INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE FASTIGIAL NUCLEUS (FN), located in the intermediate region of the cerebellar deep nuclei (CDN), is important for facilitating stressed breathing in cats and dogs but is not critical for maintaining eupneic ventilation. Electrophysiological studies in the cat have shown that constant electrical activation of the FN either increased (major effect) or decreased respiratory frequency (f), depending on the stimulating sites (11, 18, 21). On the other hand, cerebellectomy or ablation of the FN in anesthetized and decerebrated cats (17, 20, 23, 25, 27, 28) and dogs (15) did not significantly affect eupneic ventilation. FN facilitatory effects were confirmed in the respiratory response to hypoxia in which a similar attenuation of the ventilatory responses to progressive or transient hypoxia (5 breaths of pure nitrogen and sodium cyanide iv) was produced after bilateral lesions of the cat FN and cerebellectomy (26). In fact, overall cerebellar involvement in ventilatory responses to both hypoxia and hypercapnia has been reported. Mansfeld and Tyukody (13) first reported that, in the decerebrate dog, cerebellectomy depressed the ventilatory response to inhalation of 10% CO₂ as well as severe hypoxia. Consistent with this finding are more recent results obtained in anesthetized cats (25, 26) and decerebrate dogs (15) in which cerebellectomy attenuated the ventilatory responses to progressive hypercapnia or hypoxia via reducing f and/or tidal volume. The question is whether the same FN region is responsible for both ventilatory responses and, if so, whether they share a common modulating pattern.

The possible involvement of other CDN, i.e., the interposed nucleus (IN) and lateral nucleus (LCN; ventral part of which is the infracerebellar nucleus), in respiratory modulation has been studied in different experimental preparations. Respiratory-modulated neurons have been recorded in the IN of the alert cat (5) as well as in the FN (5, 12, 22). Electrical stimulation of the infracerebellar nucleus increased expiratory activities; conversely, lesions of this region dramatically inhibited expiratory activity of spinal nerves in the decerebrate preparation (9). To date, the influences of the various CDN on eupneic and stressed respiration have not been systematically compared.

Recently, immunocytochemical studies have shown that exposure to hypercapnia (29) or hypoxia (24) markedly increased fos-like immunoreactivities within the rat FN. More interestingly, a similar increase of fos expression in the FN was observed in the anesthetized rat by focal application of acidic mock cerebrospinal fluid (CSF) at the rostral ventrolateral medulla where CO₂ chemosensitive neurons are located (10). These findings strongly suggest that stimulation of central and peripheral chemoreceptors activates rat FN neurons. It is unclear, however, whether the rat CDN, particularly the FN, functionally participate in the modulation of eupneic and chemically stressed respiration.

The purpose of the present study was to answer the following questions: 1) does activation of rat FN affect respiratory motor output? 2) is FN-mediated respiratory response unique compared with other CDN and independent of the associated pressor response? and, if so, 3) are FN neurons rather than fibers of passage responsible for the response? and 4) is the same population of neurons involved in respiratory responses mediated by activation of CO₂ and O₂ receptors through a common modulating pattern? Our results show that the role of rat FN in respiratory modulation is very similar to that in the cat, i.e., primarily facilitating respiration via increasing f. FN contribution to ventilation is not secondary to the cardiovascular alteration because FN-mediated respiratory response persists after preventing or blocking the associated pressor response by using transient electrical stimulation and intravenous injection of -adrenergic blocker, respectively. The rat FN plays a unique role among CDN because 1) electrical stimulation of the IN or LCN fails to elicit any constant respiratory responses and 2) a significant attenuation of respiratory responses to hypercapnia and cyanide was observed after chemical destruction of the FN neurons but not the other CDN neurons. Compared with the response to cyanide, FN involvement in hypercapnic respiratory response seems to be much more significant. These data suggest that, among CDN, FN neurons, instead of fibers of passage, contribute to both respiratory reflexes, especially to the hypercapnic respiratory response.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

General method. The experiments were performed in 12 Sprague-Dawley rats (300-400 g) initially anesthetized with chloralose (100 mg/kg) and urethan (500 mg/kg ip). The left femoral vein and artery were cannulated, the former for anesthetic or drug administration and the latter for monitoring arterial blood pressure (ABP) and sampling arterial blood gases (model 1306 pH/blood-gas analyzer, Instrumentation Laboratory). The supplemental anesthesia (chloralose and urethan) was administered intravenously to suppress corneal and withdrawal reflexes. The trachea below the larynx was tracheotomized by blunt dissection and cannulated with a tracheal cannula connected to a one-way breathing valve. A three-way switch was attached to the inspiratory inlet of the one-way breathing valve and used to manipulate the inhaled gas mixture, i.e., either for control (30% O₂ + 70% N₂) or chemical stimulation [detailed in Chemical exposures before and after destruction of neurons within the FN (the IN and LCN)]. The core temperature was monitored with a rectal probe and maintained at 37-38°C by a heating pad and radiant heat. The animal was paralyzed by an intravenous infusion of gallamine triethiodide (0.2-0.3 mg/kg for induction, followed by a continuous infusion of 0.2 mg · kg¹ · h¹). The rat was artificially ventilated through the inspiratory and expiratory side of the one-way breathing valve by a conventional volume ventilator (model 683 rodent ventilator, Harvard). In four rats, both cervical vagi were isolated and looped with ligature for later transection. After paralysis, supplemental anesthesia was given when abnormal irregularities were observed in ABP (heart rate) and/or f and respiratory pattern under control condition.

Cerebellar craniotomy. Animals were placed in a rigid metal frame with the head fixed in a stereotaxic apparatus (Kopf). A hole (10- to 12-mm diameter) was drilled at the midline (12.5 mm posterior to the bregma) for stereotaxically inserting electrode and pipette into the CDN according to the rat brain atlas (14). Bleeding was controlled with bone wax, absorbable hemostat (Surgicel and Gelfoam), and the use of a bipolar coagulator (model 440S, Radionics). The dura was removed, and the underlying tissue was covered by cotton saturated with mineral oil.

Monitoring and recording. The right phrenic (C₅) nerve was freed in the neck, desheathed, and cut, and phrenic efferent nerve activity (PN) was recorded with bipolar silver electrodes in pools of mineral oil. Nerve signals were amplified (model P15 amplifier, Grass Instrument) and filtered (band pass 20-3,000 Hz). PN were integrated (PN) with a leaky resistance-capacitance circuit (0.1-s time constant) and monitored on a storage oscilloscope (model 5103n, Tektronix). Arterial blood was sampled after general surgery, and arterial pH was maintained within the range of 7.3-7.4. Intravenous infusion of 0.5 M sodium bicarbonate was made as needed to correct base deficiencies. End-tidal PO₂ and PCO₂ (PET_O2 and PET_CO2, respectively) were monitored via an infrared O₂-CO₂ analyzer (model 78356A, Hewlett Packard). Values of PET_O2 and PET_CO2 (breath by breath) continuously showing on the monitor were maintained at >100 and ~30 Torr via adjustment of the ventilatory volume and rate and/or adding O₂ or CO₂ into the input line except during chemical challenge. Once the parameters of ventilation were set, they were kept constant throughout the experiment. ABP, PN, PN, tracheal pressure, and PET_CO2 were monitored and recorded throughout the experiment.

Electrical stimulation of CDN. After baseline cardiorespiratory variables became stable, studies were conducted in two series of experiments. In series I (n = 5), stereotaxic coordinates were used to position a stainless steel, concentric bipolar electrode (model NE-100, Rhodes Medical Instruments) into the FN. The stimulating electrode was placed in the site where reproducible PN responses were clearly detectable during electrical stimulation. The stimulating parameters were delivered from a digital stimulator (model S8800, Grass) at the beginning of either the inspiratory or expiratory phase. The stimulating intensity was fixed (400-ms trains of 0.2-ms pulses at either 100 or 200 µA) throughout the experiment while the stimulating frequency varied (10, 20, 50, 75, 100, 150, and 200 Hz) randomly. A simulating threshold was defined as the lowest stimulating frequency at which a detectable change of PN activity was elicited. Depending on the duration, the type of stimulation was divided into transient (<2 s) or constant (up to 10 s). The placement sites of the stimulating electrodes were mapped on a grid. The same stimuli were applied after 1) administration of phenoxybenzamine (1 mg iv, -adrenergic blocker; n = 2) and 2) repositioning the electrode into the IN and the LCN or the other sites of the FN (ipsilateral or contralateral), randomly.

Chemical exposures before and after destruction of neurons within the FN (the IN and LCN). Because our preliminary data showed that reproducible respiratory responses were evoked by stimulation of the FN but not the IN and LCN, the contributions of FN neurons to respiratory chemoreflexes were tested in series II (n = 7). A dual electrode, composed of a tungsten electrode and a mounted micropipette (~15 µm ID with distance from the tungsten electrode <20 µm; detailed in Ref. 21), was placed into the FN. The former was used as a stimulating electrode and the latter for microinjection of kainic acid. When a respiratory response to electrical stimulation of the FN was elicited, the electrode site was fixed. The animal was subsequently exposed to different chemical challenges: 1) inhalation of hypercapnia (7% CO₂ + 30% O₂ + 63% N₂) until the value of PET_CO2 equaled ~55 Torr and 2) bolus injection of sodium cyanide (20-50 µg iv). The same stimuli (electrical and chemical) were repeated 1 h after bilateral microinjection of kainic acid into the FN via the mounted micropipette (1 mM in a solution of 2% Chicago sky blue in 150 mM saline, 100-150 nl for ~15 s). The volume of drug ejected was verified by viewing the meniscus through a microscope (model E3069, Melles Griot) with a calibrated reticule. To test the specificity of FN involvement in the respiratory responses to chemical challenges, the IN and LCN were lesioned by kainic acid after FN lesions in two of seven rats, and the same stimuli were repeated.

Histological examination. The rat was killed by additional anesthetic (iv) after completion of the protocols, and the brain stem and cerebellum were removed and placed in 10% Formalin. After at least 3 days of immersion fixation, the brain stem was frozen, and 50-µm sections were cut and mounted. Each slice was stained and the location of the lesions was drawn with camera lucida.

Data acquisition and analysis. Cardiorespiratory variables [i.e., ABP, PET_CO2, peak PN (PN_peak), f, minute phrenic activity (MPN; PN_peak × f), and inspiratory and expiratory duration (TI and TE, respectively)] were measured and collected. The control (baseline) values, expressed as absolute values, were obtained by averaging of the relevant variables within five breaths just before application of electrical stimuli and 1 min immediately before exposure to chemical challenges or injection of kainic acid, respectively. The responses were determined by measuring respiratory variables within a period of 1) the first breath and 3 breaths immediately after transient and constant electrical stimulation, respectively; 2) 3 breaths and 10 breaths displaying greatest responses to cyanide injection and hypercapnia, respectively; and 3) 5 breaths (immediate response) and 1 h after kainic acid injection (stabilized response). The responses were presented as percent change from control. All data were presented as means ± SE. Paired t-test was used for comparing the differences of cardiorespiratory activities between the control and response (before and after kainic acid injection). One-way ANOVA and the Student-Newman-Keuls post hoc test were used to identify significance of the differences of cardiorespiratory responses obtained under the following conditions: 1) electrical stimulation of the FN, IN, and LCN; 2) electrical stimulation of the FN with and without pretreatment of kainic acid; and 3) hypoxic and hypercapnic exposure before and after chemical lesions of the bilateral FN. A P value <0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The effect of electrical stimulation of different cerebellar nucleus. Electrical stimuli were applied in the right FN (12 trials) and subsequently in the left FN (2 trials). The stimulating thresholds to trigger the initial detectable respiratory responses were varied from 50 to 100 Hz, i.e., 50 Hz (5 trials), 75 Hz (3 trials), and 100 Hz (6 trials). Typical experimental recordings are shown in Fig. 1. Stimulation at 100 Hz dramatically elevated f associated with a pressor response, whereas 50 Hz failed to affect respiration (Fig. 1A). In contrast, 50 Hz applied in another rat led to a decrease in f, and even an apnea accompanied with a hypertension if a greater stimulation (100 Hz) was applied (Fig. 1B). The group data showing the cardiorespiratory responses to threshold stimulation of FN stimulation are listed in Table 1 (see values obtained before lesions). FN stimulation primarily led to an elevation of MPN (~30%) via increasing f rarely associated with an effect on PN_peak (3 of 14 trials). The alteration of f mainly resulted from a decrease in TE (10 of 14; ~71%; P < 0.05). Regardless of the differences in response patterns, constant electrical stimulation of the FN produced an associated pressor response (~20%; P < 0.05) that occurred at a delay of 1.5 ± 0.3 s from the onset of the evoked PN responses. These data suggest that the predominant effect of electrical activation of the FN is to modulate respiratory timing, particularly the TE. The stimulating sites within the FN and the corresponding cardiorespiratory responses are illustrated in Fig. 2. Histological data indicate that no discrete FN area was found that corresponded to a given response pattern. In addition, PN responded not only to contralateral but also to ipsilateral stimulation of the FN, as denoted on Fig. 2.

View larger version (49K): [in this window] [in a new window]   Fig. 1.   Typical experimental recordings to illustrate the effects of electrical stimulation of the fastigial nucleus (FN) on phrenic efferent nerve activity (PN). In each panel, the traces from top to bottom are arterial blood pressure (ABP), end-tidal PCO2 (PETCO2), PN, integrated PN (PN), stimulation markers, and tracheal pressure (Ptr). When the stimulus reached the threshold, respiratory frequency (f) was either increased (A) or decreased (B), as shown for 2 different rats. FN stimulation-induced PN responses were usually accompanied by hypertension. In a few cases, high stimulating frequency produced an apnea (B; 100 Hz).

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Schematic presentation of the cross section of brain stem and cerebellum depicting electrode placement sites. Open and filled symbols represent the sites where delivering stimuli prolongs and shortens expiratory time (TE), respectively; circles, TE alteration alone; triangles, TE alteration coupled with modulation of peak PN (PNpeak); *, absence of detectable respiratory responses. IN, interposed nucleus; LCN, lateral cerebellar nucleus; NA, ambiguous nucleus; 7, facial nucleus; DPGi, dorsal cochlear nucleus; PY, pyramidal tract; V4, fourth ventricle. Note: in 2 of 12 rats, contralateral and ipsilateral FN was stimulated, separately.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Experimental recording to compare the effects of electrical stimulation of the FN (A), IN (B), and LCN (C) on phrenic nerve activity. In each panel, the traces from top to bottom are ABP, PN, stimulation marker, and Ptr. Electrical stimulation (100 Hz) of the FN markedly altered PN activity associated with a hypertension. However, no apparent changes in the PN and ABP were observed when the stimulation (100-150 Hz) was applied in the IN or LCN

Hypertension and the FN-mediated respiratory responses. Because respiratory responses elicited by electrical stimulation of the FN were usually accompanied by a pressor response, experiments were performed to determine whether the FN-mediated respiratory responses were secondary to the pressor response. First, intravenous injection of phenoxybenzamine was used to block the associated pressor response. Before injection, typical FN-mediated cardiorespiratory responses were observed (Fig. 4A). After injection of phenoxybenzamine, a qualitative similar respiratory response to FN stimulation persisted without a pressor response (Fig. 4B). Second, lower stimulating frequency and transient stimulation of the FN were used to prevent the associated cardiorespiratory response. We found that transient stimulation (<2 s; Fig. 4, C and D) or utilization of a lower stimulating frequency (50 Hz; Fig. 1B) produced a dramatic respiratory response, whereas pressor response was significantly reduced or absent. Totally transient stimulation was applied in seven trials that resulted in a remarkable reduction of either TE (53 ± 9.3%; n = 5; P < 0.05) or TI (32.5%; n = 2) without significant change in ABP (115 ± 8.6 mmHg for control and 116.6 ± 10.4 mmHg for stimulation). The nature of the respiratory responses to FN stimulation in four vagotomized rats was similar to vagal intact rats. In other words, the response is independent of the integrity of the vagi. Typical examples are presented in Fig. 1B (vagal intact) and Figs. 1A and 4A (vagotomized).

View larger version (28K): [in this window] [in a new window]   Fig. 4.   Effect of the associated changes in ABP on FN-mediated PN responses. In each panel, the traces from top to bottom are ABP, PETCO2, PN, and stimulation marker (100 Hz). PN responses to FN stimulation (A) persisted in a rat when the associated pressor response was blocked by intravenous injection of phenoxybenzamine (1 mg/kg; B). PN response to FN stimulation also persisted when the associated pressor response was significantly diminished (C) or absent (D) by using transient stimulation.

Involvement of FN neurons in eupneic breathing and FN-mediated respiratory responses. An experimental record showing the effect of kainic acid injection into the FN on respiration is displayed in Fig. 5. Kainic acid injection (100 nl) produced an initial increase in f for several breaths, followed by a decrease in f associated with augmentation of PN_peak. Surprisingly, no significant changes in ABP were observed after kainic acid injection. Cardiorespiratory variables before and after microinjection of kainic acid into the FN were compared in Table 2. Kainic acid microinjection caused an immediate excitatory respiratory response characterized by a significant increase of MPN via shortening TE (increase f) with little effect on ABP. Interestingly, no significant differences of cardiorespiratory variables were found 1 h after bilateral lesions of the FN. The immediate elevation of f probably results from the initial excitatory effect of kainic acid on FN neurons. After bilateral FN lesions induced by kainic acid injections, the PN response to FN stimulation previously observed in the FN intact rat (Fig. 5A) disappeared (Fig. 5B), implicating FN neurons in the evoked respiratory responses.

View larger version (37K): [in this window] [in a new window]   Fig. 5.   Comparison of PN responses to electrical stimulation of the FN in a rat before and after microinjection of kainic acid (KA; 100 nl) into the FN. A: immediate cardiorespiratory response to microinjection of KA into the FN. Traces are ABP, PN, and a marker for microinjection duration. Microinjection of KA into the FN produced an initial increase in f for several breaths followed by a decrease in f associated with augmentation of PNpeak. Thirty minutes after bilateral lesions, f remained slower, primarily because of a prolongation of TE. B: respiratory responses to electrical stimulation before and after KA lesions were compared. Electrical stimulation of the FN at a frequency greater than or equal to the threshold (100-150 Hz) increased f (top of B), whereas these responses disappeared 1 h after injection of KA (bottom of B).

FN neurons and the respiratory responses to chemical challenges. The role FN neurons play in the respiratory response to hypercapnia and cyanide was examined. As depicted in Fig. 6, FN lesions significantly attenuated the respiratory response (MPN) to hypercapnia (PET_CO2 increased from 33.0 ± 3.4 to 54.0 ± 3.3 Torr; P < 0.05; Fig. 6A). The attenuation of respiratory response to hypercapnia primarily results from a marked decrease of TE (Fig. 6B) with a tendency to reduce respiratory amplitude (Fig. 6A). The dominant component contributing to an increase in f during hypercapnia in anesthetized rats is the reduction of TE as is evident by the comparison of inspiratory and expiratory timing (Fig. 6B). FN lesions significantly diminished the hypercapnia-induced expiratory response and thereby led to an attenuation of minute phrenic activity. With respect to the response to cyanide (Fig. 7), FN lesions attenuated the general respiratory response (Fig. 7A) via inhibition of the expiratory response (Fig. 7B). When Fig. 6 is compared with Fig. 7, it is clear that FN contribution to hypercapnic respiratory response was greater than to cyanide-elicited response. FN lesion markedly diminished f and minute ventilation in response to hypercapnia, whereas it did not significantly affect f in response to cyanide. It should be noted that FN lesions did not markedly change 1) the baseline ABP (111.0 ± 8.1 mmHg, intact; 113 ± 8.8 mmHg, FN lesioned; P > 0.05) and 2) the ABP response to hypercapnia (101.4 ± 6.7 mmHg, intact; 106.8 ± 8.0 mmHg, FN lesioned; P > 0.05) and to cyanide injection (112.5 ± 10.9 mmHg, intact; 112.5 ± 10.1 mmHg, FN lesioned; P > 0.05). In two rats, subsequent bilateral lesions of the IN and LCN did not alter respiratory motor output during eupneic and chemically stressed breathing.

View larger version (24K): [in this window] [in a new window]   Fig. 6.   Involvement of FN neurons in respiratory response to hypercapnia (PETCO2 change from 33.0 ± 3.4 Torr to 54.0 ± 3.3 Torr). Compared with control (0%), a significant increase in PNpeak, f, and MPN (A) and a decrease in TE (B) were observed during hypercapnia in both FN intact and lesioned preparations. Responses of f, MPN, and TE were profoundly attenuated after FN lesions. TI, inspiratory duration. Values are means ± SE for 7 rats. * P < 0.05 compared with control.  P < 0.05 between the responses before and after FN lesions.

View larger version (24K): [in this window] [in a new window]   Fig. 7.   Influences of FN lesions on respiratory response to intravenous injection of cyanide (CYN). Compared with control (0%), cyanide injection elevated PNpeak, f, and MPN (A) and shortened TE (B) in both FN-intact and lesioned preparations. Responses of MPN and TE were significantly depressed after FN lesions. Values are means ± SE for 5 rats. * P < 0.05 compared with control.  P < 0.05 between the responses before and after FN lesions.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The FN is unique among CDN in its ability to facilitate respiration. Our results demonstrated that electrical stimulation of the rat FN, but not of the IN and LCN, elicited consistent respiratory responses. This lack of IN and LCN involvement is supported by our preliminary observation that lesions of both the IN and LCN failed to affect both eupneic and chemically stressed breathing. This was the first investigation that explored the potential roles of individual cerebellar nuclei on respiration in the same experimental preparation. It suggests a unique involvement of the rat FN in the ability to modulate the respiratory motor output. Previous studies conducted on the cat showed that selective ablation of the IN significantly diminished the fictive cough elicited by mechanical probing of the intrathoracic trachea (23). In agreement, respiratory-modulated neurons, predominantly expiratory, have been found within the IN of alert cats (5), and electrical stimulation of this region in opossums (3) and cats (9) altered expiratory activity. In terms of the LCN, selective activation of the infracerebellar nucleus facilitates expiratory activity in the decerebrate preparation (9). We cannot rule out that the lack of IN and LCN effects noted in the rat with electrical stimulation or chemical lesions might be explicable technically on the failure to activate or inactivate the whole nucleus. Species differences must be considered because the relative size of the FN compared with the other CDN in the rat is much greater than that in the cat (14, 16). Collectively, our data clearly suggest that the FN appears to be a more significant contributor in respiratory modulation than other CDN.

The FN facilitates the respiratory response to cyanide and hypercapnia. Although previous studies examined the overall cerebellar contribution to hypoxic and hypercapnic respiratory responses, the FN influence on both reflexes has not been systematically investigated. A fundamental finding in the present study is that the rat FN is important in augmentation of respiration during hypoxia and hypercapnia. It should be emphasized that FN lesions failed to alter eupneic breathing but they dramatically depressed respiratory responses to chemical challenges. Therefore, it would appear that the facilitatory influence of the FN on respiration is triggered by severe chemical stresses. The FN contribution to the hypoxic respiratory reflex is most likely mediated via activation of peripheral chemoreceptors because thermal lesions of the bilateral FN significantly decreased the respiratory response to inhalation of five breaths of pure nitrogen and intravenous administration of cyanide (26). Furthermore, a long-term facilitation of respiratory motor drive after repeated electrical stimulation of the carotid sinus nerve in the rat disappeared after cerebellectomy (6). These findings are consistent with our observation that chemical lesions of rat FN attenuated the respiratory response to intravenous injection of cyanide. There are several lines of evidence to demonstrate the involvement of the FN in respiratory response to hypercapnia. First, cerebellectomy depressed the ventilatory response to hypercapnia in decerebrate dogs (13, 15) and anesthetized cats (25). Second, with use of extracellular recording techniques, respiratory-modulated neurons within the FN have been recorded in alert (5) and anesthetized cats (12, 22) and, interestingly, approximately one-third of respiratory-modulated neurons recorded were selectively responsive to hypercapnic challenges. Third, several authors, using c-fos immunocytochemistry in rats (29) or a magnetic resonance imaging approach in humans (4), have reported that neurons adjacent to the FN were activated by hypercapnia although these studies could not distinguish whether the labeled neurons were directly or synaptically activated by CO₂ or H⁺. Recent study shows that focal application of acidic mock cerebrospinal fluid at the rostral ventrolateral medulla, containing CO₂ chemosensitive neurons, markedly increased Fos expression in the CDN (10). This finding raises the possibility that some neurons within the FN are activated by inputs from central chemoreceptors. On the other hand, investigators using the perforated patch approach have found that some cultured cerebellar cells from postnatal rats exhibit action potentials either spontaneously or that are triggered by current injection (Dr. W. Wang, Yale University, New Haven, CT; personal communication). Neuronal firing rates were increased when the neurons were exposed to acidified mock CSF (pH 7.2) for 3-5 min, but they decreased with alkalized mock CSF (pH 7.6). These observations suggest the presence of cerebellar chemosensitive neurons although it is unknown whether these neurons are FN neurons. Collectively, the information presented above supports the postulation that the mechanisms underlying the cerebellar involvement in respiratory chemoreflexes are more complicated than that previously believed. It is possible that the cerebellum contains not only respiratory-modulated neurons that receive inputs emanating from peripheral and central chemoreceptors within the respiratory central network but also CO₂-sensitive neurons directly responsible for respiratory augmentation during hypercapnia. An interesting finding in the present study is that the FN effect on the hypercapnic respiratory response seems to be much more significant than that on the hypoxic response. FN lesion clearly reduced f and minute ventilation in response to hypercapnia, whereas it did not affect f in response to cyanide (see Figs. 6 and 7).

The predominant FN influence on respiration is the control of expiratory duration. In the present study, electrical stimulation of the rat FN enhances respiratory motor output primarily via shortening expiratory duration (elevating f). This finding is confirmed by the observation that 1) kainic acid injection into the FN produces an immediate increase in f through reduction of TE (see Fig. 5A) and 2) the response of respiratory timing (increase in f via shortening TE) to chemical challenges are attenuated (Figs. 6 and 7) after FN lesions. A cerebellar contribution to respiratory timing can also be drawn from other studies. Electrical stimulation of the FN has been shown to increase breathing rate in decerebrate cats by terminating or shortening either the expiratory and/or inspiratory duration (11, 21). Conversely, lesions within this region yielded a decrease in breathing rate (18, 19), and prolonged the abdominal electromyogram burst duration elicited by expiratory tracheal occlusion (20). Furthermore, extracellular recording demonstrated that the predominant population of respiratory-modulated neurons within the cat FN was expiratory phase related (22). Activation of these neurons prematurely terminates expiratory neuronal activity (shortening TE) recorded in the medullary ventral and dorsal respiratory groups (21). In the cat, cerebellectomy or FN lesions has been reported to lead to a reduction of f response to hypercapnia and cyanide although no further differentiation was made to determine whether inspiratory and/or expiratory duration was affected (15, 25, 26). In humans, constant electrical stimulation within the vicinity of the FN during surgical proceeding results in respiratory tachypnea (8). In some human patients with either cerebellar lesion or neurodegeneration, their f was slower or irregular, particularly during movement (2). Together, these findings show a substantial FN involvement in the modulation of respiratory timing. One may ask whether the FN is functionally linked with other pontomedullary structures responsible for regulation of respiratory timing. There has been evidence to indicate that FN neurons modulated both intrinsic and bulbospinal respiratory neuronal activity via paucisynaptic connections through which phrenic motor output was affected synchronously (21). Recent studies from our laboratory further suggest that neither the neurons of the pontine respiratory groups nor the Bötzinger complex is required for that FN-mediated respiratory response (30). The point(s) of entry of the FN influence into the respiratory network has not been established at this time.

FN-mediated respiratory response is independent of associated pressor response and vagal inputs. Constant electrical stimulation of FN usually produces a pressor response in concert with modulation of respiration (1, 11, 21). It is, therefore, questionable whether the FN stimulation-induced respiratory modulations are secondary to the pressor response. This potential confounding is ruled out in our experiment by several lines of evidence. First, the respiratory response to FN stimulation persisted after blocking (phenoxybenzamine iv) or preventing (transient stimulation) of FN stimulation-induced pressor response. Second, activation of the FN produces varied respiratory response patterns, but the pressor response occurs constantly (Fig. 1, Table 1). Third, the pressor response appeared ~2 s after the onset of the respiratory response to electrical stimulation of the FN (Figs. 1, 3, and 4). Fourth, the initial excitatory effect of microinjection of kainic acid into the FN increased f without alteration in ABP (Fig. 5A). The pressor response induced by electrical rather than chemical stimulation of the FN may be due to its activating the fibers of passage within the FN or a greater area. Functional linkage between vagal afferents and the cerebellum has been pointed out previously. Studies in anesthetized cats showed that electrical stimulation of the cervical vagus afferents produced evoked potentials in the CDN (7) and the dorsal vagal nuclei were labeled after injection of horseradish peroxidase into the FN (31). In addition, an interaction between the cerebellum and vagal inputs on expiratory muscle responses to tracheal occlusion has been reported (20, 27). However, we found that bilateral vagotomy did not affect the respiratory response to electrical stimulation of the FN, demonstrating that FN-mediate respiratory response is not dependent on vagal inputs.

Conclusion. Our results suggest that, among rat CDN, the FN neurons are uniquely involved in respiratory control primarily by modulation of expiratory timing during stressed breathing. Moreover, FN-mediated respiratory responses are independent of associated pressor response and vagal input.