The central respiratory chemoreceptor: where is it located?-Invited article.

We review previous reports on the localization of the central chemoreceptor focusing on our studies that used various experimental techniques including lesioning (brainstem transection and removal of pia mater), analyses of neuronal responses to CO(2) by electrophysiological and optical recording, mapping of CO(2)-excitable neurons by c-fos immunohistochemistry and local acidic stimulation. Among these experimental techniques, voltage imaging with calculation of cross correlation coefficients between the respiratory output activity and each pixel, i.e., correlation coefficient imaging technique, enabled us to effectively analyze imaging data without empirical signal processing. The reviewed studies have indicated that the most superficial layer of the rostral ventral medulla, i.e., the surface portions of the nucleus retrotrapezoideus/parafacial respiratory group, nucleus parapyramidal superficialis and nucleus raphe pallidus, is important in central chemoreception. We suggest that one of the major respiratory rhythm generators, i.e., the preBötzinger complex, is not chemosensitive in itself or rather inhibited by CO(2). Based on our detailed analysis of c-fos immunohistochemistry, we propose a cell-vessel architecture model for the central respiratory chemoreceptor. Primary chemoreceptor cells are mainly located beneath large surface vessels within the marginal glial layer of the ventral medulla, and surround fine penetrating vessels that branch from a large surface vessel. Respiratory neurons in the rostral portion of the ventral respiratory group could be intrinsically chemosensitive, but their role in chemoreception might be secondary. Definitive identification of chemosensitive sites and chemoreceptor cells needs further studies.

Dendritic backpropagation and synaptic plasticity in the mormyrid electrosensory lobe.

This study is concerned with the origin of backpropagating action potentials in GABAergic, medium ganglionic layer neurones (MG-cells) of the mormyrid electrosensory lobe (ELL). The characteristically broad action potentials of these neurones are required for the expression of spike timing dependent plasticity (STDP) at afferent parallel fibre synapses. It has been suggested that this involves active conductances in MG-cell apical dendrites, which constitute a major component of the ELL molecular layer. Immunohistochemistry showed dense labelling of voltage gated sodium channels (VGSC) throughout the molecular layer, as well as in the ganglionic layer containing MG somata, and in the plexiform and upper granule cell layers of ELL. Potassium channel labelling was sparse, being most abundant in the deep fibre layer and the nucleus of the electrosensory lobe. Intracellular recordings from MG-cells in vitro, made in conjunction with voltage sensitive dye measurements, confirmed that dendritic backpropagation is active over at least the inner half of the molecular layer. Focal TTX applications demonstrated that in most case the origin of the backpropagating action potentials is in the proximal dendrites, whereas the small narrow spikes also seen in these neurones most likely originate in the axon. It had been speculated that the slow time course of membrane repolarisation following the broad action potentials was due to a poor expression of potassium channels in the dendritic compartments, or to their voltage- or calcium-sensitive inactivation. However application of TEA and 4AP confirmed that both A-type and delayed rectifying potassium channels normally contribute to membrane repolarisation following dendritic and axonal spikes. An alternative explanation for the shape of MG action potentials is that they represent the summation of active events occurring more or less synchronously in distal dendrites. Coincidence of backpropagating action potentials with parallel fibre input produces a strong local depolarisation that could be sufficient to cause local secretion of GABA, which might then cause plastic change through an action on presynaptic GABA(B) receptors. However, STP depression remained robust in the presence of GABAB receptor antagonists.

Disturbance of neural respiratory control in neonatal mice lacking GABA synthesizing enzyme 67-kDa isoform of glutamic acid decarboxylase.

To examine the role of GABA in the respiratory rhythm and pattern generation in neonatal mice, we analyzed the function of the respiratory control system of 67-kDa isoform of glutamic acid decarboxylase (GAD67)-deficient neonatal mice. In these mutant (GAD67-/-) mice, GABA levels in the brainstem were reduced to about 30% of those in wild-type (GAD67+/+) mice. In in vivo preparations, ventilatory parameters were analyzed by whole body plethysmography and electromyography of intercostal muscles. GAD67-/- mice exhibited abnormal respiratory patterns, i.e. irregular respiratory rhythm, and periodic gasp-like respiration followed by shallow breathing with short inspiratory duration and apnea. In in vitro GAD67-/- brainstem-spinal cord preparations, inspiratory C4 burst duration was shorter than that in GAD67+/+ preparations. Whole cell recordings revealed that activities of inspiratory neurons in the ventral medulla of GAD67-/- mice were characterized by a short depolarization period and a paucity of firing during the inspiratory phase. Superfusion of the in vitro GAD67-/- preparation with 10 microM GABA prolonged C4 burst duration and partly restored a normal pattern of inspiration, although the restoration was limited. These results indicate that reduced GABA levels during the perinatal period induce malfunction in the respiratory control system. We suggest that GABAergic transmission is not essential for basic respiratory rhythm generation but plays an important role in the maintenance of regular respiratory rhythm and normal inspiratory pattern in neonatal mice.

Cytoarchitecture of central chemoreceptors in the mammalian ventral medulla.

We reviewed the previous reports on the fine anatomy of the mammalian ventral medulla with special attention to the cytoarchitecture of the superficial chemosensitive regions to summarize what is known, what is not yet known, and what should be studied in the future. We also reviewed studies on anatomical relationship between neurons and vessels, and morphological studies on dendrites of respiratory or chemosensitive neurons. When we compared the morphological reports on the ventral and dorsal putative chemosensitive regions, similarities were found as follows. Chemosensitive cells were often found not only near the ventral surface but near the dorsal surface of the brainstem. Dendritic projection towards the surface was a common characteristic in the ventral and dorsal chemosensitive neurons. Morphological abnormality in the brainstem of sudden infant death syndrome victims was also summarized. On the basis of the previous reports we discussed the perspective on the future study on central chemoreception. Among various unanswered questions in central chemosensitivity studies, physiological significance of surface cells and surface extending dendrites is the most important topic, and must be thoroughly investigated.

Optical recording of the neuronal activity in the brainstem-spinal cord: application of a voltage-sensitive dye.

Although there are several limitations, optical recording techniques are superior to multi-electrode mapping methods, as it is possible to record at large number of points in a small area without destroying the tissue and possible to know relative changes of membrane potentials. Optical recording techniques using voltage-sensitive dyes will be more importantly applied in the study of central respiratory control (e.g., mechanisms of respiratory rhythm generation) in the near future.

Effects of extracellular calcium and magnesium on central respiratory control in the brainstem-spinal cord of neonatal rat.

The influence of extracellular Ca2+ and Mg2+ concentrations ([Ca2+]ECF and [Mg2+]ECF, respectively) on central respiratory control was analyzed using the isolated brainstem-spinal cord of the neonatal rat. Central respiratory activity was recorded from the C4 ventral roots. The depth profile of [Ca2+]ECF below the ventral medullary surface was measured with ion-sensitive electrodes. The gradient in [Ca2+]ECF disappeared about 1 h after changing superfusate Ca2+ ([Ca2+]CSF) from 2 to 0.5 mM, but not even in 2 h after switching to Ca2+-free superfusate. High [Ca2+]CSF (4 mM) or high [Mg2+]CSF (4, 8 mM) decreased respiratory frequency (fR), whereas low [Ca2+]CSF (0.5 mM) increased fR and augmented the respiratory CO2 responsiveness. High [Ca2+]CSF as well as low [Mg2+]CSF (0.5 mM) disturbed respiratory rhythm and pattern, which were markedly restored by high CO2. The depressing effect of high [Ca2+]ECF and the stimulating effect of low [Ca2+]ECF on the medullary neuronal activity were confirmed by perforated patch recordings. These results suggest that [Ca2+]ECF and [Mg2+]ECF determine the excitability of the respiratory neuron network by modulating the neuronal surface potential, transmembrane Ca2+ influx, Ca2+-sensitive cation channel gating, and synaptic transmission. Furthermore, some of these actions appear to be antagonized by CO2/H+.

Oxygen supply and ion homeostasis of the respiratory network in the in vitro perfused brainstem of adult rats.

An in vitro arterially perfused medulla preparation of 3- to 8-week-old rats is described in which synchronous rhythmic activity (frequency 4.5 +/- 1.7 cycles/min, burst duration 3.1 +/- 1.1 s, n = 40) was recorded from hypoglossal (XII), vagal (X), or spinal (C1-2) nerves and from different classes of neurons in the region of the ventral respiratory group (VRG). Stimulation of dorsal X nerve rootlets produced a reversible blockade of rhythmic activity. Under steady-state conditions, tissue oxygen (pO2) in the VRG (depth of 600-1600 microns below the ventral surface) fell from 180 to 40 mmHg. Extracellular K+ activity (aKe) in the VRG was about 0.3 mM higher, calcium concentration ([Ca]e) did not differ, and pH (pHe) was about 0.27 units lower than in the perfusion or superfusion solution (with an aKe of 2.2 mM, a [Ca]e of 1.5 mM and a pHe of 7.4). During inspiratory XII nerve discharges, rhythmic increases of aKe by up to 0.8 mM were detected in the VRG. Perfusion of N2-gassed hypoxic solutions (5-10 min) resulted in a tissue anoxia of the VRG and a reversible cessation of rhythmic activity after 2-7 min. Such anoxia was accompanied by a rise of aKe by up to 35 mM, whereas pHe and [Ca]e fell (from mean levels of 7.17 and of 1.5 mM, respectively) by more than 0.2 pH units and 1 mM. Similar observations were made during a 2- to 5-min arrest of the perfusion pump to simulate ischaemia, whereas significantly larger changes in aKe, pHe and [Ca]e were revealed during an "ischaemia" period of 10 min. The results indicate that the rhythmic activity is generated by the functionally intact respiratory network of the VRG in which neurons are under aerobic conditions and ion homeostasis is not impaired. We conclude that the preparation is an appropriate in vitro model for the analysis of the cellular mechanisms for generation of respiratory rhythm and of metabolic perturbations like anoxia and ischaemia in the mature respiratory network.

Central hypoxic depression of respiration determined by perfusing a recipient cat's carotid bodies with blood from a donor cat.

The respiratory response to sustained hypoxia reflects the net effect of the peripheral chemoreceptor excitation and the central depression. To obtain a pure depressant effect of central hypoxia (alv. PO2, 35-75 mmHg) on respiration, the recipient cat's carotid bodies were vascularly isolated and bilaterally perfused with blood from a donor cat. Cats were anesthetized, vagotomized, paralyzed, and artificially ventilated. Alv. PCO2 was kept at a normal value (iso- and normocapnia) throughout the experiment. The minute integrated phrenic nerve activity (Min. PNA) was used as an indication of inspiratory activity. The Min. PNA at which the recipient and donor cats were both breathing room air (central normoxia and peripheral chemoreceptor normoxia) was taken as 100% (control), and Min. PNA expressed as a percentage of the control was measured in a series of three experiments (A, B, and C): experiment A (normoxia-hypoxia; central normoxia and peripheral chemoreceptor hypoxia), experiment B (hypoxia-hypoxia; central hypoxia and peripheral chemoreceptor hypoxia), and experiment C (hypoxia-normoxia; central hypoxia and peripheral chemoreceptor normoxia). At alv. PO2 of about 50 mmHg, Min. PNA was 169.7 +/- 23.6% (mean +/- SD) in experiment A, 127.3 +/- 22.0% in experiment B, and 71.6 +/- 17.3% in experiment C. Thus, a pure depressant effect of normocapnic hypoxia (alv. PO2 of 50 mmHg) was 42.4% (= 169.7-127.3) or 28.4% (= 100-71.6). At alv. PO2 of about 40 mmHg, Min. PNA was 185.2 +/- 16.8% in experiment A, 146.2 +/- 19.0% in experiment B, and 54.8 +/- 15.2% in experiment C. Thus, a pure depressant effect of normocapnic hypoxia (alv. PO2 of 40 mmHg) was 39.0% (= 185.2-146.2) or 45.2% (= 100-54.8). From these results, we concluded that a pure depressant effect of normocapnic central hypoxia (alv. PO2, 40-50 mmHg) was 28-45% and this was affected by a stimulating drive from the peripheral chemoreceptors.

Developmental changes in the hypoxia tolerance of the in vitro respiratory network of rats.

The functional relation between respiratory activity, extracellular potassium activity (aKe) and tissue oxygen pressure (pO2) was analyzed in vitro in the ventral respiratory group (VRG) of the neonatal brainstem-spinal cord (NB) and the perfused adult brainstem (AB) of rats. In the AB, an aKe increase of up to 35 mM and a reversible blockade of respiratory activity occurred during anoxia periods of 5-10 min. In the NB, respiratory activity persisted during a 60 min anoxia in CO2/HCO3(-)-buffered solutions and aKe increased by less than 1.5 mM. In both preparations, inhibition of glycolysis by iodoacetate led to an irreversible blockade of respiratory rhythm and a delayed increase of aKe by more than 15 mM. We conclude that anaerobic metabolism is sufficient for the maintenance of respiratory activity and potassium homeostasis in the brainstem of the neonatal rat, but not of the adult.

Characteristics of inspiratory inhibition by occlusion of both external carotid and basilar arteries in cats.

Effects of the occlusion of both the external carotid and basilar arteries on the inspiratory activity were studied in anesthetized, vagotomized, paralyzed, and artificially ventilated cats. Integrated phrenic nerve activity was used as an index of the inspiratory activity. Blood pressure in the lingual artery, located downstream from the occluded external carotid arteries, was measured as the arterial pressure of the upper brain stem during occlusion. The basilar artery was occluded at the boundary between the medulla and pons. Occlusions of the external carotid arteries and basilar artery suppressed the phrenic nerve activity to finally disappear within 1 min (phrenic nerve apnea, 45 out of 50 occlusions in 6 cats). The blood pressure in the upper brain stem was 16.6 +/- 5.7 mmHg (mean +/- S.D.) during occlusions. These effects of occlusion on the phrenic nerve activity were also observed during hypercapnia and hypoxia, although they were not so remarkable as those during normocapnia and normoxia. The results indicate that the upper part of the brain stem operates a profound facilitatory mechanism on the medullary inspiratory activity.

Respiratory responses to occlusion or hypercapnic blood injection of the anterior inferior cerebellar artery in cats.

To examine whether the central chemoreceptors of respiration are located in the perfused area of the anterior inferior cerebellar artery (AICA), we occluded arteries or injected hypercapnic blood into arteries in the ventral surface of the medulla in anesthetized, paralyzed, and peripherally chemodenervated cats. Phrenic nerve activity, as an index of respiratory output, was augmented by an injection of hypercapnic blood into the vertebral artery. This vertebral-injection response decreased during bilateral occlusion of AICA. However, responses of phrenic nerve activity to the occlusion of AICA were complicated; activity increased in 19 cats, did not change in 10, and decreased in 9 during occlusion. In experiments with blood of various PCO2 levels being bilaterally injected into AICA, phrenic discharges increased with increases of PCO2. During the injection of constant PCO2 blood into AICA, phrenic response to alveolar PCO2 decreased by 80% compared with the original response. From these results, the blood flow and blood PCO2 level of AICA seemed to be related to the central chemosensitivity for respiration. To examine the perfused area, the ventral surface pH of the medulla was measured with a micro-combination pH electrode (2 mm diameter). During the injection into AICA, pH in the rostral medulla depended on the PCO2 of injected blood, and pH in other areas depended on the PCO2 of systemic blood. Also, histological study of India ink injection into AICA showed that ink-filled vessels were exclusively observed in the rostral medulla. Thus, we conclude that at least part of the central chemoreceptors of respiration are located in the perfused area of AICA, that is, in the rostral medulla.

ECF pH dynamics within the ventrolateral medulla: a microelectrode study.

Using pH-sensitive microelectrodes, we evaluated pH dynamics of extracellular fluid (ECF) within the ventrolateral medulla (VLM) beneath the central chemoceptive areas in anesthetized, spontaneously breathing cats. Static ECF pH was acid in the superficial layers (less than 1 mm), compared with the overlying cerebrospinal fluid pH that became alkaline gradually during the experiments. In the deeper VLM areas (1-3 mm), no systematic gradients of ECF pH were observed. We found various, isolated regions where intravertebral artery injections of CO2-saturated saline evoked acidic shift of ECF pH in the time course analogous to ventilatory augmentation. Those responsive regions were found to be scattered not only in the superficial layers but also in the deeper VLM areas, although many nonresponsive regions were also intermingled among them. Occlusions of the principal vessels supplying the tested VLM regions diminished but failed to abolish the ECF pH responses to the CO2 loadings, suggesting a collateral blood flow by fine pial vessels. The present study suggests a possibility that the pH-dependent central chemoreceptors, if any, would be scattered in the deeper VLM areas as well as the superficial layers.

Possible locations of pH-dependent central chemoreceptors: intramedullary regions with acidic shift of extracellular fluid pH during hypercapnia.

Using liquid membrane pH microelectrodes, we evaluated rapid and transient changes in extracellular fluid (ECF) pH within the medulla during vertebral artery injections of CO2-saturated saline (0.5 ml) in anesthetized (Dial-urethane), spontaneously breathing cats. We found intramedullary regions where ECF pH shifted to the acid side in the time course analogous to respiratory excitation during the CO2 loadings: the acidic shift occurred just before the respiratory excitation. Since most of the tested regions showed no or few changes in ECF pH, the responsive regions are thought to be specific local environments fitting the central chemoreceptors. Forty (85%) out of the 47 responsive regions were found to be scattered in the ventrolateral medulla, i.e. a long narrow zone extending from the ventrolateral surface to the ventral respiratory group (VRG) areas where inspiratory or expiratory activity was frequently recorded. The responsive regions were not necessarily restricted to the superficial ventral layers. We were also able to find the responsive regions in the dorsal area ventral to the nucleus tractus solitarii, though they were fewer in number (7/47). The distributions corresponded rougly to the areas where we had previously identified the tonically firing neurons excited exclusively by stimulation of the central chemoreceptors. These results indicate a possibility that the pH-dependent central chemoreceptors, if any, would be located within the regions demonstrated in this study.

Inspiratory response to occlusion of arteries in the ventral surface of the medulla in anesthetized cats.

By the occlusion of arteries in the ventral surface of the medulla, the blood supply to the central chemoreceptor for respiration was examined in anesthetized, paralyzed and peripheral chemodenervated cats. Phrenic nerve activities (P.N.A.), as an index of the respiratory center output, increased with an injection (3 ml/min, 10 sec) of hypercapnic blood (PCO2 = 104.5 mmHg) into the vertebral artery (VA injection response). The VA injection responses during occlusion of arteries in the ventral surface of the medulla were classified into three groups: 1) The response disappeared by the bilateral occlusion of the anterior inferior cerebellar arteries (AICA) in 8 out of 29 cats. 2) The response disappeared by the occlusion of both AICA and the posterior inferior cerebellar arteries (PICA) in 9 cats. 3) The response did not disappear in spite of the additional occlusion of several branches from the basilar artery in 11 cats, although the response had diminished. These different results may be due to the complexity of the central chemosensitive structure or of the central vascular system. However, among arteries the AICA blood flow seemed to be most preferentially related to the VA injection response. Thus, at least a part of the central chemosensitive structure may be located in the area perfused by the AICA.

Effect of hypercapnic blood injection into the vertebral artery on the phrenic nerve activity in cats.

Hypercapnic blood was injected into the vertebral artery in anesthetized and paralyzed cats. The stimulating effect on the phrenic nerve activity was dependent on the injection rate, duration, and PCO2 level of the injected blood. The time delay from the start of injection to the onset of increase in phrenic nerve activity was inversely proportional to both the injection rate and the PCO2 of injected blood.

Effect of inactivation of carotid sinus nerve by cold block on phrenic nerve activity in cats.

The effect of the elimination of input via the carotid chemoreceptor on respiratory output was examined quantitatively in anesthetized, vagotomized, and paralyzed cats. The integrated phrenic nerve activity (PNA) was recorded as an indication of output of the respiratory center. Also, the elimination of the carotid chemoreflex drive was repeatedly done by a cold block of the carotid sinus nerve at various PCO2 levels during hyperoxia, normoxia, and hypoxia. The blockade induced a reduction in PNA at each PCO2 level in every PO2 group. If the highest PNA value recorded at a high PCO2 in each PO2 condition was assigned a value of 100%, the reduction of the PNA by the blockade, i.e., the respiratory effect of the carotid chemoreflex drive, would be slightly larger during normoxia (16%) than during hyperoxia (8.7%), but would be independent of PCO2. During hypoxia, this chemoreflex effect was about 40% of a low PCO2, and decreased with increments of PCO2, finally reaching about 20% of a high PCO2 level. Furthermore, the relative contribution of the carotid chemoreceptor to respiratory output, expressed as the ratio of the PNA reduction during blockade to the PNA before blockade, was inversely proportional to both PO2 and PCO2. It is concluded that the interaction of the peripheral and central chemoreceptor drive is hypoadditive at moderate and high PCO2 levels in anesthetized cats, and this interaction is emphasized by central hypoxia.

Effect of arterial [H+] on threshold PCO2 of the respiratory system in vagotomized and carotid sinus nerve denervated cats.

Phrenic nerve discharges were recorded as an output of respiratory activity in anaesthetized, vagotomized cats immobilized by gallamine and artificially ventilated with room air. 2. With the carotid sinus nerve (c.s.n.) intact or denervated, PcO2 threshold levels (Pth, CO2) were determined at arterial pH, varied between 7.0 and 7.6 ([H+] 25-100 nM) by successive intravenous infusions of 0.5 N-HCl or 1.0 M-NaHCO3. Ventilation was increased stepwise to induce a successive decrease in end-tidal PCO2. Pth, CO2 was defined as the level of end-tidal PCO2 at which phrenic discharges ceased. 3. With the c.s.n. intact, Pth, CO2 decreased linearly upon increasing arterial [H+]. The mean regression line, calculated from seven cats, was Pth, CO2 =-0.37 [H+] + 34.33. A similar inverse relationship was observed with the c.s.n. denervated. However, the slope of the regression line was significantly smaller, the mean regression line/eleven cats) being Pth,CO2 =-0.18 [H+]+ 35.06. 4. The relative contributions of arterial [H+] and PCO2 in stimulating the peripheral and central chemoreceptors could be estimated quantitatively. Arterial [H+] appears to be almost equally effective on both peripheral and central chemoreceptors; PCO2 acts exclusively on the central chemoreceptors. 5. Thus, the additive theory regarding the induction of respiratory activity by arterial [H+] and PCO2 was confirmed. In addition, the H+ drive was shown to be able to affect respiratory activity even in the absence of the peripheral chemoreceptors.

Analysis of dynamic change in phrenic nerve activity following a sudden decrease in alveolar carbon dioxide.

The effects of a sudden decrease in alveolar CO2 concentration (FCO2) on phrenic nerve activity (PNA) were studied in anesthetized and paralzyed cats. The vago-sympathetic and carotid sinus nerves were sectioned. The peak of integrated PNA was used as an index of the central inspiratory activity. 1) FCO2 decreased immediately and curvilinearly after hyperventilation. However, PNA did not change during the initial period of hyperventilation. After this time delay (Dt), the PNA began to decrease linearly over a certain period and finally disappeared. Toff, the time from the onset of hyperventilation to the disappearance of PNA, was in the range of 30-250 sec. This was related to both the level of FCO2 in a control period and the rate of decrease in FCO2 during hyperventilation, but Dt was mostly independent of these values. Mean (with SD) Dt was 10.68 +/- 7.01 sec (n = 68, from 6 cats). 2) VCO2, the quantity of CO2 eliminated through the lung during Toff, was measured in each experimental run. The VCO2 was directly proportional to the level of control FCO2 and, at a given level of control FCO2, was almost identical, irrespective of the different rates of decrease in FCO2. 3) We concluded that Dt is the time required for a change in the hydrogen ion concentration, [H+], in the brain interstitial fluids bathing the central chemosensitive structures, and that the central inspiratory activity, in the absence of the peripheral chemoreceptors, will be a single function of the [H+] in these fluids.