Expiratory bulbospinal neurons of dogs. II. Laterality of responses to spatial and temporal pulmonary vagal inputs

Transcription

1 xpiratory bulbospinal neurons of dogs.. Laterality of responses to spatial and temporal pulmonary vagal inputs MSLAV TONKOVC-CAPN, DWARD J. ZUPRKU, JURCA BAJC, AND FRANCS A. HOPP Zablocki Veterans Affairs Medical Center and Department of Anesthesiology, Medical College of Wisconsin, Milwaukee, Wisconsin Tonkovibcapin, Mislav, dward J. Zuperku, Jurica the phase (decrementing pattern). The F, of type A Bajib, and Francis A. Hopp. xpiratory bulbospinal neurons neurons augments during the first.5-l s of the phase of dogs.. Laterality of responses to spatial and temporal pul- and remains at a plateau level during the balance of the monary vagal inputs. Am. J. Physiol. 262 (Regulatory ntegraphase (augmenting/plateau pattern). The companion tive Cornp. Physiol. 31): R187-R195, Pulmonary mechanoreceptors with vagal fibers produce a combination of study demonstrated that inputs from pulmonary recepexcitation and inhibition in the majority of the expiratory bul- tors with vagal afferents have a combined excitatorybospinal (BS) neurons of dogs. Both aspects of this transpul- inhibitory effect on type D BS neurons that is Pt monary pressure-dependent neuronal response appear to be dependent. n the range from -1 to 5 mmhg, increases slowly adapting and activated at low pressure levels, suggesting in P, produce a graded linear increase in F,, while the involvement of the slowly adapting pulmonary stretch increases in Pt >4--5 mmhg produce a graded linear receptors (PSRs). The purpose of the present study was to reduction in F,. The smaller subpopulation of type A determine the contribution of different afferent pathways to each of the response components and to characterize the spatial neurons (2-3% of total) exhibit only a linear inhibiand temporal processing of ipsi-, contra-, and bilateral vagal tory F,-Pt relationship for P, >3-4 mmhg. n addition, afferent inputs by two types of BS neurons. For this purpose preliminary electromyograms (MGs) obtained from low-intensity electrical stimulation of the intact, desheathed, the T6-Ts internal intercostal muscles exhibited Ptvagus nerves was used in thiopental sodium-anesthetized para- dependent bidirectional responses similar to those of the lyzed dogs. The phrenic neurogram was used to synchronize type D BS neurons. From the nature of this bidirecboth ventilation and stimulation. During test respiratory cycles, tional phenomenon, it would appear that the excitatory pulse trains (4-5 s duration) were applied during the neural expiratory phase to each and both vagus nerves. The mean response component may be involved in the control of discharge frequency Cpn) during the stimulus period was end-expiratory lung volume and the inhibitory compoobtained from cycle-triggered histogram data. Plots of Fn vs. nent may be involved in expiratory breaking and the stimulus strength and pn vs. stimulus frequency suggest that control of the volume trajectory (18). A significant inhibition of both type D and type A BS neurons is mediated finding of these studies is that the time course of the mainly by the ipsilateral vagus nerve, and that the excitation of discharge pattern of BS neurons is highly dependent type D neurons is mediated bilaterally. These conclusions are on the instantaneous level of Pt and, to a much lesser also supported by inflation responses obtained before and after extent, on time. unilateral vagotomies. Differences in latencies and spatial and temporal summation characteristics suggest the possible The mechanisms whereby pulmonary vagal afferents involvement of different 1) types of PSRs, 2) central pathways, produce both excitation and inhibition of the same BS and/or 3) synaptic mechanisms in the biphasic response of the neuron are unknown. The main groups of pulmonary caudal ventral BS neurons to lung inflation. mechanoreceptors that may be involved include the neuronal processing; differential reflex responses; neural path- slowly adapting pulmonary stretch receptors (PSRs), ways; respiratory neurons; vagal afferents; control of breathing the rapidly adapting pulmonary receptors (RAPRs), pulmonary receptors innervated by C-fibers, and bronchial receptors with C-fibers. Because the Pt activation XPRATORY BULBOSPNAL (BS) neurons of the ventral respiratory group (VRG) located in the region of the nucleus retroambigualis provide phasic excitatory input patterns, via motoneurons, to the expiratory () muscles of the chest wall and abdomen (9, 14) and provide inhibition of thoracic inspiratory () motoneurons during the phase of the respiratory cycle. Activity arising from the pulmonary mechanoreceptors both excites and inhibits the discharge activity of the BS neurons (2, 9, 15). n the companion paper (4), we characterized the responses of two types of BS neurons, D and A, in the caudal VRG of dogs to lung inflation patterns. The two types of neurons were classified according to their spontaneous discharge patterns in the absence of lung inflation with a transpulmonary pressure (P,) of zero (pneumothorax). Typically, the discharge frequency (F,) of type D neurons is maximal early in the phase and progressively decreases throughout the balance of threshold of the PSRs, but not of the other pulmonary receptors, is in a range that corresponds with the thresholds of the excitatory and inhibitory components of the neuronal response, it is possible that the PSRs mediate both aspects of the response (4). Two types of PSRs located in the intrathoracic airways, with different recruitment thresholds and sensitivities, have been described by Miserocchi and Sant Ambrogio (17). Thus it is possible that different types of PSRs mediate the excitation and inhibition. Alternatively, different cen- tral pathways from the same PSRs may be involved. For example, the rapid excitatory component of the PSR input to the 1-p neurons of the ventrolateral nucleus of the solitary tract in cats is mediated by the ipsilateral vagus nerve, while the slower inhibitory component of the PSR input is mediated via both vagus nerves (3). n addition, different modes of processing, such as multi- R187

2 R188 XPRATORY NURONS OF DOGS: LATRALTY OF RSPONSS plicative in contrast to additive, may be involved in the production of the bidirectional phenomenon (19). A better understanding of the underlying mechanisms mediating the BS neuronal responses to lung inflation and volume would aid in delineating the role of the pulmonary mechanoreceptor control of muscles and their contribution to breathing. Accordingly, the present studies were undertaken to determine whether different afferent pathways are involved in mediating the biphasic responses and to characterize the spatial and temporal processing of ipsi-, contra-, and bilateral pulmonary afferent inputs by the BS neurons. MTHODS The methods used in the present study have been reported in detail in our companion paper (4), and only the additional pertinent methods are delineated. n 16 mongrel dogs (lo-2 kg) l- to 2-cm lengths of both cervical vagus nerves were isolated from the surrounding tissues, desheathed, placed (intact) on bipolar platinum stimulation electrodes, and submersed in pools of warmed mineral oil. n three dogs, an additional pair of recording electrodes were placed on a second length of desheathed vagus nerve, lo-12 cm cranial to the stimulating electrodes, for the purpose of measuring the conduction velocities of the electrically activated vagal afferent fibers. Protocol. The onset of the phrenic neurogram (PNG) was used to trigger a solenoid ventilator so that lung inflation occurred during neural inspiration. The duration of airflow coincided with the duration of neural inspiration (T,), and the airflow rate was adjusted to produce normocapnea (TCH,. z 5.5%). Two test lung inflation procedures, no inflation during neural inspiration and a slow augmenting ramp inflation during the neural phase (Fig. 1, A and B, respectively), were used to identify the BS neuron type as either A or D as described in the accompanying paper (4). The laterality of the BS neuronal responses was evaluated during test respiratory cycles by electrically induced, step input patterns of 4 s duration applied to the ipsilateral vagus nerve (Fig. C), to the contralateral vagus nerve (Fig. ld), and bilaterally. Constant-current stimulators were used to deliver 1~ps duration pulses with amplitudes in the range of 2-2 PA. For the spatial summation studies, the step frequency was set at 2 Hz. Stimulus strengths were expressed in multiples of a threshold level, where the threshold level was determined as that level which produced a discernible change in the time course of BS neuronal activity; this level invariably produced an increase in duration (Tn). Stimulus strengths of 1,1.2,1.5, 2, and 3 times threshold (T) level were applied in random order. However, to make comparisons with minimal time-dependent effects, the same strength (in terms of thresholds) was used ipsi-, contra-, and bilaterally before changing the strength to a new level. Based on evoked vagal compound action potentials, a strength of 3T activated only the largest myelinated fibers with minimum conduction velocities of -25 m/s, which include fibers from PSRs and, most likely, RAPRs. Because of the rather long conduction distances (lo-12 cm) and narrow stimulus pulse width (1 ps), the effect of utilization time on the conduction velocity calculation is minimal (i.e.,.1 ms utilization time out of 4. ms conduction time). For the temporal summation studies, a strength of 3T was used with step frequencies of 5, 1, 2, 4, and 8 Hz. The presentation order of the step frequencies was randomized, and a given step frequency was then applied ipsi-, contra-, and bilaterally. Cycle-triggered histograms (CTHs) of BS neuronal activity were generated from five to six test respiratory cycles for each Downloaded from by on September 18, 216 2Hz - Fig. 1. Paradigm of experimental test procedures used to identify neuron type and characterize neuronal responses. PNG, phrenic neurogram; P,, transpulmonary pressure; F,, neuronal discharge frequency calculated every 1 ms; NA, expiratory neuronal activity. A: no-inflation [during neural inspiratory () phase] test. B: augmenting ramp inflation during expiratory () phase. C: 4-s pulse train, 3 x threshold strength (3T, horizontal bar) applied to ipsilateral cervical vagus nerve during phase reduces F,,. D: 4-s pulse train, 3T (horizontal bar) applied to contralateral cervical vagns nerve during phase prolongs duration with no effect on F,,.

3 XPRATORY NURONS OF DOGS: LATRALTY OF RSPONSS RlO89 step strength and/or frequency for the ipsilateral, contralateral, data clearly demonstrate that, with the recruitment of a and bilateral conditions. Least-squares linear regression was greater number of afferent fibers as strength was used to determine the best-fit line through the portion of the increased from 1T to 3T, inhibition is produced via the CTHs bracketed by the dotted vertical lines in Fig. 3. From this ipsilateral afferents, since the same inputs applied to the analysis, the mean discharge frequency (Fn) and the slope of the activity were quantified. Plots of these variables vs. stimulus contralateral vagus nerve resulted in essentially no strength and frequency were compared for the ipsi-, contra-, and change in pattern other than that which is due to the bilateral vagal inputs. Two-way analysis of variance (ANOVA) reflex ProlOrWtion of G+ n this Particular case, the with main factors of stimulus and input pathway was used with bilateral inputs produced a greater degree of inhibition Scheffe s procedure to determine significant differences (P < than did the ipsilateral inputs alone, as can be noted by.5) in the levels of the main effects. One-way ANOVA was the vertical distance between CTHs for strengths of 1T used for each input pathway to determine the presence of sig- and 2T, in each case. nificant trends (via orthogonal polynomials) as a function of With the use of the same protocol, data from a type A stimulus level. BS neuron, as determined from inflation responses, are RSULTS shown in Fig. 3, right. The thick line control CTH illus- A comparison between responses produced by step trates the typical ramp/plateau pattern of the type A neurons. While a small degree of inhibition can be noted inflations and electrically induced step input patterns is shown for the same type D BS neuron in Fig. 2. Step for the 2-Hz contralateral step inputs, the major source inflation (Sl) produced about a 25% increase in $, above of inhibition is mediated via the ipsilateral input. No the peak level of the control pattern (thick solid line excitatory response was observed during the step input CTH, Fig. 2, middle left). Progressively higher step inflaperiod, and most CTHs exhibited a postinhibitory excition levels produced greater degrees of inhibition. The tation after the cessation of the step input. These data time courses of these responses were linear, with slopes indicate that this BS neuron was very sensitive to the that ranged from slightly negative (Sl) to slightly positive vagally mediated inhibition, since at 1.5T the bilateral (S2 and S3). lectrical activation of ipsilateral vagal input was able to essentially silence the activity. afferents at 2 Hz and 1T produced a discharge pattern of The average (&S) n values of 22 type D and 6 type A nearly constant frequency at the level of the peak control BS neurons are shown in Fig. 4. The pn-stimulus F, (Fig. 2, top right). Further increases in ipsilateral stim- strength plots for each neuron were normalized by ulus strength produced greater degrees of inhibition, expressing & as a percentage of the value obtained at the whereas the contralateral step inputs produced only 1T level [F,* = 1 x &#JlT)]. These data show that increases in F, (Fig. 2, midde right). The time courses of for both types of neuron, ipsilateral vagal fibers mediate all of these responses were essentially linear, similar to those produced by the step inflations. As with step inflations, the excitatory and inhibitory aspects of the response were observed using the electrically induced afferent patterns. The responses of a typical type D neuron to electrically induced ipsi-, bi-, and contralateral input patterns are shown as superimposed CTHs for each condition and strength (Fig. 3, left). The 2-Hz step input was applied at the onset of the phase and lasted 4 s (not shown). These 8 w 6 cn & 4-2 G w 6 Ln & 4-2 c L the inhibition. At 3T, there was an average reduction of 5% in Fn* for type D neurons, and a reduction to 4% of control for the type A neurons. Contralateral inputs to the type D BS neurons produced an average increase of 8-11% in Fn*. The average Fn* of type A neurons was unaltered by the contralateral inputs. The responses to bilateral inputs were similar to those for ipsilateral inputs, but exhibited less net inhibition, especially for type D neurons. Relative to the peak F, of control cycles (CTHs), the &JlT) t S at 2 Hz for type D neurons 4 t Fig. 2. Responses [cycle-triggered histograms (CTHs)] of a type D neuron to step inflations (left) and electrically induced, step vagal afferent inputs (right). Thick solid traces, control cycle data. Left: step P, level Sl produced a small amount of excitation, while step level S3 produced a moderate degree of inhibition. Top right: electrically induced, ipsilateral vagal afferent inputs produced increasing levels of inhibition as stimulus strength was increased. T, threshold stimulus strength. Middle right: contralateral vagal inputs produced a small increase in F,. TM (SC) TM (SC)

4 R19 XPRATOR Y NUR ONS OF DOGS: LATRALTY OF RS PONSS TYP D 12 ;1l ; 8 & 6-4 L 2 glee 12 ; 8 & 6-4 L 2 12 g1 1 8 & 6-4 L 2 t. PSLRTRAL TYP A TM (SC) Fig. 3. Responses of a typical type D neuron (left) and type A neuron (right) to electrically induced vagal inputs. CTHs (F,) for control (thick-line traces, without electrically induced input) and each input condition (strength labeled). Left: 4-s duration, 2-Hz pulse trains triggered at onset of phase were used. Dotted vertical lines, analysis period during steady-state portion of the response. Line through CTH, least squares, best fit, used to quantify response slope and average F, (F,) during analysis period. Right: stimulus duration was 5 s. was 8.1 t 3.2, 81.7 t 4., and 81.8 t 3.% for the ipsi-, the slopes increased to Hz/s and were not differcontra-, and bilateral inputs, respectively. For type A ent from each other for strengths from 1.2T to 3T, but neurons, Fn(lT) was 82.4 t 4.1,86.2 t 3.3, and 75.5 t 4.8 were positive (i.e., >O; P <.5). Such small increases in of the peak control F, for the ipsi-, contra-, and bilateral activity slope contribute very little (51%) to the overall inputs, respectively. response. The time-dependent effects of the step inputs were An example of the effects of temporal summation of quantified via the slope of the activity during the steady- vagal afferent-mediated inputs for a type D neuron is state portion of the step response (e.g., Fig. 3, regression given in Fig. 5. At a stimulus strength of 3T, various step line slopes). The activity slopes of 21 type D neurons at frequencies were applied for each of the input path 1T were all negative or decrementing (i.e., CO; P <.1) conditions. For the ipsilateral inputs, the same degree of with mean values (&S) of t 3.5, -9.3 t 2.9, and inhibition was produced at all step frequencies. Con t 1.1 Hz/s for ipsi-, contra-, and bilateral inputs, tralateral inputs produced graded excitatory responses respectively. As stimulus strength increased from 1T to with almost a doubling of F, at 4 Hz relative to that at 3T, the slopes rapidly approached an asymptotic value of 1 Hz. For bilateral step inputs, it appears that the conzero. For all three input pathway conditions in type A tralaterally mediated excitatory effect tends to offset the neurons, the response in terms of activity slopes was ipsilaterally mediated inhibitory effect, as can be noted similar. The average slopes of six type A neurons at 1T by comparing the F* vs. stimulus frequency plots of bilatwere not different from zero for ipsi- and contralateral era1 and ipsilateral in Fig. 5, right. The average (M) inputs, while the average activity slope for bilateral effect on & of increasing step input frequency for 15 type inputs was 1.1 t.4 Hz/s. As stimulus strength increased, D and 3 type A neurons is shown in Fig. 6. These data NP D NURONS 12 1 TYP A NURONS Fig. 4. Average k S normalized & vs. stimulus strength relationships for 21 type D (left) and 6 type A (right) neurons. Fn for individual neurons has been normalized (&*) relative to the 1T strength for each of 3 input conditions. C, contralateral inputs; B, bilateral inputs;, ipsilateral inputs. l. 2 N= STMULUS STRNGTH STMULUS STRNGTH (X THRSHOLD) (X THRSHOLD)

5 XPRATORY NURONS OF DOGS: LATRALTY OF RSPONSS R ; 6-4 -c 1 -B Y t 2 1. BLATRRL - 8 _ NPUTS 8 - z - 6- = Fig. 5. Responses of a type D neuron to changes in step frequency of electrically induced vagal afferent inputs. Stimulus strength, 3T. Horizontal bar, pulse-train duration. Left: step frequency (Hz) for each response (CTH). Right: corresponding plots of P, vs. stimulus frequency for each input condition TM (SC) STMULUS FRQUNCY (HZ) were normalized relative to the & at 1T (2 Hz), as was are shown for ipsilateral inputs (top PSTHs) and conthe case for the stimulus strength data. For type D neu- tralateral inputs (middle PSTHs). n each case, the botrons, no significant trends or differences (P <.5) in F,* tom plots in Fig. 7 show the effects of stimulus strength were found as a function of step frequency for both ipsi- from the 2-Hz step input patterns on the average F, and bilateral inputs, whereas increases in contralateral during the stimulus period. The &stimulus strength plot step frequency produced significant increases in &* (P < for the type D neuron of Fig. 7A clearly shows that the.1). For type A neurons, a trend analysis indicated ipsilateral inputs reduced fin while the contralateral that increases in ipsilateral step frequency produced a inputs increased &. The PSTH (ipsilateral) indicates a significant graded reduction in F,* (P < O.Ol), whereas marked reduction in the probability of firing, with a there was no significant change in F,* with stimulus fre- latency of -7.5 ms and a duration >2 ms. The correquency for contra- or bilateral inputs. For both neuron sponding PSTH for contralateral inputs indicates a types, the responses to contralateral inputs were signifi- marked increase in the probability of firing at a latency of cantly greater (P <.1) than those to either ipsi- or 22.5 ms; a small early reduction also occurred. These bilateral inputs. directional changes in firing probability are in agreement Short time-scale neuronal responses were analyzed with those of the pn-strength plots (Fig. 7A, bottom). The using poststimulus time histograms (PSTHs), in which &-strength plot for the type D neuron of Fig. 7B, bottom, each vagal nerve stimulus pulse was used to trigger the indicates that ipsilateral inputs produced inhibition, analysis at time. Typically, the synchronizing effect of while little or no change is indicated for the contralateral each stimulus was followed over a period of 4 ms. ach inputs. The corresponding PSTH (ipsilateral) shows a PSTH contained data obtained over the range of stimulus long-lasting reduction in the probability of firing after a strengths (lt-3t). xamples of PSTHs for two type D latency of 5.5 ms, while an increase in the firing probaneurons (Fig. 7, A and B) and one type A neuron (Fig. 7C) bility occurred after a latency of ~32 ms. For the con N=3 c 1 / T 12- z A 5 loo- $t 8- - *c A 6- Fig. 6. Average t S normalized Fn (Fn*) vs. step stimulus frequency for 15 type D (left) and 3 type A (right) neurons. Stimulus strength, 3T. J, data for each neuron were normalized with respect to corresponding & value at 1T and 2 Hz STMULUS FRQUNCY (Hz) -r STMULUS FRQUNCY (Hz)

6 R192 XPRATORY NURONS OF DOGS: LATRALTY OF RSPONSS PSLRTRAL PSLATRRL ps.4!? g.ekl t= z.3 t.1 t y -82 z Sam c -2 Y ti b a.1 n.1 Ln i% PSLATRAL ii L TM (tlsc)!2 t-4-1 ;;; f ii: << 1 CONTRALATRAL a / Tfl : ns c: STlMULUS STRNGTH STMULUS STRNGTH STMULUS STRNGTH Fig. 7. Poststimulus time histograms (PSTHs) for 2 type D neurons (A and B) and a type A neuron (C). Ordinate: frequency of occurrence given as spikes/trigger. Bottom illustration in each panel is corresponding Fn vs. stimulus strength plot for each neuron. PSTHs are based on 2-Hz stimulus frequency and strengths from 1T to 3T. Bin width,.5 ms. Superimposed thick line traces are phase-adjusted, 5-ms running averages used to identify trends. Horizontal dashed lines, average level of initial 5 ms of PSTH. A: ipsilateral, 776 triggers, 1,437 spikes; contralateral, 1,391 triggers, 2,359 spikes; B: ipsilateral, 1,32 triggers, 1,499 spikes; contralateral, 1,56 triggers, 3,571 spikes. C: ipsilateral, 1,281 triggers, 1,722 spikes; contralateral, 1,678 triggers, 4,323 spikes. tralateral inputs, the PSTH indicates a reduction followed by an increase in firing probability that may account for the lack of a net change in the &strength plot. The PSTHs of a typical type A neuron, which was strongly inhibited by ipsilateral inputs but little affected by contralateral inputs, are shown in Fig. 7C. A longlasting decrease in the firing probability occurred at a latency of -8.5 ms for ipsilateral inputs, while a smaller increase in probability was found for the contralateral inputs after a latency of -1 ms. As indicated by these PSTH examples, a combination of inhibitory and excitatory effects coexist for both ipsiand contralateral inputs. To estimate the onset latencies of reductions or increases in the probability of firing, the moving averages were used to detect significant trends; however, the PSTHs were used to estimate the actual onset latencies. Differences between the PSTH and corresponding moving-average values were used to calculate confidence limits (+2.3 Ss; 95%) that were applied to the mean level of the initial 5 ms of each PSTH. A trend was considered significant if the moving average passed below or above the confidence limits. For ipsilateral inputs, a significant reduction in firing probability was found to occur in 17 of 18 neurons, with an average (MD) onset latency of 7.1 t 1.6 ms. n 17 of 18 neurons, contralateral inputs produced a decrease in the probability of firing, with an average latency of 8.5 t 2.8 ms. A paired analysis based on 14 neurons indicated that there was no significant difference between latencies for the two pathways. A significant increase in the firing probability above the average level of the initial 5 ms of each PSTH was found in 12 of 18 neurons for ipsilateral inputs and 15 of 18 neurons for contralateral inputs. The latency values were 25.4 t 4.4 and t 4.9 ms for ipsi- and contralatera1 inputs, respectively. A limitation to the detection of an increase in probability due to excitation earlier than ms is that such excitation may be reduced or masked if it occurred during the period of reduced probability of firing, which invariably started at short latency and persisted for X-15 ms. Four examples of the effects of unilateral vagotomy on type D BS neuronal responses to lung inflations are shown in Fig. 8. Panel A and B represent the effects of ipsilateral (to the BS neuron) vagal nerve transections, while panels C and D represent the effects of contralateral vagal nerve transections. The slow ramp inflation in Fig. 8A, left, produced a graded reduction in &, which was absent after ipsilateral vagal transection. n Fig. 8B, the quiescent periods in the F, record represent periods of central inhibition of BS neuronal activity. The PNG is absent for technical reasons. A period of tonic BS neuronal activity occurred before the onset of the slow Pt ramp. As Pt increased, F, increased as P, passed through the lower Pt range, and a graded reduction in F, occurred above the threshold for inhibition. After ipsilateral vagal nerve transection, BS neuronal activity remained tonic; the inhibitory aspect of the response was eliminated, while a small amount of excitation appeared to remain. For Fig. 8, C and D, the control neuronal responses showed the typical biphasic pattern of graded excitation

7 XPRATORY NURONS OF DOGS: LATRALTY OF RSPONSS R193 2n9 1 j ; NTACT +, + CONTRA. VAGOTOMY NTACT + 111) CONTRA. VAGOTOMY followed by graded inhibition. Transection of the contralateral vagus nerves had no affect on the excitatory responses; however, some reduction in inhibition occurred. These data support the conclusion derived from the electrically induced pulmonary vagal afferent-mediated responses, which suggests that the major source of inhibition is mediated by the ipsilateral vagus nerve. n addition, these data also suggest that excitation is also carried, at least in part, via the ipsilateral vagal fibers of the pulmonary mechanoreceptors. DSCUSSON The main conclusions suggested by the present results are 1) that the inhibition of both type D and type A BS neurons (caudal VRG) by pulmonary mechanoreceptors is mediated predominantly by the ipsilateral vagus nerve, and 2) that the excitation of type D BS neurons (type A exhibited no excitation) is mediated bilaterally (Figs. 4 and 8). With regard to the latter point, contralaterally mediated excitation was clearly illustrated using electrically induced activation of the vagal afferents (Fig. 4, left). The existence of ipsilaterally mediated excitation is suggested by the observation that transection of the ipsilateral vagus nerve eliminated the excitatory component (as well as the inhibitory component) of the neuronal response to lung inflation for the neuron shown in Fig. 8B. n addition, inflation-mediated excitation was not eliminated after contralateral vagal nerve transections (Fig. 8, C and D). With the electrically induced ipsilateral inputs, the observed strong inhibitory effect may have masked the excitatory effect, while the weaker inhibitory effects mediated by contralateral inputs did not obscure the observed excitatory response. The existence of contralaterally mediated inhibition is suggested by the small reductions in the inhibitory component of the neuronal D 1 SC Fig. 8. ffects of unilateral vagotomy on expiratory bulbospinal neuronal responses to augmenting ramp inflations during phase of test cycles in 4 different preparations. Dashed vertical lines separate pre- and postvagotomy records in each case. A and B: effects of ipsilateral vagotomy. C and D: effects of contralatera1 vagotomy. responses after contralateral vagal nerve transection (near the peak of the slow inflations in Fig. 8, C and D). The analysis of poststimulus time histograms of spike data has indicated that ipsilateral and contralateral pulmonary afferents are capable of producing both inhibition and excitation (e.g., Fig. 7B), though with differing degrees of functional effectiveness. A monotonic reduction in BS neuronal discharge frequency was found to occur with the recruitment of more ipsilateral afferent fibers as the stimulus strength was increased from 1T to 3T, indicating that there is a considerable degree of spatial summation involved (Fig. 4). As stated in MTHODS, a strength of 3T appears to activate mainly those fibers with conduction velocities (CVs) >25 m/s, that is, the largest myelinated fibers. Since the maximum level of inhibition was not achieved at 3T (Fig. 4), it appears that smaller-diameter myelinated fibers (CVs >25 m/s) also contribute to the spatial summation. n the case of the contralaterally mediated excitation of type D BS neurons, an increase in stimulus strength >1.5T did not produce a further increase in excitation (Fig. 4, left), suggesting that the recruitment of the smaller-diameter myelinated fibers was either not contributory or may have counterbalanced a further increase in F,. n addition, increases in step input frequency at 3T, especially >2 Hz, produced marked increases in the contralaterally mediated excitation, while the ipsilaterally mediated inhibition was essentially unchanged (Figs. 5 and 6, left). Thus, while the recruitment of afferent fibers with stimulus strengths >1.5T was ineffective, those fibers that were recruited (5 1.5T) were able to increase BS neuronal excitation via temporal summation up to a step frequency of -8 Hz. n the step frequency range of 5 to 2 Hz, temporal summation also increased the level of ipsilaterally mediated inhibition for both type D and

8 R194 XPRATORY NURONS OF DOGS: LATRALTY OF RSPONSS type A BS neurons (Fig. 6). Thus it appears that a combination of both spatial and temporal summation is involved in the mediation of excitation and inhibition, but the strength and frequency ranges are different for each effect. The neuronal responses produced by electrically induced vagal afferents appear to be qualitatively similar to those produced by lung inflation (Fig. 2). The graded reduction in F, with an increase (T-3T) in stimulus strength (Fig. 4) is similar to that for the plots of F, vs. P, for P, > 4-5 mmhg (e.g., Fig. 8 of Ref. 4). Contralateral inputs produced an increase in F, of type D neurons in agreement with that which occurred for P, < 4-5 mmhg (4). The negative activity slopes of the type D decrementing patterns were reduced and became positive as stimulus strength increased, a finding similar to that observed with increases in step P, (4). For type A neurons, increases in stimulus strength and step P, both resulted in small positive increases in the activity slopes. Based on these similarities, it would appear that many of the same afferents that were activated by lung inflation were also activated by electrical vagal stimulation in the range of 1T to 3T. The companion study (4) suggested that the most likely candidate mediating both the excitatory and inhibitory components of the BS neuronal responses to lung inflation appears to be the PSRs. This was based on the fact that the activation threshold Pt of PSRs corresponds to the Pt thresholds observed in the neuronal responses, and that the activation threshold for RAPRs was greater than those required to account for the responses. n addition, the relatively slow adaptation rate of the neuronal responses to step inflation patterns also supports a role for the PSRs. Thus, while a stimulus strength that activates myelinated fibers with conduction velocities ~25 m/s may activate RAPR fibers, their contribution to the reflex control of BS neurons appears to be minimal. f it is assumed that the PSRs are involved, then the literature, based on experiments in cats, suggests some possible neuronal pathways over which the excitatory and inhibitory components of the neuronal reflex may be relayed. Most PSR afferents project to the medial nucleus tractus solitarii (nts), ventrolateral nts, and to an area dorsolateral to the solitary tract (11). ach of the individual PSR afferents appears to project to only one of the subnuclei (6). Pump or P-cells, which are second-order neurons in the PSR pathway (5), are also found in these same three nts regions (1). n addition, 1-p neurons receive monosynaptic inputs from PSRs (l), but because they are silent during the phase, it would appear that they are not involved, which suggests that the P-cells are most likely involved. The P-cells located dorsolateral to the tract have axonal projections to the contralateral caudal medial and commissural subnuclei of the nts, while P-cells in the ventromedial nts do not appear to have contralateral projections (1). The ipsilateral axonal projections of the P-cells are not known. The P-cells located in a discrete region medial to the solitary tract in rats appear to be necessary for the production of the Breuer-Hering reflex (S), which appears to indirectly affect the BS neuronal responses. Based on conduction latencies, it is possible that the P-cells directly project to the caudal BS neurons. For example, the average latency of inhibition, obtained from poststimulus time histograms, was 7.1 ms for the ipsilatera1 vagal inputs and 8.5 ms for the contralateral inputs. The shortest conduction time from the cervical vagus, 8-12 cm caudal to the base of the skull, to the nts in dogs is 4-5 ms (unpublished observations). For ipsilateral inputs, the balance of transmission time to the caudal VRG, 2-3 ms, could be accounted for by two synaptic delays (.5-.6 ms each) and the conduction time of the P cell (~6 mm+ 3 m/s = 2 ms; Ref. 1). Thus the inhibitory component of the BS neuronal response may be directly relayed to the caudal BS neurons via the P-cells. An alternative pathway from the PSRs to the caudal BS neurons may include neurons of the Botzinger complex, some of which are excited ( decrementing) and others ( augmenting) inhibited by lung inflation (12); these effects may be relayed via the P-cells. Caudal BS neurons receive monosynaptic inhibitory inputs from neurons in the ipsilateral Botzinger complex (13); excita- tory inputs have also been demonstrated (7). The latency from the Botzinger complex neuron somatic spikes to the onset of the inhibitory postsynaptic potential in the caudal BS neurons is -2.5 ms (13). The average latencies of vagal afferent-mediated excitation were 25.4 and 24.1 ms for the ipsi- and contralatera1 vagal inputs, respectively. These latencies are more than twice that of those associated with the inhibitory effect. Thus a different path and/or more interneurons may be involved, or a different type of synaptic mechanism, such as modulation of chemodrive to caudal BS neurons, may be operative. CTHs of caudal BS neuronal activity also indicated that, upon the rapid removal of an inhibitory/excitatory producing level of lung inflation, there is a postinhibitory period of excitation suggestive of the slower dynamics of the pulmonary afferent-mediated excitation to these neurons (e.g., Fig. 4 and the positive ramp of the companion paper). The physiological role for BS neuronal laterality may be that it provides a mechanism for balancing right and left lung volumes and/or lobar airflow. Such imbalances may occur because of changes in lateral posture or partial lobar bronchial obstructions. Regional lung volume distribution is determined by the interaction of lung and thoracic cavity shapes, with both structures influenced by gravity (16). Vertical gradients in Pt, regional lung volume, ventilation, and perfusion have been demonstrated for the lateral decubitus position and other positions. Since the axons of BS neurons cross over the midline in the medulla and descend in the contralateral spinal tracts, differences in lung volume or P, on one side relative to those of the opposite lung will result in changes in contralateral muscle drive and regional pleural pressure. For example, since P, levels at end-inspiration are capa- ble of producing inhibition in BS neurons, a retarded rate of deflation in one lung or lobe would reduce the

9 XPRATORY NURONS OF DOGS: LATRALTY OF RSPONSS R195 amount of BS neuronal drive, via net inhibition, to the contralateral motoneurons and muscles and, increase ipsilateral drive, via net excitation, mediated by the contralateral BS neurons, thereby counteracting the imbalance in lung volume. Another example of laterality is found in the PSR-mediated effects on the dorsal respiratory group 1-p neurons, where ipsilateral PSR inputs produce a rapid excitation together with a slow inhibition, whereas contralateral inputs produce only the slow inhibition (3). Because of the crossed projection of these 1-p bulbospinal neurons, a relative overinflation of one lung could lead to a relative increase in drive to the contralateral muscles, acting to balance the level of inflation in both lungs. n summary, the data of the present study suggest that the facilitation of caudal VRG BS neurons is mediated by pulmonary mechanoreceptors arising from both lungs, whereas the inhibitory component of the neuronal response is mainly dependent on receptors in the ipsilatera1 lung. This inhibition would manifest itself in the activity of the contralateral spinal motoneurons. There appears to be a considerable, and greater, degree of convergence of the afferent inputs in the production of inhibition in contrast to excitation. Different latencies, spatial summation, and temporal summation characteristics suggest the possibility that different types of PSRs, different central pathways, and/or different synaptic mechanisms are involved in the biphasic response of the caudal VRG BS neurons to lung inflation. The authors gratefully thank Jack Tomlinson for his technical assistance. This work was supported by the Dept. of Veterans Affairs Medical Research Funds and the Dept. of Anesthesiology of the Medical College of Wisconsin, Milwaukee. J. Bajic and M. Tonkovic-Capin are postgraduate fellows from the Univ. of Zagreb School of Medicine in Split, Croatia. Address for reprint requests:. J. Zuperku, Research Service 151, Zablocki VA Medical Ctr., Milwaukee, W Received 4 April 1991; accepted in final form 1 December RFRNCS 1. Averill, D. B., W.. Cameron, and A. J. Berger. Monosynaptic excitation of dorsal medullary respiratory neurons by slowly adapting pulmonary stretch receptors. J. Neurophysiol. 52: , Bajic, J., F. A. Hopp, and. J. Zuperku. Characterization of medullary neuronal responses to lung inflation patterns (Abstract). Physiologist 31: Al73, Bajic, J.,. J. Zuperku, and F. A. Hopp. Processing of pulmonary afferent input patterns by respiratory 1-p neurons. Am. J. Physiol. 256 (Regulatory ntegrative Comp. Physiol. 25): R379- R393, Bajic, J.,. J. Zuperku, M. Tonkovic-Capin, and F. A. Hopp. xpiratory bulbospinal neurons of dogs.. Control of discharge patterns by pulmonary stretch receptors. Am. J. Physiol. 262 (Regulatory ntegrative Comp. Physiol. 31): R175-R186, Berger, A. J. Dorsal respiratory group neurons in the medulla of cat: spinal projections, responses to lung inflation and superiorlaryngeal nerve stimulation. Brain Res. 135: , Berger, A. J., and D. B. Averill. Projection of single pulmonary stretch receptors to solitary tract region. J. Neurophysiol. 49: , Bongianni, F., M. Corda, G. Fontana, and T. Pantaleo. xpiration-related neurons in the caudal ventral respiratory group of the cat: influences of the activation of Botzinger complex neurons. Brain Res. 526: , Bonham, A. C., and D. R. McCrimmon. Neurones in a discrete region of the nucleus tractus solitarius are required for the Breuer- Hering reflex in rat. J. Physiol. Lond. 427: , Cohen, M.., J. L. Feldman, and D. Sommer. Caudal medullary expiratory neurone and internal intercostal nerve discharges in the cat: effects of lung inflation. J. Physiol. Lond. 368: , Davies, R. O., L. Kubin, and A.. Pack. Pulmonary stretch receptor relay neurones of the cat: location and contralteral medullary projections. J. Physiol. Lond. 383: , Donoghue, S., M. Garcia, D. Jordan, and K. M. Spyer. The brain-stem projections of pulmonary stretch afferent neurones in cats and rabbits. J. Physiol. Lond. 322: , zure, K., and M. Manabe. Decrementing expiratory neurons of the Botzinger complex. xp. Brain Res. 72: , Jiang, C., and J. Lipski. xtensive monosynaptic inhibition of ventral respiratory group neurons by augmenting neurons in the Botzinger complex in the cat. Brain Res. 81: , Kirkwood, P. A., and T. A. Sears. Monosynaptic excitation of thoracic expiratory motoneurones from lateral respiratory neurones in the medulla of the cat. J. Physiol. Lond. 234: 87P-89P, Koepchen, H. P., D. Klussendorf, and U. Philipp. Mechanisms of central transmission of respiratory reflexes. Acta Neurobiol. xp. 33: , Lai-Fook, S. J., and J. R. Rodarte. Pleural pressuuare distribu# tion and its relationship to lung volume and interstitial pressure. J. Appl. Physiol. 7: , Miserocchi, G., and G. Sant Ambrogio. Responses of pulmo nary stretch receptors to static pressure inflations. Respir. Physiol. 21: 77-85, Remmers, J.., and D. Bartlett, Jr. Reflex control of expiratory airflow and duration. J. Appl. Physiol. 42: 8-87, Zuperku,. J., and F. A. Hopp. Control of discharge patterns of medullary respiratory neurons by pulmonary vagal afferent inputs. Am. J. Physiol. 253 (Regulatory ntegrative Comp. Physiol. 22): R89-R82, 1987.