JN Fuel your research with LabChart
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

J Neurophysiol 79: 2383-2393, 1998;
0022-3077/98 $5.00
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Koos, B. J.
Right arrow Articles by Kruger, L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Koos, B. J.
Right arrow Articles by Kruger, L.

The Journal of Neurophysiology Vol. 79 No. 5 May 1998, pp. 2383-2393
Copyright ©1998 by the American Physiological Society

Thalamic Locus Mediates Hypoxic Inhibition of Breathing in Fetal Sheep

Brian J. Koos, Andrew Chau, Masahiko Matsuura, Oscar Punla, and Lawrence Kruger

Departments of Obstetrics and Gynecology and Neurobiology, Nicholas S. Assali Perinatal Research Laboratory, Brain Research Institute, UCLA School of Medicine, Los Angeles, California 90095-1740

    ABSTRACT
Abstract
Introduction
Methods
Results
Discussion
References

Koos, Brian J., Andrew Chau, Masahiko Matsuura, Oscar Punla, and Lawrence Kruger. Thalamic locus mediates hypoxic inhibition of breathing in fetal sheep. J. Neurophysiol. 79: 2383-2393, 1998. The effects of lesions rostral to the brain stem on breathing responses to hypoxia were determined in chronically catheterized fetal sheep (>0.8 term). These studies were designed to test the hypothesis that the diencephalon is involved in hypoxic inhibition of fetal breathing. As in normal fetuses, hypoxia inhibited breathing with transection rostral to the thalamus or transection resulting in virtual destruction of the thalamus but sparing most of the parafascicular nuclear complex. Neuronal lesions were produced in the fetal diencephalon by injecting ibotenic acid through cannulas implanted in the brain. Hypoxic inhibition of breathing was abolished when the lesions encompassed the parafascicular nuclear complex but was retained when the lesions spared the parafascicular nuclear region or when the vehicle alone was injected. A new locus has been identified immediately rostral to the midbrain, which is crucial to hypoxic inhibition of fetal breathing. This thalamic sector involves the parafascicular nuclear complex and may link central O2-sensing cells to motoneurons that inhibit breathing.

    INTRODUCTION
Abstract
Introduction
Methods
Results
Discussion
References

In fetal sheep (>0.8 term), breathing occurs in episodes associated with rapid eye movements (REM) and low-voltage electrocortical activity. While breathing occurs during this REM-like sleep state, it is normally absent in high-voltage states that otherwise resemble quiet sleep (Dawes et al. 1972). Although stimulated by respiratory acidosis, acute reductions in fetal PaO2 of 8-10 mmHg eliminate breathing activity almost completely (Boddy et al. 1974). This virtual absence of breathing occurs in fetuses after denervation of the peripheral arterial chemoreceptors and bilateral section of the vagosympathetic trunks (Koos and Sameshima 1988), but the depression is abolished by supramedullary brain stem transection (Dawes et al. 1983; Koos 1985). Thus hypoxic inhibition results from the direct effects of hypoxia on the fetal brain.

Previous brain lesion studies indicated that the inhibitory mechanism includes regions in the rostral pons (Dawes et al. 1983; Gluckman and Johnston 1987) or caudal mesencephalon (Dawes et al. 1983) but not in the cerebral cortex (Dawes et al. 1983; Ioffe et al. 1986). More recently, transection of the midbrain immediately caudal to the red nucleus in one fetus arrested the inhibitory effects of hypoxia on breathing, indicating that regions rostral to the brain stem may contribute to hypoxic respiratory depression (Koos et al. 1992).

This study was designed to test the hypothesis that the neural network mediating hypoxic inhibition involves a contribution rostral to the midbrain. The findings implicate a specific sector of the caudal thalamus, a region associated with sleep regulation postnatally, in mediating hypoxic inhibition. A brief description of some of these observations has been given previously (Koos et al. 1996).

    METHODS
Abstract
Introduction
Methods
Results
Discussion
References

Twenty pregnant ewes (singleton and twin, Rambouillet-Columbia breed) were operated on under halothane (2-5% in O2) anesthesia at 120 days of gestation (~0.8 term). Polyvinyl catheters were inserted in the right brachiocephalic trunk, carotid artery, and external jugular vein of the fetus with other catheters placed in the trachea and amniotic sac (Koos et al. 1987). Bipolar stainless steel electrodes were implanted on the lateral orbital ridge to record eye movements and on the cranial dura to record electrocortical activity.

Brain transection

In five fetuses, the rostral brain was transected by inserting a 3-mm wide metal spatula through a trephine hole in the parietal bone. The point of insertion varied from 1 to 7 mm caudal to the coronal suture with the spatula approximately directed toward the vertical plane of the coronal suture at the base of the brain. These preliminary studies helped establish the approximate region of the diencephalon essential to be intact for preserving hypoxic inhibition.

Guide cannula placement

In 15 fetuses, two stainless steel guide cannulas (CMA/Microdialysis, Stockholm) with a shaft length of 2.0 cm (1.0 mm OD) were stereotaxically directed (Koos et al. 1994b) toward the caudal diencephalic region identified by the brain transection studies as containing neurons inhibitory to breathing. The cannulas were inserted ~3 mm lateral to midline with the tip extending 1.0 cm below parietal bone and were anchored to the calvaria with dental acrylic that, in turn, was glued with cyanoacrylic ester cement to a silicone rubber window in the uterus and flank of the ewe (see Fig. 1 of Koos et al. 1994b). The guide cannulas were exteriorized through the silastic rubber window, which allowed placement of needles for microinjections in the fetal brain several days after surgery (Koos et al. 1994b). All surgical procedures and experiments received prior approval from the Chancellor's Animal Research Committee.


View larger version (34K):
[in this window]
[in a new window]
  		FIG. 1. Outline drawings of representative Nissl-stained histological sections at a level optimally depicting lesion(s) that failed to abolish hypoxic inhibition of breathing. Each specimen is indicated by animal number followed by section number. Fetus 937 represents a sagittal section of a thalamic transection. The remaining cross-sectional drawings represent lesions induced by ibotenic acid. Needle tracks are black, heavy gliosis is dark gray or stippled black when necrotic, lighter gliosis is lightly stippled. The magnification bar represents 1 mm. AL, anterior limbic cortex; Arc, arcuate n.; AD, anterodorsal n.; AM, anteromedial n.; AV, anteroventral n.; cc, corpus callosum; cg, central (periaqueductal) gray; cm, central medial n.; cp, cerebral peduncle; epi, epiphysis (pineal); Hb, habenular n. (l, lateral; m, medial); IC, inferior colliculus; MD, mediodorsal n.; mt, mammillothalamic tract; pc, paracentral n.; Pf, parafascicular n. (includes n. centrum medianum); Po, posterior n.; Pta, anterior pretectal n.; R, thalamic reticular n.; re, reuniens n.; rh, rhomboid n.; SC, superior colliculus; sn, substantia nigra; sPf, subparafascicular n.; VB, ventrobasal n.; ZI, Zona incerta; VA, ventral anterior n.

Fetal arterial and tracheal pressures were measured with pressure transducers (Cobe Laboratories, Lakewood, CO) and were referenced to amniotic fluid pressure. Heart rate was determined by a cardiotachometer triggered by the arterial pulse pressure. Tracheal and arterial pressures, heart rate, electrooculogram (EOG), and electrocorticogram (ECoG) were displayed on a chart recorder (Model 7E, Grass Instruments, Quincy, MA). Heart rate, arterial, and tracheal pressures were sampled at 100 Hz with breathing movements identified from on-line monitoring of tracheal pressure (Koos et al. 1992). Minute averages of heart rate, mean arterial pressure, inspiratory time, breath interval, and breath amplitude were stored on disk. Arterial blood gas tensions and pH were determined with the use of blood gas electrodes (Instrumentation Laboratories, Lexington, MA) with values corrected to fetal temperature (39.5°C).

Experimental protocol

The fetus and ewe were allowed at least 4 days to recover before starting experiments. Fetal breathing responses to isocapnic hypoxia were determined by having the ewe breathe for 1 h a hypoxic gas mixture (9% O2, 3% CO2, 88% N2) (Koos et al. 1987). Only one hypoxia study was conducted per day to minimize carry-over effects.

After establishing hypoxic inhibition, ibotenic acid (IBO) was used to produce lesions in the fetal diencephalon. By stimulating glutamate receptors, IBO causes local destruction of neurons, while leaving fibers of passage and the vasculature relatively intact (Guldin and Markowitsch 1982; Schwarcz et al. 1979). A stainless steel needle (shaft: 0.66 mm OD, 38 mm long) connected to microbore tubing (1.2 µl/10 cm) was inserted in a guide cannula, placing the tip in the diencephalon. Using a microinjection pump (CMA/200, CMA/Microdialysis, Stockholm), synthetic cerebrospinal fluid (CSF) (Koos 1985) containing IBO (0.19 M) was infused at 0.33 µl/min for a total volume per injection of 0.4-2 µl. These injection volumes were within the range used for creating localized lesions in adult cats (Denoyer 1986; Marini et al. 1992), which have a brain volume considerably smaller than that of near-term fetal sheep. Because multiple microinjections have been beneficial in previous studies (Jarrard and Meldrum 1993), repeated bilateral symmetrical microinjections were generally carried out with the interval between them varying from 0.5 to 4 h and not exceeding four injections on a single day. One or more days after IBO administration, fetal breathing responses to hypoxia were again determined.

Control experiments were performed in five fetuses in which CSF alone was microinjected in the approximate sector where IBO had abolished hypoxic inhibition in other fetuses. These microinjections were performed using the same coordinates and in the same manner as those for IBO: two injections per side in four fetuses and three per side in one fetus. Fetal breathing responses to isocapnic hypoxia were determined before and after administering the vehicle for IBO.

Brain preservation and analysis

In 9 of 11 animals, the ewe and fetus were killed 3-8 days after IBO administration; 2 fetuses were killed 2 days after IBO microinjections. The fetal brains were fixed in situ by arterial perfusion with a buffered 10% Formalin solution. Brains that had been transected were cut in the sagittal plane at 50 µm, with every third section stained with Cresyl violet. Brains with IBO lesions were cut in coronal sections at 35 µm, and alternate sections were stained with Cresyl violet. Lesion size was determined by cell destruction and gliosis, which was evident even in the two fetuses euthanized 2 days after IBO administration. Because IBO creates clearly defined brain lesions, the boundaries of destruction were easily identified.

The anatomic account of brain lesions was provided by a neuroanatomist (Kruger) blind to study results and was based on the description of the sheep thalamus by Rose (1942), but using altered nomenclature to conform more closely to contemporary usage. Recognition of the centrum medianum is only achieved in those species where it is surrounded by a fibrous capsule, resulting in the current practice of including it with the parafascicular nucleus (Pf), Rose's "postmedial group," distinct from the more rostral medial, midline nuclei, which he called the "central commissural system," the nn rhomboidalis, reuniens, and centralis medialis of most authors (see Kruger et al. 1995). We also recognize the subparafascicular nucleus (sPf) of modern descriptions, acknowledging that an implied homology or functional grouping has not been established. Rose also recognized a n interventralis, which we have labeled nn ventromedial (VM) and reuniens (re); the latter (re) is called medioventral in some descriptions. Rose called the tactile nucleus of the sheep the "ventrobasal complex" (VB) comprised of external and arcuate divisions; his arcuate division includes a distinctive small-celled nucleus, the homologue of the gustatory nucleus in other species often called the parvocellular subnucleus (VPMpc) (Paxinos and Watson 1986).

Data analysis

Breathing was identified from characteristic negative deflections in tracheal pressure. Because of the episodic nature of fetal breathing, the number of minutes per hour of breathing was determined as a measure of breathing activity. Breathing was determined to be present if it occurred in >= 20 s of each 1-min epoch (Koos and Matsuda 1990). The incidence of REM was determined in a similar manner.

The expected incidence of breathing in normal fetuses during hypoxia was determined from responses observed in 18 fetuses, 9 from previous studies (Koos et al. 1987, 1994) and 9 from the present investigation before the administration of IBO. According to the 95% prediction limits for breathing incidence in these fetuses, the lower limit of breathing expected during normoxia was 14 min/h, representing 23% of the time; the upper limit predicted during hypoxia was 10 min/h, corresponding to 17% of time. On the basis of these limits, breathing was judged to be inhibited if breathing occurred <12 min/h (20% of the time) during hypoxia.

Statistical analysis

For ECoG activity, REM, and breathing, the incidence during the 4-h control period was averaged for comparison with the respective values measured during the experiment. Mean values of physiological measurements for individual epochs during and after hypoxia were compared with control measurements preceding hypoxia using repeated measures of analysis of variance (ANOVA) methods with post hoc comparison of means performed using Tukey's least-significant difference criterion. For hypoxia studies, repeated measures of ANOVA were also carried out over time with the presence or absence of brain lesions as the within-animal factors. Single comparisons of means were carried out using Student's t-test. A log transformation of the data was carried out before analysis when it normalized a skewed distribution. Differences were considered significant at the P < 0.05 level. All values are expressed asmeans ± SE.

    RESULTS
Abstract
Introduction
Methods
Results
Discussion
References

Brain transection

Three of the five transected fetal brains were preserved for histological examination. In one fetus, a large rostral diencephalic transection extended to the base of the hypothalamus, which destroyed enough of the thalamic vasculature to render most of the thalamus necrotic but fortuitously spared the habenular nuclei and much of the caudal portion of the parafascicular nuclei, although there was a rostral cap of heavy gliosis with mild gliosis in the subparafasicular sector (Fig. 1). The section was through the anterior thalamic nuclei of the second fetus, and it was 2-3 mm rostral to the thalamus in the third. The other two fetuses did not breathe, presumably because of severe trauma caused by brain interruption. Unfortunately, this could not be confirmed by histological evaluation because the fetuses died before the brains could be preserved satisfactorily.

In the three fetuses studied, arterial blood gases andpH were within normal limits (PaO2, 20.3 ± 1.2 mmHg,mean ± SE; PaCO2, 44.7 ± 2.4 mmHg; pHa, 7.351 ± 0.010), as was the mean breathing incidence (23 ± 3.6 min/h), before inducing hypoxia. Two types of responses to isocapnic hypoxia (PaO2, ~12 mmHg) were observed. In the first, breathing persisted during hypoxia [control (c): 28, hypoxia (h): 25 min/h] at reduced amplitude in a fetus whose brain was transected through the anterior thalamic nuclei (982). In the second response variant, hypoxia reduced the incidence of breathing (<9 min/h) with transection rostral to the thalamus (393) or with the interruption destroying almost all the thalamus except for most of the parafascicular region (937, Fig. 1). The latter was the largest thalamic lesion that damaged a sector of the parafascicular region yet spared hypoxic inhibition.

Although other thalamic neurons may play a role in modulating breathing during hypoxia (as in 982), the observations in fetus 937 revealed that a fortuitously preserved, relatively small sector of the posterior thalamus, identified in sagittal sections in which virtually the entire thalamus was destroyed, appeared to contain neurons essential to reducing the incidence of fetal breathing during hypoxia. Guided by this remarkable observation based on a very large lesion, subsequent studies were aimed at producing relatively small lesions in the caudal thalamus using "excitotoxic" IBO injections to destroy neurons without concomitant interruption of fiber tracts.

IBO injections

Microinjections of IBO were performed in 11 of the 15 fetuses with guide cannulas with control injections of the vehicle (CSF) in the remaining 4 fetuses. Fetal breathing responses to hypoxia were determined before and after the administration of this glutamate receptor agonist.

EFFECTS OF HYPOXIA BEFORE LESIONS. The effects of hypoxia were determined satisfactorily before IBO administration in 9 of 11 fetuses studied. During the control period, the mean preductal arterial blood gas and pH values were within the normal range (PaO2, 24.2 ± 1.3 mmHg; PaCO2, 49.8 ± 1.1 mmHg; pH, 7.343 ± 0.010), whereas during hypoxia, mean PaO2 fell by ~10 mmHg, and was associated with a slight (~3 mmHg) fall in PaCO2 accompanied by a progressive decline in pH to 7.289 ± 0.019. Fetal heart rate averaged 170 ± 5.6 beats/min during the control period and was not significantly affected by hypoxia. During the phase of acute O2 deprivation, the mean arterial pressure rose significantly by 3.9 ± 1.8 mmHg compared with the control value of 40.9 ± 2.5 mmHg.

As shown in Figs. 2 and 3, low-voltage ECoG averaged 35 ± 4.7 min/h during the control period and 23 ± 4.2 min/h during hypoxia, differences that were not statistically significant (P = 0.06). Hypoxia also failed to significantly alter the incidence of high-voltage electrocortical activity (control: 16 ± 3.3; hypoxia: 19 ± 1.7 min/h).


View larger version (23K):
[in this window]
[in a new window]
  		FIG. 2. Ibotenic acid (IBO) lesions that abolished hypoxic inhibition. Solid line depicts incidence of rapid eye movements and breathing in 5 fetuses before administering IBO. Satisfactory electrocorticogram (ECoG) recordings were obtained in only 1 fetus in this group; therefore the incidence of high (HV) and low (LV) ECoG is not shown for these experiments. Dotted line represents ECoG (5 fetuses), rapid eye movements (6 fetuses), and breathing (7 fetuses) responses after IBO injections. *P < 0.05 compared with the respective control mean. †P < 0.05 compared with mean value during hypoxia before IBO administration.


View larger version (24K):
[in this window]
[in a new window]
  		FIG. 3. IBO lesions with retention of hypoxic inhibition. The incidence of HV- and LV-ECoG, rapid eye movements, and breathing is shown before () and after (···) IBO administration. Hypoxia is shown by the stippled vertical bar. *P < 0.05 compared with the respective control mean.

REM averaged 30 ± 1.9 min/h during the control period but fell significantly to 6 ± 4.7 min/h during hypoxia (Figs. 2 and 3). After fetal PaO2 was returned to normal levels, the incidence of REM returned to values within the range of those recorded during the control period.

Fetal breathing averaged 30 ± 4.1 min/h during the 4 h preceding the onset of acute O2 deprivation. Hypoxia nearly eliminated breathing activity; this respiratory depression was reversed when fetal PaO2 returned to normal values (Figs. 2 and 3). Mean inspiratory time, breath interval, and breath amplitude averaged 0.52 ± 0.03 s, 1.51 ± 0.12 s, and2.3 ± 0.1 mmHg, respectively, and were not significantly affected by hypoxia in the 9 of 11 animals studied satisfactorily.

IBO LESIONS. IBO was injected into the brain of 11 fetal sheep. Two to four microinjections were performed bilaterally at one site on two to three separate days in four fetuses (633, 198, 69, and 23) and at two different sites separated by 3 mm along the axis of the needle in two fetuses (617 and 139). In one fetus (489), only two symmetrical injections of IBO were performed. The number and sites of injections were limited by technical problems in this difficult preparation. For example, fetus 308 received three microinjections through only the left guide cannula.

HYPOXIC RESPONSE AFTER LESIONS. Fetal responses to isocapnic hypoxia were tested in 11 fetuses at 1-8 days after completion of the IBO injections. Before hypoxia, fetal arterial blood gases and pH were within normal limits (PaO2, 24.2 ± 1.4 mmHg; PaCO2, 50.0 ± 0.8 mmHg; pH,7.353 ± 0.007); during isocapnic hypoxia PaO2 was reduced by ~10 mmHg with an associated decline in pH to7.271 ± 0.019. Similar cardiovascular responses to hypoxia occurred in fetuses irrespective of lesion site; thus the responses of all lesioned fetuses are presented together. Within 10 min of inducing hypoxia, mean heart rate fell by >30 beats/min from the control of 169 ± 5.9 beats/min and remained significantly reduced during the first 40 min of O2 deficiency, a response that is expected in fetuses at this gestational age (Boddy et al. 1974; Koos et al. 1987). Mean arterial pressure increased by ~4 mmHg during hypoxia compared with the control of 47.2 ± 2.2 mmHg.

Fetal breathing was generally unaffected by the IBO lesions, but prolonged episodes of breathing were observed after the microinjections in two fetuses (69 and 617), with virtually continuous breathing for up to 9 and 15 h, respectively. In these fetuses, breathing was present during high-voltage ECoG, and this activity appeared to be independent of REM. Episodic breathing returned within 24 h after the injections. The "control" values of Figs. 2 and 3 show that the incidence of high- and low-voltage ECoG, REM, and breathing activity after IBO administration under normoxic conditions did not differ significantly from those before administration of the excitotoxin. On the basis of breathing responses to acute oxygen deficiency (see METHODS), the fetuses could be separated into two groups: those breathing >20% of time during hypoxia (i.e., hypoxic inhibition abolished) and those breathing <20% of time (hypoxic inhibition retained).

Hypoxic inhibition abolished. In 7 of the 11 lesioned fetuses, the incidence of fetal breathing was not significantly affected by hypoxia. In these fetuses, hypoxia did not alter the mean incidence of low- and high-voltage ECoG activity (Fig. 2). Although the overall incidence of REM was not affected by hypoxia, responses varied in individual fetuses. For example, eye movements persisted with increased amplitude during virtually the entire hour of hypoxia in 617, whereas they increased in amplitude, but not incidence, during hypoxia in 23. Acute O2 deprivation reduced the amplitude and incidence of eye movements in 489 (Fig. 4).


View larger version (15K):
[in this window]
[in a new window]
  		FIG. 4. ECoG, electrooculogram (EOG), and tracheal pressure measurements (PT) during hypoxia in fetus 489.

Mean breathing incidence during hypoxia was 36 ± 6.0 min/h, similar to the control mean of 35 ± 6.9 min/h. Breathing generally remained episodic during hypoxia, although it was present during virtually the entire hour of hypoxia in 617. During hypoxia, breathing occurred with low- and high-voltage ECoG in three fetuses with larger lesions (23, 617, and 633), but only in low-voltage states in fetus 198 and in fetus 489 with the smallest lesion (Fig. 4). The ECoG was not recorded in the other two fetuses.

The normal association of fetal breathing with low-voltage ECoG activity indicates that breathing is affected by changes in electrocortical state apart from direct effects of hypoxia. In fetus 489 (Fig. 4), breathing was present during ~92% of the low-voltage period during hypoxia, which was virtually the same (94%) as that for the normoxic control period. Thus the slight reduction in breathing during hypoxia (normoxia: 33 min/h; hypoxia: 22 min/h) appeared to be associated with a decrease in the incidence of low-voltage ECoG activity (normoxia: 33 min/h, hypoxia: 24 min/h), as seen in Fig. 4. Hypoxia did not significantly affect mean inspiratory time [control (c): 0.58 ± 0.02 s; hypoxia (h): 0.55 ± 0.03 s], breath interval (c: 1.27 ± 0.27; h: 1.51 ± 0.29 s) or breath amplitude (c: 3.0 ± 0.3; h: 2.3 ± 0.1 mmHg).

The smallest lesion that eliminated hypoxic inhibition is illustrated in photomicrographs from fetus 489 (Fig. 5). The bilateral lesions are largely symmetrical, exhibiting extensive gliosis and virtual elimination of neurons in a large portion of Rose's (1942) medial group, but most specifically his "postmedial" group [Rose's separated medial group, which includes the mediodorsal (MD) and paratenial nuclei are only minimally involved]. A small ventrolateral sector of the MD nucleus was also degenerated on both sides. The postmedial group of Rose appears to constitute the critical zone for this analysis (detailed comments reconciling it with customary current usage is given in METHODS). The n reuniens (re) (or medioventral in some descriptions) is also largely involved, whereas the n ventromedialis (VM) is minimally damaged on only one side. Thus the lesion in 489 destroys a large portion of the parafascicular (Pf) nuclear complex (including the lateral wing constituting the centrum medianum and the caudal ventral subparafascicular component), whereas sparing most of VM, arcuate n. (Arc), and MD. A sector of the midline nuclei, especially central medial (cm), paracentral (pc), rhomboid (rh), and re are extensively destroyed in their caudal extent, but large portions of these are intact in rostral sections. Given the large size of the Pf complex in sheep, the lesion in this crucial specimen is remarkable in encompassing most of this structure with relatively minor involvement of small sectors of adjacent nuclei (Fig. 5).


View larger version (164K):
[in this window]
[in a new window]
  		FIG. 5. Photomicrographs of 2 levels (A ~0.4 mm rostral to B) of Nissl-stained transverse sections of the thalamus in 489 illustrating the smallest thalamic lesion that eliminated hypoxic inhibition (magnification bar = 1 mm). Regions of IBO destruction are intensely gray due to neuronal destruction and extensive gliosis. Pul, pulvinar n.; VM, ventromedial n.

There were six additional lesions that interfered with hypoxic inhibition; all involving substantial damage to the parafascicular complex bilaterally (Fig. 6), in crudely approximate symmetry, but also exhibiting some gliosis in adjoining nuclei. The accuracy of the localization, even when nuclear outlines are obscured by extensive neuronal degeneration, is secured by the ease of identifying the fasciculus retroflexus (rf), from which the surrounding Pf nucleus derives its name. Although the ventromedial, arcuate, and anterior pretectal nuclei are involved in some of the lesions, they are spared in others and thus do not appear to be implicated in hypoxic inhibition.


View larger version (46K):
[in this window]
[in a new window]
  		FIG. 6. Outline drawings of Nissl-stained transverse sections at an optimal level depicting lesions that abolished hypoxic inhibition of breathing. Light gliosis is lightly stippled; heavy gliosis is dark gray. Needle tracks are black. Animal number is followed by section number(s). Magnification bar represents 1 mm. AH, amygdalo-hippocampal area; cl, central lateral nucleus; fo, fornix; mm, medial mammillary n.; ot, optic tract; Ptm, medial pretectal area; rf, retroflex fascicle.

Special attention was paid to the nearby ventral tegmentum of the pons because of a recent report (Waites et al. 1996) implicating the red nucleus in the neonatal hypoxic response, but in no case did any lesion or trace of gliosis encroach within 2 mm of the red nucleus, although the ventrocaudally subjacent central gray is slightly involved in some (but not all) of the larger lesions involving the subparafascicular nucleus.

In fetuses 69 and 617, histological analysis also showed that the thalamic lesions were several millimeters dorsal to the hypothalamus. Although clearly beyond the limits of the most generous estimate of the lesions, activated caudal hypothalamic neurons (Ryan and Waldrop 1995) cannot be eliminated as the cause of the continuous breathing induced by IBO in these fetuses.

Hypoxic inhibition retained. The second group consisted of four fetuses in which IBO administration failed to alter the depressing effects of hypoxia on fetal breathing (Fig. 3). In these fetuses, hypoxia did not significantly affect the incidence of low- or high-voltage ECoG. Hypoxia inhibited eye movements, except in fetus 163 with lesions underlying the cingulate gyrus in which eye movements were continuously present with increased amplitude during the entire 60 min of acute O2 deficiency.

Those brain lesions that did not substantially interfere with hypoxic inhibition (Fig. 1) enabled more secure delimitation of the critical zone, but the most valuable case was 937, the transection that spared most of the parafascicular group with destruction of most of the remainder of the thalamus. Two lesions that involved smaller portions of the Pf complex were largely asymmetric (308 and 66) and were also without effect. A large rostral lesion destroying much of the midline nuclei on one side (511) and another involving part of the intralaminar nuclei (66) also were ineffective. A control lesion (163) of the overlying anterior limbic cortex (AL), which was damaged extensively by needle penetrations for all thalamic lesions, also failed to alter hypoxic inhibition. Thus lesions of adjacent structures without extensive destruction of the parafascicular group were eliminated for implication in hypoxic inhibition.

The anatomic analysis indicates that the stereotaxic placement of lesions in the fetal diencephalon were only crudely reproducible. Because twins can be smaller than singletons, some of the variability in the lesions derived from the use of both singletons and twins for this study; but much of the difficulty is due to the inherent imprecision in fetal stereotaxic methodology. Despite these limitations and the necessarily small number of cases, we were able to abolish hypoxic inhibition of breathing in 7 of 11 fetuses.

Microinjections of CSF

Bilateral microinjections of synthetic CSF alone were carried out in the thalamus in 5 of the 11 fetuses with guide cannulas. The site of injection was determined by needle tracks on histological examination. Isocapnic hypoxia (PaO2 ~12 mmHg) inhibited both REM and breathing activity before and after microinjections of CSF, the vehicle for IBO. The needle tracks terminated in Pf (2 fetuses), lateral to sPf (1 fetus), sPf (1 fetus), mediodorsal n (1 fetus), and ventrobasal n (1 fetus). Thus the injections were within (2 fetuses) or proximate to (3 fetuses) the specific sector of the brain (Pf) that had been identified by IBO lesions as being implicated in hypoxic inhibition of fetal breathing. This conclusion is further strengthened by one observation (139) in which CSF injections in Pf, as indicated by the position of needle tracks, were without effect, but lesions produced by subsequent bilateral injections of IBO acid using the same depth of needle insertion as for CSF injections abolished hypoxic inhibition of fetal breathing.

    DISCUSSION
Abstract
Introduction
Methods
Results
Discussion
References

Thalamus and breathing

HYPOXIC INHIBITION. The present work reveals that the thalamus is involved in hypoxic inhibition of fetal breathing. The fortuitous results with brain transection suggested that the central core of the caudal thalamus, containing the parafascicular nuclear complex, was important for the depression of breathing, and subsequent experiments with large and small lesions induced by IBO provided consistent results: bilateral lesions that encompassed the parafascicular region abolished the inhibition, while those that left this region relatively intact did not.

As part of the rostral pole of the "reticular activating system," the "nonspecific" intralaminar nuclei participate in tonic cortical activation associated with increased discharge rates in desynchronized states of wakefulness and REM sleep (Bentivoglio et al. 1991). The parafascicular portion of the intralaminar region also integrates sensorimotor function (Bentivoglio et al. 1991). Receiving fibers from hypothalamic nuclei regulating sleep (Simerly and Swanson 1988) and raphe nuclei, cells within the parafascicular area project to the cerebral cortex, striatum, subthalamus, and back to the raphe nuclei. Monosynaptic projections extend to other brain stem nuclei, including midbrain cuneiformis, ventral parabrachial nucleus, deep mesencephalic, and gigantocellular reticular nuclei (see Mileikowsky et al. 1994). Electrical stimulation of Pf inhibits spontaneous motor activity (Mileikowsky et al. 1994), and the present findings indicate that this thalamic sector depresses fetal breathing during acute O2 deficiency.

RELATION TO PERIPHERAL ARTERIAL CHEMORECEPTOR INPUT. Although hypoxia excites the carotid bodies in the fetus (Blanco et al. 1984), this stimulation does not increase respiratory drive, as indicated by the reduction in breathing incidence and its failure to increase the rate and amplitude of breathing activity (Sameshima and Koos 1986). Moreover, denervation of the peripheral arterial chemoreceptors does not alter the inhibitory effects of hypoxia on fetal breathing (Koos and Sameshima 1988). Thus in fetal life, afferent input from the carotid bodies during hypoxia appears to be gated out of respiratory regulation.

But hypoxia increases the rate and amplitude of breathing in fetuses in which hypoxic inhibition has been eliminated by electrolytic lesions of the pons (Gluckman and Johnston 1987), transection of the pons/caudal mesencephalon (Dawes et al. 1983; Koos 1985), or disruption of the rostral midbrain, just caudal to the red nucleus (Koos et al. 1992). Following such lesions, stimulation of breathing results from hypoxic excitation of the carotid bodies because it is eliminated by denervation of the peripheral arterial chemoreceptors (Johnston and Gluckman 1993; Koos et al. 1992). Thus abolishing hypoxic inhibition allows expression of the respiratory chemoreflex normally observed postnatally.

In contrast, IBO lesions that allowed breathing to continue during hypoxia (i.e., Pf lesions) changed neither the rate nor amplitude of breathing during acute O2 deprivation. This lack of respiratory stimulation would suggest that the thalamic locus involved in hypoxic inhibition is not part of the central mechanism that dissociates fetal breathing from hypoxic stimulation of the carotid bodies.

RELEVANCE TO POSTNATAL RESPIRATION. Although normally a respiratory stimulant, hypoxia reduces ventilation postnatally under certain conditions. The magnitude of the ventilatory response depends on the integrity of the peripheral arterial chemoreflexes, site of lesions relative to the red nucleus, and the maturity of the animal.

Decortication in cats with intact carotid bodies increases respiratory responses to hypoxia attributable to a lack of inhibitory input from the cortex and the facilitatory influence of the diencephalon (Tenney and Ou 1977). Hypoxic ventilatory responses in conscious cats were virtually the same as those in animals with midcollicular decerebration, indicating that the descending inhibitory and facilitatory influences are closely balanced, having a net minimal effect with disruptions caudal to the red nucleus.

A hypoxia-induced decline in ventilation can be most clearly observed in acutely glomectomized animals in which O2 deficiency reduces both tidal volume and frequency (Watt et al. 1942). In animals with chronically denervated carotid bodies, this hypoxia-induced reduction in ventilation is either blunted (Martin-Body et al. 1986) or eliminated altogether with a return of the hyperventilatory response (see Nolan et al. 1995). An increase in respiratory frequency largely accounts for the increased ventilation, which has been attributed to a direct effect of hypoxia on the brain as well as to enhanced input from other peripheral chemoreceptors (Martin-Body et al. 1986). The direct central effect may involve excitatory input from O2-sensing neurons in the ventrolateral medulla and the caudal hypothalamus (Nolan et al. 1995).

In rats with bilateral carotid nerve denervation, the hypoxia-induced decrease in ventilation is caused primarily by a reduction in respiratory frequency (Martin-Body 1988). But midcollicular transections that pass through the red nucleus change the hypoxic respiratory response to stimulation by increasing both frequency and tidal volume. Hypoxia also increases ventilation in glomectomized rats with brain transections that appear to be rostral to the red nucleus but caudal to the parafascicular region of the thalamus (Martin-Body 1988); this stimulation results from a progressive increase in tidal volume because frequency declined.

In ventilated cats with bilateral section of the vagi and carotid body nerves, posthypoxia changes in phrenic nerve activity include facilitation following mild hypoxia and depression after more severe O2 deprivation (Gallman andMillhorn 1988). Decerebration performed rostral to the superior colliculi and, presumably, the red nucleus, abolished facilitatory but not inhibitory responses; disruptions through or caudal to the red nucleus eliminated both facilitation and inhibition (Gallman and Millhorn 1988). Further work with electrical and glutamate stimulation localized a region inhibitory to respiration near or within the red nucleus (Gallman et al. 1991).

Although present during prolonged hypoxia in adults (Easton et al. 1986), central inhibition in intact animals is more clearly evident in the biphasic respiratory response of the newborn to acute O2 deficiency. Ventilation initially increases, but it declines toward control values after 2-3 min of O2 deprivation (Cross and Warner 1951). Denervation of the arterial chemoreceptors eliminates the initial stimulation and enhances the fall in ventilation (Schweiler 1968), indicating that the decreased ventilation results from a direct effect of hypoxia on the brain. The decline in ventilation is greater in altricial species and is modulated by several factors, including a fall in metabolic rate (Mortola et al. 1989).

This hypoxia-induced reduction in ventilation persists with decerebration performed rostral to the superior colliculus and thus the red nucleus (Hanson and Williams 1989; Schweiler 1968) but not with midcollicular transections that pass through or caudal to the red nucleus (Hanson and Williams 1989; Martin-Body and Johnston 1988). As in adult animals, electrical (Ackland et al. 1995) or glutamate stimulation of the red nucleus depresses phrenic nerve activity in young rabbits (see Waites et al. 1996). Recently, Waites and associates (1996) have reported that bilateral obliteration of only a portion of the red nucleus abolishes the second-phase decline in phrenic activity during isocapnic hypoxia, suggesting that the red nucleus is part of an inhibitory network. But the only example illustrated by these authors (their Fig. 3) reveals a poorly stained, thick (150 µM) frozen section in which there is, at best, only a partial lesion of a small dorsomedial sector of the red nucleus; clearly most of the red nucleus is intact.

Other brain stem loci apparently participate in the respiratory response of the neonate to hypoxia. For example, the second-phase ventilatory decline in anesthetized newborn lambs is eliminated by unilateral focal cooling of a dorsal pontine region proximate to the locus ceruleus (Moore et al. 1996).

In summary, hypoxia decreases ventilation in acutely glomectomized animals and as a secondary effect in normal newborns---responses that are eliminated by transections through or caudal to the red nucleus and by localized lesions of the red nucleus. Supracollicular transections with disruptions rostral to the red nucleus do not substantially alter the biphasic response to hypoxia in the newborn. Thus structures caudal to supracollicular transection but including the red nucleus are critically involved in reducing breathing activity during acute O2 deficiency in the newborn. Because of the lack of detailed histological analysis in studies with supracollicular disruptions, the actual site of the lesion, relative to the rostral midbrain and posterior thalamus, has understandably not been clearly delineated in small rat and rabbit brains. The red nucleus contains large motoneurons with projections to bulbospinal neurons near the ambiguous nuclei and cervical cord; consequently, the red nucleus may inhibit phrenic nerve activity through these rubrobulbar and rubrospinal pathways (see Waites et al. 1996). Because there is no evidence that the red nucleus is associated with rapid sensory responses, the central chemosensitive locus that detects low O2 tensions and triggers the response likely resides elsewhere.

The parafascicular nuclear complex consists of small neurons associated with sensory or integrative function (Bentivoglio et al. 1991; Rose 1942). In fetal sheep (>0.8 term), Pf lies a few millimeters dorsal and rostral to the red nucleus and immediately rostral to the periaqueductal gray. The latter receives projections from posterior hypothalamic nuclei that stimulate breathing during hypoxia (Ryan and Waldrop 1995), indicating that Pf is contiguous with structures modulating respiratory responses to hypoxia. Furthermore, this thalamic locus is presumably caudal to the supracollicular transections that left the biphasic respiratory response intact in newborn animals (Hanson and Williams 1989; Schweiler 1968). Thus there is reason to suspect that Pf plays a crucial role in hypoxic inhibition in the newborn and possibly the adult.

The neural substrate mediating hypoxic inhibition in the fetus has some similarities to that in the newborn. For example, hypoxic inhibition of breathing is eliminated by lesions of the ventrolateral pons (Gluckman and Johnston 1987), a region probably containing the rubrospinal tract (see Gallman et al. 1991; Waites et al. 1996); and the inhibitory response is abolished by transections immediately caudal to the red nucleus (Koos et al. 1992). But in contrast to postnatal studies, our results clearly indicate that the thalamic Pf sector is involved in hypoxic inhibition of fetal breathing. This discrepancy may be more apparent than real because of the limited spatial resolution of lesions in the very small brains of newborn rats and rabbits. Because the late stage fetal sheep brain is much larger than that of a newborn rodent, eliminating the red nucleus as a critical site is easily achieved, as illustrated by the largest transection lesions (937 and 982), nearly 1 cm distant from the red nucleus.

Other possibilities may also explain the apparent discrepancy between the newborn and fetus in the site of lesions that eliminate the depressing effects of hypoxia on respiratory activity. For example, decerebrations in newborns severely disrupt the neurocircuitry relating the brain stem to the thalamus and other forebrain structures, destroying pathways mediating the regulatory effects of sleep and probably other factors on respiration but do not necessarily exclude the involvement of more rostral circuits in hypoxic inhibition.

Fetal breathing is distinguished from respiration after birth by its episodic occurrence and the lack of modulation by Hering-Breuer reflexes, peripheral arterial chemoreceptors, and normal arterial blood gases (Dawes et al. 1972). Pathways mediating hypoxic inhibition of breathing may differ postnatally, and future studies should determine whether the red nucleus is involved in hypoxic inhibition of fetal breathing and whether Pf is crucial to the hypoxia-induced decline in newborn ventilation.

Mechanism and specificity of lesions

IBO destroys cells by activating N-methyl-D-aspartate (Ranson and Stec 1988) and metabotropic receptors (Houamed et al. 1991), leading to a rise in cytoplasmic Ca2+ levels (Saura et al. 1995). This cellular destruction apparently occurs without significantly affecting fibers of passage (Hastings et al. 1985; Schwarcz et al. 1979). Although IBO causes demyelination under some conditions, this is a much less extensive area than that of neuronal destruction (Stellar et al. 1991). Thus some fiber tract demyelination cannot be excluded as a factor in the loss of hypoxic inhibition. Neurons and/or fiber tracts related to hypoxic depression of breathing may have been injured by inserting the probe itself into the brain, but this seems unlikely because our control CSF injections into Pf and sPf did not abolish hypoxic inhibition.

Transient localized edema may have accompanied the acute brain injury induced by IBO, causing a functional defect larger than that outlined by the zone of gliosis or degeneration, but this seems unlikely because 1) huge diencephalic lesions, which were near but did not include Pf failed to abolish hypoxic inhibition and 2) excitotoxin-induced edema resolves within 2-4 days (Olney 1971; Olney et al. 1979), and we noted that hypoxic inhibition was abolished up to 8 days after IBO administration.

In summary, neuronal lesions of the diencephalon in fetal sheep have led to identification of a new locus immediately rostral to the brain stem, which is critically involved in hypoxic inhibition of fetal breathing. This sector encompasses the Pf complex, which is closely associated with the reticular activating system and sensorimotor function and maylink central O2-sensing cells to motoneurons that inhibitbreathing.

    ACKNOWLEDGEMENTS

  We thank S. Sampogna for the histological preparations.

  This study was supported in part by National Institute of Child Health and Human Development Grant HD-18478 and Institute of Neurological Disorders and Stroke Grant NS-5685.

    FOOTNOTES

  Address for reprint requests: B. J. Koos, Dept. of Obstetrics and Gynecology, 22-177 CHS, UCLA School of Medicine, Los Angeles, CA 90095-1740.

  Received 29 September 1997; accepted in final form 29 January 1998.

    REFERENCES
Abstract
Introduction
Methods
Results
Discussion
References
0022-3077/98 $5.00 Copyright ©1998 The American Physiological Society



This article has been cited by other articles:


Home page
 J. Appl. Physiol.Home page
P. M. Macey, M. A. Woo, K. E. Macey, T. G. Keens, M. M. Saeed, J. R. Alger, and R. M. Harper
Hypoxia reveals posterior thalamic, cerebellar, midbrain, and limbic deficits in congenital central hypoventilation syndrome
J Appl Physiol, March 1, 2005; 98(3): 958 - 969.
[Abstract] [Full Text] [PDF]

Home page
 J. Neurophysiol.Home page
R. M. Harper, P. M. Macey, M. A. Woo, K. E. Macey, T. G. Keens, D. Gozal, and J. R. Alger
Hypercapnic Exposure in Congenital Central Hypoventilation Syndrome Reveals CNS Respiratory Control Mechanisms
J Neurophysiol, March 1, 2005; 93(3): 1647 - 1658.
[Abstract] [Full Text] [PDF]

Home page
 J. Appl. Physiol.Home page
B. J. Koos, Y. Kawasaki, A. Hari, F. Bohorquez, C. Jan, J. Roostaeian, C. L. Wilson, and L. Kruger
Electrical stimulation of the posteromedial thalamus modulates breathing in unanesthetized fetal sheep
J Appl Physiol, January 1, 2004; 96(1): 115 - 123.
[Abstract] [Full Text] [PDF]