| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1Zablocki Veterans Affairs Medical Center; and 2Department of Anesthesiology, Medical College of Wisconsin, Milwaukee, Wisconsin
Submitted 15 December 2003; accepted in final form 8 December 2004
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
; Schwarzacher et al. 1995; Smith et al. 1991). In vitro studies show that this site contains a group of inspiratory (I) interneurons with intrinsic bursting properties (Butera et al. 1999a; Johnson et al. 1994; Lieske et al. 2000). Excitatory synaptic connections among these neurons suggest that these neurons constitute the core of a bilaterally distributed pacemaker network that is the origin of respiratory rhythm generation demonstrated in vitro (Butera et al. 1999b; Koshiya and Smith 1999). In in vivo models, microinjections of D,L-homocysteic acid (DLH) into pre-BC produce a marked tachypnea characterized by a decrease in both inspiratory and expiratory durations (TI and TE) (Chitravanshi and Sapru 1999; Monnier et al. 2003; Solomon et al. 1999; Wang et al. 2002). Selective unilateral ablation of neurokinin-1 receptor-expressing (NK1R) neurons in the pre-BC abolishes the tachypneic response to DLH on the ablated side. In awake adult rats, selective and nearly complete bilateral destruction of pre-BC NK1R neurons results in an ataxic breathing pattern with hypercapnic and hypoxic blood gases and inadequate or paradoxical responses to challenges such as hyperoxia, hypoxia, and anesthesia (Gray et al. 2001). Accordingly, it appears that the pre-BC plays a key role in respiratory rhythm generation.
Several models have been proposed to explain the mechanisms underlying rhythm generation. These include neural networks, (Balis et al. 1994; Duffin et al. 1995; Ogilvie et al. 1992; Richter et al. 1986; Rybak et al. 1997; Zuperku et al. 1982b), interconnected pacemaker neurons (Butera et al. 1999b), a hybrid-pacemaker-network system (Feldman et al. 1990; Smith et al. 2000), and recently a maturational network-burster model (Richter and Spyer 2001). However, new findings appear to question some of the basic tenets of these models (Feldman et al. 2003), and the issue of the neural substrates that underlie rhythmogenesis remains unresolved.
In the mature mammal, respiratory-related neurons with several types of discharge patterns can be recorded in the pre-BC in vivo. These include neurons that discharge during the I phase, the E phase, and across the two phases (Connelly et al. 1992; McCrimmon et al. 2001; Schwarzacher et al. 1995; Sun et al. 1998). The neuron types can be classified further according to the time course of their patterns, for example augmenting, decrementing, and E-I and I-E phase spanning. The various models have attempted to assign a role in rhythm generation to each neuron type, but these roles have yet to be verified.
The response characteristics of pre-BC neurons to systematic changes in phase timing produced by pulmonary afferent inputs have received limited attention (Hayashi et al. 1996), although reflexly induced changes in phase timing and neuronal responses to peripheral chemoreceptor (Morris et al. 1996), baroreceptor (Lindsey et al. 1998), and airway defense receptor (Shannon et al. 2000) stimulation have been studied. The discharge patterns of those neurons that play a key role in generation and control of phase timing should exhibit some aspect of their pattern that is consistently related to TI or TE. This relationship can be revealed when these phase durations are systematically altered. Only neurons the patterns of which tightly correlate with changes in phase timing are likely to play a primary role in rhythmogenesis. The objectives of the present studies were to characterize the types of neurons, their relative frequency of occurrence, their axon projections, and their responses to changes in phase timing produced by pulmonary afferent inputs in an in vivo canine model.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Surgical procedure
Dogs were intubated with a cuffed endotracheal tube and mechanically ventilated with an air-O2-isoflurane mixture. The surgical procedures, monitoring, and maintenance of body homeostasis have been previously described in detail elsewhere (Dogas et al. 1998). Briefly, after cannulating the femoral artery (for blood pressure recording and blood-gas sampling) and vein (for continuous infusion of maintenance fluids and drugs), a bilateral pneumothorax was performed to reduce motion artifacts. The animals were placed in a Kopf (model 1530) stereotaxic apparatus and then decerebrated (Tonkovic-Capin et al. 1998). This procedure leads to an anatomically well-defined, midcollicular decerebration. After completion of the decerebration, isoflurane was discontinued, and the dogs were ventilated with an air-O2 mixture and maintained in hyperoxic normocapnia (PO2>400 mmHg, PCO2 35–45 mmHg). Body temperature was maintained at 37 ± 0.5°C with a servo-controlled heating pad. With the dog's head ventrally flexed 38° from the horizontal plane, the medulla was tilted an average of 17° from the horizontal plane with the rostral aspect lower than the caudal aspect. The dorsal surface of the medulla oblongata was exposed by an occipital craniotomy. Prior to neuronal recording, the neuromuscular blocker pancuronium (0.1 mg/kg, followed by 0.1 mg · kg–1 · h–1) was given to reduce motion artifacts.
Phrenic nerve activity was recorded from the central end of the desheathed right C5 rootlet. The phrenic neurogram (PNG) was obtained from the moving-time average (100 ms) of the amplified phrenic nerve activity and was used to produce timing pulses corresponding to the beginning and end of the inspiratory phase. The spike-triggered averaging technique was used to test for the presence of laryngeal motoneurons within the region of the pre-BC. The superior and inferior laryngeal nerves were carefully isolated, and the central ends were desheathed and prepared for recording of efferent activity. The laryngeal nerves were placed on bipolar platinum hook electrodes and submersed in warm mineral oil. Spike-triggered averages of amplified efferent nerve activity (0.2- to 3-kHz, bandwidth) were obtained using a Nicolet NIC-370 digital signal averager triggered by the unit activity of single neurons (e.g., Fig. 1A).
|
Multi-barrel micropipettes, (10–30 µm composite tip diameter), consisting of one recording barrel containing a carbon filament and drug barrels, were used for extracellular neuronal recordings and pressure ejections. The ejected volumes were measured via changes in meniscus height with a x100 microscope equipped with a reticule. The location of each recorded neuron was registered in coordinates relative to the obex (rostral), midline (lateral), and dorsal surface at point of micropipette entry (depth). Micropipette penetrations were started at a predetermined position based on maximum tachypneic responses to DLH injections that had been obtained in dogs of similar size in a preliminary study (e.g., Fig. 2). Because of potential toxic effects of DLH, due to relative large injection volumes (
20 nl) and concentration (20 mM) required to produce such responses, confirmation of the tachypneic response was only done after all neuronal recordings were completed. With respect to the initial position, penetration locations for the current study of single neurons were incremented in a grid-like manner using 250-µm steps in the rostral-caudal and medial-lateral directions ±750 µm. After all neuronal recordings were completed, a DLH injection was made near the center of the region from which the recordings were made. This was then followed by picoejection of Pontamine Sky Blue dye (1% solution, 50–100 nl) to mark the location. After the animals were killed, each medulla was removed and placed in 4% paraformaldehyde for 24 h before being frozen (–70°C). Frozen transverse sections (30 µm) were cut, stained with neutral red, and cover slipped. Sections were examined light microscopically to identify the dye marker, and subsequent photomicrographs were made to establish the coordinates relative to obex (e.g., Fig. 3).
|
|
Statistical analyses
Neuron location coordinate values were compared with each other with a one-way ANOVA procedure (StatView, SAS Institute, Cary, NC). The 
2 test was used to test whether the observed frequency of a given neuron subtype was different from expected, based on the hypothesis that all neuron subtypes are equally likely to be observed. The 2 test was also used to test whether there was a preferred response direction, assuming that all three outcomes, the same, opposite, or no response to vagal stimulation, were equally probable. Differences were considered significant for P < 0.05. The Fisher's protected least significant difference (PLSD) was used for multiple comparisons. Values are expressed as means ± SE unless indicated otherwise.
Protocol 1
The purpose of protocol 1 was to determine the types and frequency of occurrence of neurons within the pre-BC, based on their spontaneous discharge patterns, and the number of neurons that are cranial motor- or bulbospinal neurons. This study was carried out in a group of 15 dogs that were vagotomized to eliminate changes in discharge patterns and phase timing due to slowly adapting pulmonary stretch receptor (PSR) inputs. Both spike-triggered averaging and spinal antidromic activation procedures were performed for each recorded neuron. Those neurons, which gave negative spike-triggered averaging and antidromic activation results, were assumed to be propriobulbar. The coordinates for the location of the pre-BC region were verified at the end of the recording session by the presence of the tachypneic response produced by DLH picoejection.
Protocol 2
The purpose of protocol 2 was to determine the response characteristics of pre-BC neurons to changes in phase timing produced by PSRs. A second group of 31 dogs were used for this purpose. In some preparations, a programmable ventilator synchronized by the PNG was used to produce various inflation patterns during either the I or E phase of test cycles. These were separated by several control cycles to prevent swings in PaCO2 and blood pressure. In most preparations, electrically induced PSR activity from the central ends of the sectioned vagus nerves was used to produce phase-timing changes. This procedure allows constant ventilation and avoids oscillations in PaCO2. It also minimizes changes in blood pressure. A PNG-synchronized programmable stimulation system (Hopp et al. 1983) was used to deliver pulse train stimuli to the ipsilateral, contralateral and both vagus nerves, throughout the I or E phase. Stimulus strength (20-µs pulse duration, 100–150 µA) was adjusted to mimic the Hering-Breuer reflexes, that is, I-phase stimuli shorten the I phase without altering the PNG time course, and E-phase stimuli prolong the E phase. Stimulus parameters were adjusted so that the magnitude of the reflex change in phase timing was the same for both vagus nerves. In some cases, a perfect match could not be completely achieved. Figure 4 shows a typical stimulus protocol for a late E neuron (Fn). The test stimulus patterns (traces 3 and 4) were separated by baseline cycles with a ramp frequency input during the I phase, which shortened TI, and a low-level tonic (5 Hz) input during the E phase. Inserting such baseline input patterns between test and no-stimulus control cycles mimics baseline conditions in preparations with intact vagus nerves, where test lung-inflation and no-inflation cycles are used to alter baseline PSR inputs. These baseline patterns condition the breathing pattern, producing more normal breathing rates, and produce memory effects that can carry over into subsequent breath cycles (Clark and von Euler 1972; Zuperku and Hopp 1985). However, in this protocol, both no-stimulus control cycles and test cycles are preceded by comparable amounts of conditioning, thus allowing comparisons under similar conditions. The test sequence included no vagal input control cycles, contralateral, ipsilateral, and bilateral step-frequency vagal input patterns during the E phase. This sequence was repeated three to five times for each step frequency and used for the CTHs. Paired comparisons between test and control CTH patterns were based on an equal number of cycles for both conditions. Step frequencies of 10, 20, 40, 80, 120, and 160 Hz were used with a quasi-random order.
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
The discharge patterns of 211 respiratory related neurons in the region of the pre-BC were classified into 13 categories. The I neurons were classified into 6 subtypes: I decrementing (dec), pre-Idec, pre-Iaug, Iaug, I parabolic (p), and late-I neurons as shown by the examples in Figs. 5 and 8. The E neurons were classified into 5 subtypes: Edec, Late-E, Eaug, Eparabolic (e.g., Fig. 6), and pre-Edec, which begins to discharge late in the I phase (e.g., Fig. 15, left). Five neurons were E bulbospinal neurons within the region of the Bötzinger Complex as indicated by their more rostral and lateral location relative to all other neuron types (e.g., Fig. 6, bottom). These neurons were antidromically activated from the spinal cord and yielded positive collisions. Two I neurons with augmenting patterns (Iaug) neurons were identified as superior laryngeal motoneurons, and one Iaug neuron was identified as an inferior laryngeal motoneuron. The remaining 203 neurons (96%) were designated as propriobulbar. Approximately 56% (115/206) of the pre-BC region neurons were I neurons and 
43% (89/206) were E neurons (Fig. 7). Two neurons without spontaneous activity were excited only by lung inflation and designated as pump neurons. The pre-I neurons were the most numerous making up 37% of the I neurons and 20% (42/206) of all I and E neurons. Based on discharge pattern, the pre-I neurons were divided into pre-Idec and pre-Iaug neurons, where preIaug made up 75% of the pre-I neurons. The Edec and late-E neurons were the most numerous making up 55% (49/89) of the E neurons.
|
|
|
|
|
Neuronal responses to vagal afferent input
The responses of 42 I and 37 E neurons to reflex changes in either TI or TE produced by electrically induced PSR inputs or by lung inflation (14 I and 8 E) were analyzed. Three qualitatively different types of response were observed when the CTHs associated with the step frequency inputs were compared with those of the no-input control cycles: excitation, inhibition, and no response. In terms of this study, excitation is defined as an increase in discharge frequency, which may be due to synaptic excitation or to disinhibition, whereas inhibition is defined as a decrease in discharge frequency, which may be due to synaptic inhibition or disfacilitation. In all cases, the afferent input during the I phase shortened TI, while those during the E phase prolonged TE.
I neurons
The decrementing slope of Idec neurons increased with the decrease of TI. This type of inhibition was also seen in the preIdec neurons (Fig. 8). Because the pre-Idec and Idec neurons responded similarly, they were pooled for analysis. All 10 neurons belonging to these two subtypes exclusively exhibited inhibitory responses to all three afferent inputs (Fig. 9, top). The inverse relationship between TI and the decrementing slope of Idec and pre-Idec neurons, which leads to the gradual decrease in their discharge frequency and hence a decrease in their presumed inhibitory effect on E neurons, is consistent with their possible role in the control of the I phase duration by promoting or enabling I offswitching.
|
56% of the cases, whereas the response direction of 86% of the preIaug neurons to the bilateral inputs was the same as the ipsilateral response direction (Fig. 10).
|
50%) or did not respond (50%) to either the ipsi- or contralateral inputs (Figs. 9; e.g., and 11). Most Iparabolic neurons did not respond to any of the afferent inputs (Fig. 9). In the only late-I neuron with a complete protocol, all three inputs produced excitation (Fig. 11). Two of four late-I neurons were excited by contralateral inputs. Bilateral inputs inhibited one of three, had no effect on one of three, and excited one of three late-I neurons (Fig. 9).
|
The decrementing slope of the Edec neurons decreased with the increase in TE (e.g., Fig. 12). All Edec neurons studied responded in the same way to all three inputs (Fig. 13) with the exception of one in which the contralateral input had no affect on the time course of the discharge pattern. In the latter case, the neuron continued to discharge at the level of activity that was present at the end of the control cycle. With the PSR-induced increase in TE, the slope of the Eaug neurons decreased, but their resulting peak frequency was equal to or greater than the peak Fn of control cycles with the contralateral inputs (e.g., Fig. 12, Eaug). In all cases, ipsilateral inputs produced inhibition, whereas bilateral inputs were excitatory in most cases (Fig. 13). The responses of the Eparabolic neurons were inconsistent; however there was a tendency for these neurons to be excited by the contralateral inputs (Figs. 12 and 13).
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
The changes in the discharge patterns of the Idec and Edec neurons were consistently related to the experimentally-induced changes in phase timing, whether they were produced by ipsi-, contra-, or bilateral PSR inputs. With the exception of the late-I neurons, I neurons were not excited by the PSR inputs (Fig. 9), whereas contra- and bilateral inputs excited most E neurons (Fig. 13). Ipsi- and bilateral inputs inhibited most preIaug neurons. Iaug neurons were either inhibited (50%) or did not respond (50%), whereas most Iparabolic neurons were unresponsive. Late-E neurons, which made up the largest number of E neurons studied with PSR inputs, were mainly inhibited by ipsilateral PSR inputs; however, they were excited by contralateral inputs (Fig. 13). In 90% of the cases, their responses to bilateral inputs were in the same direction as those to the ipsilateral inputs (Fig. 10). The responses of the Eparabolic neurons were not consistently related to the three inputs, although contralateral inputs produced small amounts of excitation in their discharge patterns.
Location of pre-BC in dogs
Our main criterion for locating the pre-BC in dogs was the tachypneic response produced by localized microinjections of the excitatory amino acid DLH. This criterion is based on similar responses produced by DLH injections in the pre-BC of other species such as rats and cats (Chitravanshi and Sapru 1999; McCrimmon et al. 2000; Monnier et al. 2003; Solomon 2002; Solomon et al. 1999; Wang et al. 2002). DLH microinjections 0.5 mm rostral or caudal to the center of the pre-BC produce bradypnea, whereas those approximately 0.5 mm more dorsal or ventral or 0.5 mm more medial or lateral are without effect (Monnier et al. 2003). Based on preliminary DLH mapping studies in our dog model, the "tachypneic" region appears to be comprised of a 1- to 2.5-mm section of the ventral respiratory neuron column, located from 1–1.5 mm caudal to 1–1.5 mm rostral to the rostral pole of the inferior olive (4.2–6.5 mm rostral to obex), and includes the ventral portion of the retrofacial nucleus, especially at the more rostral extent of this response region (e.g., Fig. 3).
In dogs with weights similar to the ones in our study (7–11 kg), Fukuda and Koga (1997b) defined the pre-BC on the basis of the location of the pre-I neurons, which made up only 8.6% (14/163) of the total neurons they recorded from in the region of pre-BC. They estimated that the pre-BC extended from 4.2 to 5.2 mm rostral to obex. However, they also showed that the more numerous Idec neurons (35.6% or 58/163) were distributed from 3.5 to 5.5 mm rostral to obex. Thus it is possible that the pre-BC extends over a greater rostral-caudal extent, including the region from which we evoke DLH-induced tachypneic responses. In other studies by the same investigators (Fukada and Koga 1997a; Koga et al. 1998), they suggest that the Bötzinger complex extends from 6 to 7 mm rostral to obex. Because the average rostral coordinate for our neuronal study was 5.5 mm rostral, our recordings are likely to have been from the rostral portion of the pre-BC. It should also be noted that dog size can influence the location and extent of intra-medullary nuclei. For example, Wallach et al. (1983), using 15–25 kg dogs, showed the rostral pole of the inferior olivary nucleus at >7.0 mm rostral to the obex. The subtype composition of our recorded neurons and their propriobulbar nature are consistent with the phenotype of pre-BC neurons or at least, neurons located in a region that contains neuronal circuitry capable of controlling I and E phase timing.
Neuronal axon projections
Other studies have shown that the pre-BC consists mainly of propriobulbar neurons (Dobbins and Feldman 1994; Ellenberger and Feldman 1990). Our results are in agreement with these findings. Only 5 of 211 neurons were found to be bulbospinal neurons. These neurons were located more rostral and lateral to the region that produced tachypnea after DLH injections. Three other neurons were found to be laryngeal motoneurons. Most neurons (203 of 216) could neither be antidromically activated from the spinal cord nor did they register responses with spike-triggered averaging of the superior and inferior laryngeal nerve activities. We conclude that most of these neurons are propriobulbar, although some may have been cranial motoneurons with axons in the pharyngeal branch of the vagus nerve (Bianchi et al. 1995) or possibly within the glossopharyngeal nerve. The pharyngeal branch of the vagus nerve is inaccessible in the dog and could not be isolated. We also did not record from the glossopharyngeal nerve as it only contains very few motoneurons that innervate the stylopharyngeus muscle deep within the pharynx (Martin 1996).
Neuronal discharge patterns
This study found that the canine pre-BC region contains a wide variety of I, E, and phase spanning neurons with discharge patterns similar to those previously reported by others in rats (McCrimmon et al. 2001; Sun et al. 1998), cats (Connelly et al. 1992; Schwarzacher et al. 1995) and dogs (Fukuda and Koga 1997b).
Pre-I neurons
The pre-I neurons were the most numerous (42 of 115 = 37%) of the recorded I neurons. The onset of activity of the preI neurons preceded the onset of phrenic activity by >100 ms, and they continued to discharge throughout the I phase. About 74% had augmenting patterns while the remaining 26% had decrementing patterns. The discharge patterns of preI neurons with a decrementing pattern are somewhat similar to the type I neurons reported in other species (Connelly et al. 1992; Guyenet and Wang 2001; McCrimmon et al. 2001; Schwarzacher et al. 1995; Sun et al. 1998) as well as in dogs (Fukuda and Koga 1997b). We did not encounter any biphasic-activity typical of the type II preI neurons, rather the pre-Iaug neurons had an augmenting or augmenting/plateau pattern with peak discharge frequencies occurring near the end of the I phase.
While most pre-Iaug neurons showed some inhibition to the PSR inputs near the end of the I phase, this inhibition was relatively small compared with the PSR effects on the pre-Idec neurons (e.g., Fig. 8).
Other studies have indicated that a greater number of pre-I neurons are found in the pre-BC compared with the Bötzinger complex or rostral VRG. The percentage of pre-I neurons within the pre-BC of other preparations has been reported to be 3–11% for cats (Connelly et al. 1992; Schwarzacher et al. 1995), 33–47% for rats (Guyenet and Wang 2001; McCrimmon et al. 2001; Sun et al. 1998), 19% for mice (Paton 1996), and 10% for dogs (Fukuda and Koga 1997b). It is possible that differences between our study results and those of the latter study (36.5 vs. 10%) may be due to sampling differences and that the center of our recording site was on average 0.6 mm more rostral than the center of the site reported in that study. The site that resulted in a maximum tachypneic response to DLH microinjections defined the coordinates used for this study.
Comparisons of neuron subtype composition among species
The results of this study suggest that differences exist among dogs, cats, rats, and mice in the proportion of neuron subtypes that comprise the pre-BC. Not all subtypes have been reported for each species, but some of the more common subtypes have been reported for all four species (Chitravanshi andSapru 1999; Connelly et al. 1992; Fukuda and Koga 1997b; Guyenet and Wang 2001; Hayashi et al. 1996; McCrimmon et al. 2001; Paton 1996; Schwarzacher et al. 1995; Sun et al. 1998). The frequency distribution is: rat > dog > mouse > cat for pre-I neurons, cat dog rat for Iaug neurons, dog > cat > rat for Idec neurons, dog cat mouse for late-I neurons, rat mouse > cat dog for Edec neurons, rat > cat dog for pre-E neurons, mouse > dog > cat > rat for Eaug/parabolic neurons, and cat dog for late-E neurons. A notable difference was rat >> cat for pre-I neurons. In rats, 37% of neurons were Edec and no Eaug neurons were found.
There were some differences in the subtype composition between our study and another canine study where a substantial number of pre-BC neurons (n = 183) were recorded (Fukuda and Koga 1997b). We found more preI neurons (37 vs. 10%) but few Idec neurons (4 vs. 34%). We also found more E neurons of all subtypes (44 vs. 18%). These differences may be due to our sampling site being 0.6 mm more rostral than that of the other canine study.
Direct and indirect effects of vagal afferent inputs
Electrically evoked pulmonary vagal input patterns elicit highly reproducible changes in phase timing that mimic those of the Hering-Breuer reflexes (Feldman and Gautier 1979) and have unique advantages (D'angelo 1985; Trenchard 1977; Zuperku et al. 1982a). Because the vagi are transected, ventilation remains constant despite the induced changes in central respiratory phase timing, preventing oscillations in blood gases. Compared with lung inflation mediated inputs, highly controllable, graded input patterns can be produced, and the effects of ipsi-, contra-, and bilateral inputs can be isolated and compared.
The central processing of the vagal afferent inputs, as seen by its effects on the neuronal discharge patterns, appears to occur in at least two different modes. In one mode, the neuronal discharge patterns are either compressed or expanded along the time axis with the changes in phase duration, but the morphology of the pattern is preserved. Examples of this are shown by the responses of the Idec and Edec neurons of Figs. 8 and 12, respectively, and the Eaug pattern in response to the contralateral input (Fig. 12). Mechanistically, this may be due to the neuronal patterns being sculptured by elements of the phase timing circuitry. In addition, this may be combined with presynaptic, long time-constant processing of the afferent input (Bajic et al. 1989; Zuperku and Hopp 1985, 1987; Zuperku et al. 1982a). The second mode of vagal afferent processing appears to be mediated by rapid mechanisms and is more directly transmitted to the neurons, possibly over oligosynaptic pathways. The effect of this mechanism, which appears to occur in combination with the more indirect slower processing, manifests itself as a parallel upward or downward shift in the neuronal discharge patterns. For example, this effect can be seen by comparing the contra- and ipsilateral responses of the Eaug neuron of Fig. 12 and the late-E neurons of Fig. 14.
Responses of neurons with decrementing patterns
The response characteristics of the Idec and Edec neurons were consistently related to the changes in phase timing produced by ipsi-, contra-, or bilateral afferent PSR inputs (Figs. 9 and 13). The negative slope of the Idec and preIdec neurons increased in proportion to the decrease in TI, while the negative slope of the Edec neurons decreased in proportion to the increase in TE. These findings are consistent with the responses of Idec and Edec neurons previously reported for other species, even though such neurons may not have been within the pre-BC, and they are also consistent with their proposed role in phase timing (Cohen 1979; Cohen and Feldman 1977; Ezure 1990; Feldman and Cohen 1978; Hayashi et al. 1996; Zuperku and Hopp 1987). In rat Edec neurons, excitatory postsynaptic potentials are produced by electrically evoked vagal afferent volleys acting via a paucisynaptic pathway, suggesting that increases in Edec activity produced by lung inflation are mediated by excitation rather than disinhibition (Hayashi et al. 1996). Alternatively, there is evidence for reciprocal inhibition between Edec and Eaug neurons (Shen et al. 2003), which suggests the possibility that an inflation mediated inhibition of Eaug activity (Feldman and Cohen 1978; Manabe and Ezure 1988) could contribute to the increase in Edec activity.
In most network models, it has been assumed that reciprocal inhibition between pools of Idec and Edec neurons provides the main substrate for rhythm generation (Balis et al. 1994; Duffin et al. 1995; Ogilvie et al. 1992; Richter et al. 1986; Rybak et al. 1997; Zuperku et al. 1982b). Recent studies have shown that these neurons are either GABAergic or glycinergic (Ezure et al. 2003; Okazaki et al. 2001) supporting their inhibitory role (Ezure and Manabe 1988; Lindsey et al. 1987; Shen et al. 2003). The results of this study support a control-of-timing role for these neurons.
Responses of Eaug and late-E neurons
Late-E neurons (29% of pre-BC region E neurons) may be similar to the E2 augmenting neurons reported in other studies (Cohen 1969; Cohen et al. 1985; Hayashi et al. 1996; Schwarzacher et al. 1995; Shen et al. 2003). However, their discharge pattern is different from the E bulbospinal neurons of the canine caudal VRG, which begin to discharge immediately at the end of the I phase and continue to discharge throughout the E phase. Most of the canine E bulbospinal neurons have decrementing patterns, and they are excited at low transpulmonary pressures and strongly inhibited at higher transpulmonary pressures (Bajic et al. 1992; Tonkovic-Capin et al. 2000). The Eaug and late-E neurons of the pre-BC do not exhibit an inflation pressure-dependent bidirectional response. There are, however, similarities among the pre-BC region and bulbospinal E neurons in terms of the laterality of the responses. In both cases, the ipsilateral vagal input produces mainly inhibition, whereas the contralateral vagal input either has no effect or produces excitation (e.g., Figs. 12 and 14) (Tonkovic-Capin et al. 1991).
A consistent finding for most late-E neurons was that, within the same neuron, ipsilateral inputs produced inhibition while contralateral inputs produced excitation, however the response to bilateral inputs was inhibitory (Figs. 10 and 13). This suggests a dominant role for the ipsilateral inputs. With intact vagal nerves, lung inflation also produced inhibition (e.g., Fig. 15), which is consistent with previous reports of Eaug neurons in the VRG and Bötzinger complex (Feldman and Cohen 1978; Manabe and Ezure 1988). Because an ipsilateral input inhibits ipsilateral late-E neurons while at the same time exciting contralateral late-E neurons, it is possible that this mechanism adjusts E muscle activities in response to uneven inflation of the lung. If the control of these neurons serves a compensatory function, then perhaps the late-E neurons function in a pattern-generating capacity rather than in the control of phase timing. The reflex effects on the discharge patterns of these propriobulbar neurons may be relayed to other postsynaptic brain stem respiratory neurons controlling, in turn, their discharge patterns. Because both ipsi- and contralateral inputs delay the onset of late-E activity, they most likely have their maximum effect in the second half of the E phase.
Responses of other pre-BC region neurons
About 50% of the Iaug neurons were inhibited to a small degree near the end of the I phase with all three afferent inputs, whereas the time courses of the other 50% were unaltered but terminated with the ending of the I phase. These responses are similar to those previously reported for I
or I(0)/I(–) neurons but unlike the I
or I(+) neurons (Cohen and Feldman 1984). The time courses of I neurons with parabolic patterns were mainly unaffected by the vagal afferent inputs. Thus it appears that Iaug and Iparabolic neurons are not involved with phase timing.
The only I neurons to be excited by vagal afferent inputs were the late-I neurons, although this did not happen in every case (Figs. 9 and 11). It has been previously suggested that the late-I neurons may play a role in the graded I inhibition and irreversible termination of the I phase (Baker and Remmers 1980; Cohen and Feldman 1984; Cohen et al. 1993; Haji et al. 2002).
Summary
The canine pre-BC region consists of a heterogeneous mixture of I and E neuron subpopulations similar to those reported in other species. In general, the neuronal responses to pulmonary vagally mediated afferent inputs were similar to those of other species, even though those neurons were recorded in regions outside of the pre-BC. Thus it appears that the response properties of the pre-BC neurons are not unique. However, the late-E neurons of the canine pre-BC region are not found in the caudal VRG. The time courses of the discharge patterns of I and E neurons with decrementing patterns were consistently related to the duration of I and E phases, respectively. These were the only neuron subpopulations to have this property, which suggests an important role in phase timing control.
|
GRANTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: E. J. Zuperku, Research Service/151, Zablocki VA Medical Center, Milwaukee, WI 53295 (E-mail: ezuperku{at}mcw.edu)
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Bajic J, Zuperku EJ, Tonkovic-Capin M, and Hopp FA. Expiratory bulbospinal neurons of dogs: I. Control of discharge patterns by pulmonary stretch receptors. Am J Physiol Regulatory Integrative Comp Physiol 262: R1075–R1086, 1992.
Baker JP and Remmers JE. Temporal correlation of graded reversible inspiratory inhibition with discharge patterns of late inspiratory neurons located in the dorsal respiratory group in cats. Brain Res 200: 331–340, 1980.[CrossRef][ISI][Medline]
Balis UJ, Morris KF, Koleski J, and Lindsey BG. Simulations of a ventrolateral medullary neural network for respiratory rhythmogenesis inferred from spike train cross-corelation. Biol Cybern 70: 311–327, 1994.[ISI][Medline]
Bianchi AL, Denavit-Saubie M, and Champagnat J. Central control of breathing in mammals: neuronal circuitry, membrane properties, and neurotransmitters. PhysiolRev 75: 1–45, 1995.
Butera RJ Jr, Rinzel J, and Smith JC. Models of respiratory rhythm generation in the pre-Bötzinger complex. I. Bursting pacemaker neurons. J Neurophysiol 81: 382–397, 1999a.
Butera RJ Jr, Rinzel J, and Smith JC. Models of respiratory rhythm generation in the pre-Bötzinger complex. II. Populations of coupled pacemaker neurons. J Neurophysiol 81: 398–415, 1999b.
Clark FJ and vonEuler C. On the regulation of depth and rate of breathing. J Physiol 222: 267–295, 1972.
Chitravanshi VC and Sapru HN. Phrenic nerve responses to chemical stimulation of the subregions of ventral medullary respiratory neuronal group in the rat. Brain Res 821: 443–460, 1999.[CrossRef][ISI][Medline]
Cohen MI. Discharge patterns of brain-stem respiratory neurons during h-b reflex evoked by lung inflation. J Neurophysiol 32: 356–374, 1969.
Cohen MI. Neurogenesis of respiratory rhythm in the mammal. Physiol Rev 59: 1105–1173, 1979.
Cohen MI and Feldman JL. Models of respiratory phase-switching. Fed Proc 36: 2367–2374, 1977.[ISI][Medline]
Cohen MI and Feldman JL. Discharge properties of dorsal medullary inspiratory neurons: relation to pulmonary afferent and phrenic efferent discharge. J Neurophysiol 51: 753–776, 1984.
Cohen MI, Feldman JL, and Sommer D. Caudal medullary expiratory neurone and internal intercostal nerve discharges in the cat: effects of lung inflation. J Physiol 368: 147–178, 1985.
Cohen MI, Huang W-X, Barnhardt R, and See WR. Timing of medullary late-inspiratory neuron discharges: vagal afferent effects indicate possible off-switch function. J Neurophysiol 69: 1784–1785, 1993.
Connelly CA, Dobbins EG, and Feldman JL. Pre-Bötzinger complex in cats: respiratory neuronal discharge patterns. Brain Res 590: 337–340, 1992.[CrossRef][ISI][Medline]
D'angelo E. Verification of a model for the mechanisms controlling expiratory duration in rabbits under various conditions. Respir Physiol 59: 239–264, 1985.[CrossRef][ISI][Medline]
Dobbins EG and Feldman JL. Brainstem network controlling descending drive to phrenic motoneurons in rat. J Comp Neurol 347: 64–86, 1994.[CrossRef][ISI][Medline]
Dogas Z, Krolo M, Stuth EA, Tonkovic-Capin M, Hopp FA, McCrimmon DR, and Zuperku EJ. Differential effects of GABAA receptor antagonists in the control of respiratory neuronal discharge patterns. J Neurophysiol 80: 2368–2377, 1998.
