SUMMARY1. The effect of systemic hypoxia was tested in anaesthetized, immobilized, thoracotomized and artificially ventilated cats with peripheral chemoreceptor afferents either intact or cut. Extracellular recordings from different types of medullary respiratory neurones and intracellular recordings from stage 2 expiratory neurones were made to determine the hypoxia-induced changes in neuronal discharge patterns and postsynaptic activity as an index for the disturbances of synaptic interaction within the network.2. The general effect of systemic hypoxia was an initial augmentation of respiratory activity followed by a secondary depression. In chemoreceptordenervated animals, secondary depression led to central apnoea.3. The effects of systemic hypoxia were comparable with those of cerebral ischaemia following occlusion of carotid and vertebral arteries.4. In chemoreceptor-denervated animals, all types of medullary respiratory neurones ceased spontaneous action potential discharge during hypoxia.5. Reversal of inhibitory postsynaptic potentials (IPSPs) and/or blockade of IPSPs was seen after 2-3 min of hypoxia.6. During hypoxia, the membrane potential of stage 2 expiratory neurones showed a slight depolarization to -45 to -55 mV and then remained stable.7. The neurone input resistance increased initially and then decreased significantly during central apnoea.8. Rhythmogenesis of respiration was greatly disturbed. This was due to blockade of IPSPs and, in some animals, to more complex disturbances of phase switching from inspiration to expiration.9. Central apnoea occurred while respiratory neurones were still excitable as shown by stimulus-evoked orthodromic and antidromic action potentials.10. The results indicate that the medullary respiratory network is directly affected by energy depletion. There is indication for a neurohumoral mechanism which blocks synaptic interaction between respiratory neurones in chemoreceptorintact animals. MS 9022 D. W. RICHTER AND OTHERS

A mathematical model of the three-phase respiratory network proposed by Richter et al. (News Physiol. Sci. 1: 109-112, 1986) is developed and its properties are examined. The model reproduces the experimentally determined trajectories of membrane potential for the five physiologically distinct types of neurons included. Stepwise parameter changes can produce a respiratory rhythm with only two separate electrophysiological phases, result in apnea, or produce more complex patterns of firing. The phase-resetting behavior of the model was obtained with perturbing stimuli and is comparable to experimentally determined phase-resetting data. There is reasonable agreement between model predictions and experimental results. In the model, the properties of the phase singularity make termination of the respiratory rhythm by an appropriately timed perturbation virtually impossible, which is in agreement with experimental observations. The rhythm can be stopped by alterations that simulate the effect of input from the superior laryngeal nerve; the rhythm is locked in the postinspiratory phase. We conclude that our results are consistent with the concept of a network oscillator as the source of the respiratory rhythm.

SUMMARY1. In urethane or Nembutal anaesthetized and artificially ventilated Wistar rats, respiratory neurones of the ventrolateral medulla oblongata were analysed in extracellular (n = 74) and intracellular (n = 43) recordings.2. Some respiratory neurones were identified as bulbospinal by their antidromic response to spinal cord stimulation at the C4 level. The neurones examined were not antidromically excited by vagal nerve stimulation.3. Based on their discharge pattern in relation to efferent phrenic and vagal nerve activity, six types of respiratory neurones were classified: early-inspiratory, throughout-inspiratory, late-inspiratory, post-inspiratory, expiratory, and phasespanning expiratory-inspiratory neurones. 4. Analysis of postsynaptic activities and IPSP reversal following chloride injection revealed post-inspiratory and expiratory inhibition in inspiratory neurones, a pronounced early-inspiratory and a relatively weak expiratory inhibition in postinspiratory neurones, and an early-inspiratory and post-inspiratory inhibition in expiratory neurones.5. In phase-spanning expiratory-inspiratory neurones the post-inspiratory inhibition was strong and effectively blocked action potential discharge. Expiratory-inspiratory neurones were quite similar to the group of inspiratory neurones, but seemed to receive tonic excitatory inputs not shunted by weak expiratory inhibition. This pre-inspiratory discharge was readily blocked by weak negative DC injection.6. Under conditions of experimental hypoxia, or long lasting lung inflation and non-inflation, post-inspiratory neurones displayed a second burst of discharge at the end of the expiratory phase in addition to their longer lasting post-inspiratory discharge.7. We conclude that in the rat the central respiratory rhythm is organized in three (inspiratory, post-inspiratory, expiratory) phases, and that synaptic interaction within the medullary respiratory network of the rat occurs similarly to that described for the cat.

SUMMARY1. Termination of inspiration is an essential component of respiratory rhythm generation and its perturbation can result in apneusis, i.e. significant prolongation of inspiratory activity. In an effort to further analyse inspiratory termination mechanisms, we studied the p'ostsynaptic events in respiratory neurones during apneustic respiratory periods, and compared them to normal respiratory cycles.2. Experiments were performed in pentobarbitone-anaesthetized, paralysed, thoracotomized cats ventilated with a constant volume or a cycle-triggered constant pressure pump. Apneusis, separated by normal cycles, was induced as follows: the animal was ventilated by a cycle-triggered pump that normally inflated the lungs during the inspiratory burst of phrenic nerve discharge. The NMDA-receptor blocker MK-801 [(+ )-5-methyl-10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5,10-iminemaleate] (0-3-0-7 mg/kg) was administered intravenously, and, for designated breaths, inflation of the lungs was withheld during neural inspiration.3. Membrane potential trajectories of forty-one late expiratory (E-2) and eight postinspiratory (PI) neurones of the caudal ventral respiratory group were analysed before and/or after MK-801 administration, during normal and apneustic periods.4. Before MK-801 administration, withholding lung inflation caused modest (10-20%) lengthening of the inspiratory period; after MK-801 administration, withholding inflation caused apneusis. Provided that the lungs were inflated during the inspiratory phase, the temporal pattern of phrenic nerve, recurrent laryngeal nerve and membrane potential trajectories of E-2 and PI neurones were not significantly altered by MK-801. Apneusis following NMDA-receptor blockade produced consistent changes in the synaptic activation patterns of E-2 neurones. In particular, the slow late inspiratory-related depolarization pattern of E-2 neurones was consistently retarded during apneustic inspiratory phases when compared to normal inspiratory phases. This was due to continuation of Cl--mediated synaptic inhibition of E-2 neurones. Superior laryngeal nerve stimulation stopped apneusis and sustained membrane hyperpolarization of E-2 neurones similar to lung inflation. 6. We conclude that: (i) the prolonged inhibition of E-2 neurones during apneusis is indicative of the process responsible for the prolongation of the inspiratory phase.(ii) Synaptic interactions between medullary respiratory neurones and spinal respiratory motoneurones producing burst patterns of motor output do not require NMDA-receptor-mediated pathways. (iii) Pathways used by pulmonary and laryngeal afferents to terminate inspiration also do not require NMDA-receptormediated function. (iv) NMDA-receptor controlled pathways necessary for inspiratory termination seem to be activated in the absence of lung inflation and to involve the pontine respiratory group of neurones. (v) MK-801-induced apneusis affects the interaction between early-inspiratory and late-inspiratory neurones resulting in oscillations.

The state of activation of normal human intestinal mononuclear cells obtained from transplant donors was studied. Compared with PBMC, freshly isclated intestinal mononuclear cells expressed significantly more cell surface activation antigens on both B and T lymphocytes. Intestinal mononuclear cells contained significant numbers of immunoglobulin secreting cells immediately after cell separation. This population included CD5-positive B cells that secreted predominantly IgA. Cells from the large bowel consistently revealed higher numbers of IgA secreting cells than cells from the small bowel. Thus, intestinal B cells are markedly activated in vivo compared with PBMC and this increased activation correlates with increased spontaneous antibody secretion. B cells from the large intestine are more highly activated and secrete more antibody than do cells from the small intestine. The intestinal lamina propria lymphoid compartment exhibits a heightened state of activation that may be important for its distinct role in mucosal defense.