glance, the fact thatmany lower vertebrates breathe episodically may be interpreted as a reflection of importantdifferences in their respiratory control system, given the almost dogmatic notion that mammalsbreathe continuously, and that any other type of breathing pattern is either pathological or anPage 243experimental artefact. However, there is growing evidence suggesting that many groups ofmammals can produce episodic breathing if the respiratory drive is reduced sufficiently. Thismay occur naturally during hibernation or euthermic sleep or, as it has been documented in somecases, when the chemical drive to breathe is reduced artificially (Bartoli et at., 1974). Thisfurther supports the notion that the respiratory control system is highly conserved amongstvertebrates, and a better understanding of the mechanisms that dictate the production ofbreathing episodes in all vertebrate classes will help elucidate this question.Page 244BIBLIOGRAPHYPage 245Adrian, E.D. and F.J.J. Buytendijk. (1931). Potential changes in the isolated brain stem of thegoldfish. J. Physiol. London. 71:121-135.Ballam, G.O. and E.L. Coates. (1989).Effect of upper airway CO2 pattern on ventilatoryfrequency in tegu lizards. Am. J. Physiol. 26(1): R156-R161.Ballintijn, C. M. (1987). Evolution of central nervous control of ventilation in vertebrates. In:The neurobiology of the cardiorespiratory system. Edited by E.W. Taylor. ManchesterUniversity Press. pp.1-27.Bartoli, A., E. Bystrzycka, A. Guz, S.K. Jam, M.I.M. Noble and D. Trenchard. (1973).Studies of the pulmonary vagal control of central respiratory rhythm in the absence of breathingmovements. J. Physiol. London. 230: 449-465.Bartoli, A., B.A. Cross, A. Guz, S.K. Jam, M.I.M. Noble and D.W. Trenchard. (1973). Avagal reflex modifying ventilation activated by CO2 in the lungs. J. Physiol. London. 234:93P-94P.Bartoli, A., B.A. Cross, A. Guz, S.K. Jam, M.I.M. Noble and D.W. Trenchard. (1974).Theeffect of carbon dioxide in the airways and alveoli on ventilation; A vagal reflex studied in thedog. J. Physiol. London. 240: 91-109.Berger, A.J. (1990). Recent advances in respiratory neurobiology using in vitro methods. Am.J. Physiol. 259: L24-L29.Bianchi, A.L., M. Denavit-Saubié and 3. Champagnat. (1995). Central control of breathing inmammals: Neuronal circuitry, membrane properties, and neurotransmitters. PhysiologicalReviews. 75(1): 1-45.Boutilier, R.G. (1988). Control of arrhythmic breathing in bimodal breathers: Amphibia. Can.J. Zool. 66: 6-19.Boutilier, R.G. (1989). Diving Physiology: Amphibians. In: Comparative PulmonaryPhysiology: Current concepts. Edited by S.C. Wood. Marcel Dekker Inc. New York. PP 677-692.Boutilier, R.G., D.J. Randall, G. Shelton and D.P. Toews. (1979). Acid-base relationships inthe blood of the toad, Bufo marinus I. The effects of environmental CO2. J. Exp. Biol. 82:331-344.Boutilier, R.G. and D.P. Toews. (1977). The effect of progressive hypoxia on respiration in thetoad Bufo marinus. J. Exp. Biol. 68: 99-107.Boutilier, R. G. and D . P. Toews. (1981). Respiratory, circulatory and acid-base adjustments tohypercapriia in a strictly aquatic and predominantly skin-breathing urodele, Cryptobranchusalleganiensis. Respir. Physiol. 46(2): 177-192.Page 246Branco, L.G.S., M.L. Glass and A. Hoffmann. (1992). Central chemoreceptor drive tobreathing in unanesthetized toads, Bufo paracnemis. Respir. Physiol. 87: 195-204.Branco, L.G.S., M.L. Glass, T. Wang and A. Hoffmann. (1993). Temperature and centralchemoreceptor drive to ventilation in toad (Bufo paracnemis). Respir. Physiol. 93(3): 337-346.Brett, S.S. and G. Shelton. (1979). Ventilatory mechanisms of the amphibian, Xenopus laevis;the role of the buccal force pump. J. Exp. Biol. 80: 25 1-269.Brockhaus, J., K. Ballanyi, J.C. Smith and D.W. Richter. (1993). Microenvironment ofrespiratory neurons in the in vitro brainstem-spinal cord of neonatal rats. J. Physiol. London.462: 421-445.Burggren, W. W. (1984). Transition of respiratory processes during amphibian metamorphosis:from egg to adult. In: Respiration and Metabolism of Embryonic Vertebrates; edited by R.S.Seymour. Dordrecht: Dr. W . Junk Publishers, pp.31-53.Burggren, W. W. (1988). Cardiovascular responses to diving and their relation to lung andblood oxygen stores in vertebrates. Can. J. Zool. 66: 20-28.Burggren, W.W. (1989). Lung structure and function. In: Comparative Pulmonary Physiology:Current Concepts, edited by S.C. Wood and C. Lenfant. New-Yrok: Dekicer. 39:153-192.Burggren, W. W. and M. Doyle. (1986). Ontogeny of regulation of gill and lung ventilation inthe bullfrog, Rana catesbeiana. Respir. Physiol. 66(3): 279-291.Burggren, W.W. and A.W. Pinder. (1991). Ontogeny of Cardiovascular and RespiratoryPhysiology in Lower Vertebrates. Ann. Rev. Physiol. 53: 107-135.Burggren, W.W. and N.H. West. (1982). Changing respiratory importance of gills, lungs andskin during metamorphosis in the bullfrog Rana catesbeiana. Respir. Physiol. 47(2): 15 1-164.Carison, A.J. and A.B. Luckhardt. (1920). Studies on the visceral sensory nervous system I.Lung automatism and lung reflexes in the frog (R. pipiens and R. catesbiana). Am. J. Physiol.54: 55-95.Castellini, M.A., W.K. Milsom, R.J. Berger, D.P. Costa, D.R. Jones, M. Castellini, L.D. Rea,S. Bharma and M.B. Harris. (1994). Patterns of respiration and heart rate during wakefulnessand sleep in elephant seal pups. Am. J. Physiol. 266: R863-R869.Cherniack, N.S., N.N. Stanley, P.G. Tuteur, M.D. Altose and A.P. Fishman. (1973). Effectsof lung volume changes on respiratory drive during hypoxia and hypercapnia. J. Appi. Physiol.35(5): 635-641.Clark, F.J. and C. Von Euler. (1972). On the regulation of depth and rate of breathing. J.Physiol. London 222: 267-295.Page 247Coates, E.L. and G.O. Ballam. (1989). Breathing and upper airway C02 in reptiles: role of thenasal and vomeronasal systems. Am. J. Physiol. 25: R91-R97.Coates, E.L. and G.O. Ballam. (1990). Olfactory receptor response to C02 in bullfrogs. Am.J. Physiol. 27(5): R1207-R1212.Coates, E.L., R.A. Furilla, GO. Ballam and D. Bartlett Jr. (1990). A decrease in nasal CO2stimulates breathing in tegu lizards. Respir. Physiol. 86: 65-75.Coyle, J.T., M.E. Molliver and M.J. Kuhar. (1978). In situ injection of kaimc acid: a newmethod for selectively lesioning neuronal cell bodies while sparing axons of passage. J. Comp.Neurol. 180: 301-324.Cross, B.A., P.W. Jones and A. Guz. (1980). The role of vagal afferent information duringinspiration in determining phrenic motoneurone output. Respir. Physiol. 39(2): 149-167.De Jongh, H.J. and C. Gans. (1969). On the mechanism of respiration in the bullfrog, Ranacatesbeiana: a reassessment. J. Morph. 127: 259-290.Dejours, P. (1988). Respiration in water and air: Adaptations Regulation Evolution. Elsevier,New York. 179p.De Marneffe-Foulon, C. (1962). Contribution a l’étude du mécanisme et du contrôle desmouvements respiratoires chez Rana. Ann. Soc. Roy. Zoo!. Belgique 92(1): 81-132.Denavit-Saubié, M., D. Riche, J. Champagnat and J.C. Velluti. (1980). Functional andmorphological consequences of kainic acid microinjections into a pontine respiratory area of thecat. Neuroscience. 5: 1609-1620.Douse, M.A. and G.S. Mitchell. (1990). Episodic respiratory related discharge in turtle cranialmotoneurons: in vivo and in vitro studies. Brain Res. 536: 297-300.Douse, M.A. and G.S. Mitchell. (1992a). Effects of vagotomy on ventilatory responses to CO2in alligators. Respir. Physiol. 87: 63-76.Douse, M.A. and G.S. Mitchell. (1992b). Episodic breathing in alligators: role of sensoryfeedback. Respir. Physiol. 87: 77-90.Eldridge, F.L. (1971). Relationship between phrenic nerve activity and ventilation. J. Appi.Physiol. 221: 535-543.Eldridge, F .L. (1975). Relationship between respiratory nerve and muscle activity and muscleforce output. J. App!. Physiol. 39: 567-573.Page 248Eldridge, F.L. and D.E. Milihorn (1986). Oscillation, gating and memory in the respiratorycontrol system. In: Handbook of Physiology. Respiration, edited by A.P. Fishman, N.S.Cherniack and J.G. Widdicombe. Bethesda, MD: Am. Physiol. Soc., 1986, sect. 3, vol. I,chapt. 3, p.93-114.Evans, B.K. and G. Shelton. (1984). Ventilation inXenopus laevis after lung or carotid labyrinthdenervation. Proc. IUBS: Comp. Physiol. Biochem. 1: A 75.Feder, M.E. (1983). Responses to acute aquatic hypoxia in larvae of the frog Rana berlandieri.J. Exp. Biol. 104: 79-95.Feldman, J.L. 1986. Neurophysiology of respiration in mammals. In: Handbook of physiology,Sect. 1. The nervous System. Vol. 4. Edited by Bloom, F.E. American Physiological Society,Bethesda, MD pp. 463-524.Feldman, J.L. and H. Gautier. (1976). Interaction of pulmonary afferents and pneumotaxiccenter in control of respiratory pattern in cats. J. Neurophysiol. 39: 31-44.Feldman, J.L., J.C. Smith, H.H. Ellenberger, C.A. Connelly, G. Liu, J.J. Greer, A.D. Lindsayand M.R. Otto. (1990).Neurogenesis of respiratory rhythm and pattern: emerging concepts. Am.J. Physiol. 28: R879-R886.Galante, R.J., A.I. Pack and A.P. Fishman. (1989). Control of lung breath duration in theAfrican lungfish. Faseb J. 3(3): A856.Galante, R.J., E. Smith, L. Kubin and A.I. Pack. (1992). Developmental changes in respiratoryneural ouput in larval form of Rana catesbeiana. Soc. Neurosci. Abstract. 58.13.Gans, C. (1970). Strategy and sequence in the evolution of the external gas exchangers ofectothermal vertebrates. Forma Functio 3: 6 1-104.Glass, M.L. and K. Johansen. (1976). Control of breathing inAcrochordusjavanicus, an aquaticsnake. Physiol. Zool. 49: 328-340.Gottlieb, G. and D.C. Jackson. (1976). Importance of pulmonary ventilation in respiratorycontrol in the bullfrog. Am. J. Physiol. 230: 608-613.Grippi, M.A., A.I. Pack, R.O. Davies and A.P. Fishman. (1985). Adaptation to reflex effectsof prolonged lung inflation. J. Appi. Physiol. 58: 1360-1371.Halliday, T.R. and H.P.A. Sweatman (1976). To breathe or not to breathe: the newt’s problem.Anim. Behav. 24:551-561.Infantino, R.L. (1989). Ontogeny of gill and lung ventilatory response to oxygen and carbondioxide in the bull frog Rana catesbeiana. Am. Zool. 29:57A (Abstract).Page 249Ishii, K. and K. Ishii (1970). Efferent innervation to the chemoreceptor of the carotid labyrinthof the toad. Tohoku J. Exp. Med. 102: 113-119.Ishii, K., K. Ishii, and T. Kusakabe (1985). Chemo- and baroreceptor innervation of the aortictrunk of the toad bufo vulgaris. Respir. Physiol. 60: 365-375.Issa, F. Q. and J. E. Remmers. (1992). Identification of a subsurface area in the ventral medullasensitive to local changes in CO2 J. App!. Physiol. 72: 439-446.Jackson, D.C. (1978). Respiratory control in air breathing ectotherms. In: Regulation ofVentilation and Gas Exchange; Academic Press, Inc., pp.93-130.Jackson, D.C. and B.A. Braun. (1979). Respiratory control in bullfrogs: cutaneous versuspulmonary response to selective CO2 exposure. J. Comp. Physiol. B 129: 339-342.Jesse, M.J., C. Shub and A.P. Fishman. (1967). Lung and gill ventilation of the African lungfish. Respir. Physiol. 3: 267-287.Johansen, K. and C. Lenfant. (1968). Respiration in the African lungfish, Protopterusaethiopicus. II. Control of breathing. J. Exp. Biol. 49: 453-468.Jones, D.R. and C. Chu. (1988). Effect of denervation of carotid labyrinths on breathing inunrestrained Xenopus laevis. Respir. Physiol. 73: 243-256.Jones, D.R. and W.K. Milsom. (1982). Peripheral receptors affecting breathing andcardiovascular function in non-mammalian vertebrates. J. Exp. Biol. 100: 59-91.Jones, D.R. (1970). Experiments on amphibian respiratory and circulatory systems. In:Experiments in Physiology and Biochemistry; Vol III, edited by G.A. Kerkut. New York:Academic Press, pp. 233-293.JUch, P.J.W. and C.M. Ballintijn. (1983). Tegmental neurons controlling medullary respiratorycentre activity in the carp. Respir. Physiol. 51(1): 95-107.Kelsen, S.G., M.D. Altose and N.S. Cherniack. (1977). Interaction of lung volume andchemical drive on respiratory muscle EMG and respiratory timing. .1. Appi. Physiol. 42(2):287-294.Kinney, J.L. and F.N. White. (1977). Oxidative cost of ventilation in a turtle, Pseudemysfloridana. Respir. Physiol. 31: 327-332.Kogo, N., S.F. Perry and J.E. Remmers. (1994). Neural organization of the ventilatory activityin the frog, Rana catesbeiana I. J. Neurobiol. 25(9): 1067-1079.Kogo, N. and J.E. Remmers. (1994). Neural organization of the ventilatory activity in the frog,Rana catesbeiana. II. J. Neurobiol. 25(9): 1080-1094.Page 250Kruhøffer, M., M.L. Glass, A.S. Abe and K. Johansen. (1987). Control of breathing in anamphibian Bufo paracnemis: effects of temperature and hypoxia. Respir. Physiol. 69: 267-275.Kuhlmann, W.D. and M.R. Fedde. (1979). Intrapulmonary receptors in the bullfrog: sensitivityto CO2. J. Comp. Physiol. 132A: 69-75.Larsell, 0. (1924). The nucleus isthmi of the frog. J. Comp. Neurol.36:309-322.Liao, G.S., R.J. Galante, A.P. Fishman, L. Kubin and A.I. Pack. (1994). Blockade of smallconductanceCa2-activ ted K channels (sKj affects gill and lung ventilation in tadpole, Ranacatesbeiana), brainstem in vitro. The Physiologist, 37(5):24.9.Lillo, R.S. (1980). Heart rate and blood pressure in bullfrogs during prolonged maintenancein water at low temperature. Comp. Biochem. Physiol. 65A:251-253.Lumsden, T. (1923). The regulation of respiration. part I. J. Physiol. London. 58: 82-91.Maclntyre, D.H. (1975). Respiratory mechanisms and acid base regulation in Bufo marinusmarinus (L.). M.Sc. Dissertation. Acadia University, Wolfville, N.S.Maclntyre, D.H. and D.P. Toews. (1976). The mechanics of lung ventilation and the effects ofhypercapnia on respiration in Bufo marinus. Can. J. Zool. 54: 1364-1374.McArthur, M.D. and W.K. Milsom. (1991a). Ventilation and respiratory sensitivity ofeuthermic Columbian and golden-mantled ground squirrels (Spennophilus columbianus andSpermophilus lateralis) during the summer and winter. Physiol. Zool. 64(4): 92 1-939.McArthur, M.D. and W.K. Milsom. (1991b). Changes in ventilation and respiratory sensitivityassociated with hibernation in Columbian (Spennophilus columbianus) and golden-mantled(Spermophilus lateralis) ground squirrels. Physiol. Zool. 64(4): 940-959.McLean, H.A., G.S. Mitchell and W.K. Milsom. (1989). Effects of prolonged inflation onpulmonary stretch receptor discharge in turtles. Respir. Physiol. 75: 75-88.McLean, H.A. (1992). Endogenous respiratory rhythm in isolated brainstem preparations. Ph.D.Thesis, Dept. of Medical Science, Univ. of Calgary.Martin, H.N. (1878). The normal respiratory movements of the frog, and the influence upon itsrespiratory centre of stimulation of the optic lobes. J. Physiol. London 1: 131-170.Milsom, W.K. (1989). Comaprative aspects of vertebrate pulmonary mechanics. In:Comparative Pulmonary Physiology: Current Concepts, edited by S.C. Wood and C. Lenfant.New-Yrok: Dekker. 39:587-619.Page 251Milsom, W.K. (1990a). Control and co-ordination of gas exchange in air breathers. In:Advances in Comparative and Environmental Physiology, Vol.6; edited by R.G. Boutilier.Berlin: Springer-Verlag, pp.347-400.Milsom, W.K. (1990b). Mechanoreceptor modulation of endogenous respiratory rhythms invertebrates. Am. J. Physiol. 28: R898-R910.Milsom, W.K. (1991). Intermittent breathing in vertebrates. Annu. Rev. Physiol. 53:87-105.Milsom, W.K. and P. Chan. (1986). The relationship between lung volume, respiratory driveand breathing pattern in the turtle, Chiysemys picta. J. Exp. Biol. 120: 233-247.Milsom, W.K. and D.R. Jones. (1977). Carbon dioxide sensitivity of pulmonary receptors inthe frog. Experientia 33: 1167-1168.Mitchell, G. S. (1987). Effects of hypoxemia on phrenic nerve responses to static lung inflationin anesthetized dogs. Respir. Physiol. 67: 183-195.Mitchell, G. S. (1992). In vitro studies of respiratory control: An overview. In: RespiratoryControl: Central and Peripheral Mechanisms, edited by D.F. Speck, M.S. Dekin, W.R.Revelette and D . T. Frazier. The University Press of Kentucky, pp. 30-33.Mitchell, G.S., B.A. Cross, T. Hiramoto and P. Scheid. (1980). Effects of intrapulmonary CO2and airway pressure on phrenic activity and pulmonary stretch receptor discharge in dogs.Respir. Physiol . 41(1): 29-48.Mitchell, G.S., B.A. Cross, T. Hiramoto and P. Scheid. (1982). Interactions between lungstretch and Pa02 in modulating ventilatory activity in dogs. J. Appi. Physiol. 53(1): 185-191.Mitchell, G.S., M.A. Douse and K.T. Foley. (1990). Receptor interactions in modulatingventilatory activity. Am. J. Physiol. 28: R91 1-R920.Mitchell, G. S. and B. D. Selby. (1987). Effects of carotid denervation on interactions betweenlung inflation and PaCO2 in modulating phrenic activity. Respir. Physiol. 67: 367-378.Mitchell, G.S. and B.D. Selby. (1988). Ventilatory responses to lung inflation and arterial CO2in halothane-anesthetized dogs. J. Appi. Physiol. 64: 1433-1438.Mitchell, G.S. and E.H. Vidruk. (1987). Effects of hypercapnia on phrenic and stretch receptorresponses to lung inflation. Respir. Physiol. 68: 310-330.Murakoshi, T. and M. Otsuka. (1985). Respiratory reflexes in an isolated brainstem-lungpreparation of the newborn rat:possible involvement of -y-aminobutyric acid and glycine.Neurosci. Lett. 62: 63-68.Page 252Naifeh, K.H., S.E. Huggins and H.E. Hoff. (1971a). Study of the control of crocodilianrespiration by anesthetic dissection. Respir. Physiol. 12: 251-260.Naifeh, K.H., S.E. Huggins and H.E. Hoff. (1971b). Effects of brain stem section onrespiratory patterns of crocodilian reptiles. Respir. Physiol. 13: 186-197.Nattie, E.E. and A. Li. (1990). Fluorescence location of RVLM kainate microinjections thatalter the controlof breathing. J. Appi. Physiol. 68(3): 1157-1166.Nattie, E.E., A. Li and W.M. St.John. (1991). Lesions in retrotrapezoid nucleus decreaseventilatory output in anesthetized or decerebrate cats. Am. J. Physiol. 71: 1364-1375.Nicholson, C. (1985). Diffusion from an injected volume of a substance in brain tissue witharbitrary volume fraction and tortuosity. Brain Res. 333: 325-329.Nielsen, B. (1961). On the regulation of the respiration in reptiles. I. The effect of temperatureand CO2 on the respiration of lizards (Lacerta). J. Exp. Biol. 38: 301-314.Nieuwenhuys, R. and P. Opdam. (1976). Structure of the brain stem. In: Frog Neurobiology:A Handbook, edited by R. Llinas and W. Precht. Springer Verlag, Berlin, Heidelberg, New-York. pp. 911-855.Norris, D.O. (1985). Vertebrate endocrinology. Second Edition. Lea and Febiger, Philadelphia.5OSp.Oka, K. (1958a). The influence of the transection of the brain upon the respiratory movementof the frog. J .Physiol .Soc .Japan 20(7): 513-519.Oka, K. (195 8b). Further studies on the localization of the respiratory centers of the frog. J.Physiol. Soc. Japan 20(7): 520-524.Okada, Y., K. Mückenhoff, G. Holtermann, H. Acker and P. Scheid. (1993a). Depth profilesof pH and P02 in the isolated brain stem-spinal cord of the neonatal rat. Respir . Physiol . 93(3):315-326.Okada, Y., K. Mückenhoff and P. Scheid. (1993b). Hypercapnia and medullary neurons in theisolated brain stem-spinal cord of the rat. Respir . Physiol . 93(3): 327-336.Opdam, P., M. Kemali and R. Nieuwenhuys. (1976). Topological analysis of the brain stem ofthe frogs Rana esculenta and Rana catesbeiana. J. Comp. Neur. 165: 307-332.Pack, A.I. , R.J. Galante, R.E. Walker, L.K. Kubin and A.P. Fishman. (1992). Comparativeapproach to neural control of respiration. In: Respiratory Control: Central and PeripheralMechanisms, edited by D.F. Speck, M.S. Dekin, W.R. Revelette and D.T. Frazier. TheUniversity Press of Kentucky, pp. 52-57.Page 253Pack, A.I., R.J. Galante and A.P. Fishman. (1990). Control of interbreath interval in theAfrican lungfish. Am .J .Physiol. 28(1): R139-R146.Pack, A.I., R.J. Galante and A.P. Fishman. (1992). Role of lung inflation in control of airbreath duration in African lungfish (Protopterus annectens). Am .J .Physiol . 262: R879-R884.Pinder, A.W. and W.W. Burggren. (1986). Ventilation and partitioning of oxygen uptake in thefrog Rana pliens: effects of hypoxia and activity. J .Exp .Biol. 126: 453-468.Portner, H.0., L.M. MacLatchy and D.P. Toews. (1991). Metabolic responses of the toad Bufomarinus to environmental hypoxia: an analysis of the critical P02 Physiol. Zool. 64(3): 836-849.Randall, D.J. Burggren, W.W. Farrell, A.P. and M.S. Haswell. (1981). The evolution of airbreathing in vertebrates. Cambridge University Press. 133p.Richter, D.W. (1982). Generation and maintenance of the respiratory rhythm. J .&p. Biol. 100:93-107.Richter, D.W., D. Ballaantyne and J.E. Remmers. (1986). How is the respiratory rhythmgenerated? A model. NIPS 1: 109-112.Rovainen, C. M. (1985). Respiratory bursts at the midline of the rostral medulla of the lamprey.J. Comp. Physiol. 157A: 303-309.St-John, W. M. (1977). Integration of peripheral and central chemoreceptor stimuli by pontineand medullary respiratory centers. Federation Proc. 36:2421-2427.Sakakibara, Y. (1978). Localization of CO2 sensor related to the inhibition of the bullfrogrespiration. Jpn .J .Physiol . 28: 721-735.Sakakibara, Y. (1984a). The pattern of respiratory nerve activity in the bullfrog. Jpn. J. Physiol.34: 269-282.Sakakibara,Y. (1 984b). Trigeminal nerve activity and buccal pressure as an index of totalinspiratory activity in the bullfrog. Jpn .J .Physiol . 34: 827-838.Schlaefke, M. E., J. F. Kille and H. H. Loeschcke. (1979). Elimination of central chemosensitivityby coagulation of a bilateral area on the ventral medullary surface in awake cats. Pflugers Arch.378: 231-241.Shelton, G. (1959). The respiratory centre in the tench (Tinca tinca L.). J. Exp. Biol. 36:19 1-202.Shelton, G. (1961). The respiratory center in the tench (Tinca Tinca L.) II. Respiratory neuronalactivity in the medulla oblongata. J. Exp. Biol. 38:79-92.Page 254Shelton, 0. (1985). Functional and evolutionary significance of cardiovascular shunts in theamphibia. In: Cardiovascular Shunts; edited by K. Johansen and W.W. Burggren. Munlcsgaard:Alfred Benzon Symposium 21, pp.100-120.Shelton, 0. and R.G. Boutilier. (1982). Apnoea in amphibians and reptiles. J. Exp. Biol. 100:245-273.Shelton, G. and P.C. Croghan. (1988). Gas exchange and its control in non-steady-state systems:the consequences of evolution from water to air breathing in the vertebrates. Can. J. Zool. 66:109-123.Shelton, G., D.R. Jones and W.K. Milsom. (1986). Control of Breathing in EctothermicVertebrates. In: Handbook of Physiology: Section 3, The Respiratory System; edited by A.P.Fishman. Bethesda,MD: American Physiological Society, pp.857-909.Sherrington, C.S.. (1891). Note on Cheyne-Stokes breathing in the frog. J. Physiol. London 12:292-298.Smatresk, N.J. (1990). Chemoreceptor modulation of endogenous respiratory rhythms invertebrates. Am. J. Physiol . 259(28): R887-R897.Smatresk, N.J. and A. W. Smits. (1991). Effects of central and peripheral chemoreceptorstimulation on ventilation in the marine toad, Bufo marinus. Respir. Physiol. 83(2): 223-238.Smith, E.G., R.J. Galante, S. Lahiri and A.I. Pack. (1991a). Respiratory responses to CO2 inthe brainstem preparation of Rana catesbeiana. The Physiologist. 34(4): 266.Smith, J.C., H.H. Ellenberger, K. Ballanyi, D.W. Richter and J.L. Feldman.(199 lb) . Pre-Botzinger complex: a brainstem region that may generate respiratory rhythm inmammals. Science 254: 726-729.Smith, J.C. and J.L. Feldman. (1987). In vitro brainstem-spinal cord preparations for study ofmotor systems for mammalian respiration and locomotion. J. Neurosci. Meth. 21: 321-333.Stehouwer, D .J. (1987). Effect of tectomy and decerebration on spontaneous and elicitedbehaviour of tadpoles and juvenile frogs. Behav. Neurosci. 101(3): 378-384.Senn, D.G. (1972). Development of tegmental and rhombencephalic structures in a frog (Ranatemporaria L.). Acta anat. 82: 525-548.Stuesse, S.L., W.L.R. Cruce and KS. Powell. (1984). Organization within the cranial IX-Xcomplex in ranid frogs: a horseradish peroxidase transport study. J. Comp. Neurol. 222: 358-365.Suzue, T. (1984). Respiratory rhythm generation in the in vitro brain stem-spinal cordpreparation of the neonatal rat. J. Physiol. London. 354:135-152.Page 255Taglietti, V. and C. Casella. (1966). Stretch receptors stimulation in frog’s lungs. Pflugers Arch.292: 297-308.Taglietti, V. and C. Casella. (1968). Deflation receptors in frog’s lungs. Pflugers Arch . 304:8 1-89.Taylor, E.W. (1989). Nervous control of ventilation and heart rate in elasmobranch fish, amodel for the study of the central neural mechanisms mediating cardiorespiratory interactionsin mammals. In: Non-Mammalian Animal Models in BIomedical Research. Edited by by A.D.Woodhead, CRC Press, Florida, pp. 161-183.Taylor, A.C. and J.J. Kollros. (1946). Stages in the normal development of Ranapipiens larvae.Anat. Rec. 94: 7-24.Toews, D.P. (1971). Factors affecting the onset and termination of respiration in thesalamander, Amphiuma tridaclylum. Can. J. Zool. 49(9): 1231-1237.Toews, D. and D. Maclntyre. (1978). Respiration and circulation in an apodan amphibian. Can.J. Zool. 56(4):998-1004.Tortora, G.J. (1992). Principles of human anatomy, sixth edition. Harper Collin, Publishers,Van Vliet, B. N. and N.H. West. (1986). Cardiovascular responses to electrical stimulation ofthe recurrent laryngeal nerve in conscious toads (Bufo marinus). J. Comp. Physiol.156B :363-375.Van Vliet, B.N. and West, N.H. (1987). Static and dynamic response characteristics ofpulmocutaneous arterial baroreceptors in the toad, Bufo marinus. Manuscript. 32 pp and 14figuresVan Vliet, B. N. and N.H. West. (1992). Functional characteristics of arterial chemoreceptorsin an amphibian (Bufo marinus). Respir. Physiol. 88(1,2): 113-127.Vitalis, T. Z. and W . K. Milsom (1986). Mechanical analysis of spontaneous breathing in thesemi-aquatic turtle, Pseudemys scripta. J. Exp. Biol. 125: 157-171.Vitalis, T.Z. and G. Shelton. (1990). Breathing in Rana pipiens: the mechanism of ventilation.J. Exp. Biol. 154: 537-556.Walker, R., R.J. Galante, A.P. Fishman and A.I. Pack (1990). Effect of Gaba on gill and lungventilation in an in vitro isolated brainstem preparation in the tadpole. The Physiologist. 33(4):A-35.Page 256Wang, T., L.G.S. Branco and M.L. Glass. (1994). Ventilatory responses to hypoxia in the toadBufo paracnemis before and after a decrease in haemoglobin oxygen-carrying capacity. J. Exp.Biol. 186: 1-8.West, N.H. and W.W. Burggren. (1982). Gill and lung ventilatory responses to steady-stateaquatic hypoxia and hyperoxia in the bullfrog tadpole. Respir. Physiol. 47(2): 165-176.West, N.H. and W . W. Burggren. (1983). Reflex interactions between aerial and aquatic gasexchange organs in larval bullfrogs. Am. J. Physiol. 244(13): R770-R777.West, N.H. and W.W. Burggren. (1984). Factors influencing pulmonary and cutaneous arterialblood flow in the toad, Bufo marinus. Am. J. Physiol. 247(16): R884-R894.West, N.H., P.J. Butler and R.M. Bevan. (1992). Pulmonary blood flow at rest and duringswimming in the green turtle, Chelonia mydas. Physiol. Zool. 65(2): 287-3 10.West, N.H. and D.R. Jones. (1975). Breathing movements in the frog Rana pipiens I. Themechanical events associated with lung and buccal ventilation. Can. J. Zool. 53: 332-344.West, N.H., A.W. Smits and W.W. Burggren. (1989). Factors terminating nonventilatoryperiods in the turtle, Chelydra serpentina. Respir. Physiol. 77(3): 337-350.West, N.H., Z.L. Topor and B.N. van Vliet. (1987). Hypoxemic threshold for lung ventilationin the toad. Respir. Physiol. 70: 377-390.West, N.H. and B.N. van Vliet. (1992). Sensory mechanisms regulating the cardiovascular andrespiratory systems. In: Environmental Physiology of the Amphibians; edited by M. Feder andW. Burggren. Chicago: University of Chicago, pp.151-182.Woidring, S. (1965). Interrelation between lung volume, arterial CO2 tension, and respiratoryactivity. J. Appi. Physiol. 20: 647-652.Page 257


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items