INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

IT IS WIDELY ACCEPTED that the transport of food and drink through the esophagus depends on neuroregulatory mechanisms residing in the lower brain stem (3, 5, 16, 28). The most detailed information on the associated esophagomotor circuitry has come from neuroanatomical studies in the rat. In this species, the peripheral axons of vagal sensory neurons located in the nodose ganglion form specialized terminal structures (multilamellar intraganglionic bodies) in esophageal myenteric ganglia that are postulated to serve as tension receptors (22). The central projections of these afferents terminate organotopically in a discrete part of the nucleus of the solitary tract (NTS) known as the subnucleus centralis (NTS_C) (1, 3). NTS_C neurons emit a dense and topographically discrete projection to the esophagomotor portion of the nucleus ambiguus (3, 9), corresponding to the compact formation (AMB_C), which in turn innervates the striated muscle of the esophageal external muscle tunic (6). By virtue of their dual role as second-order sensory and premotoneurons, NTS_C cells thus form the major internuncial link between afferent and efferent fibers subserving esophageal reflex peristalsis (3-5, 9).

To date neither physiological nor pharmacological aspects of reflex esophageal peristalsis have been investigated in the rat. The purpose of the present study was 1) to characterize vagally mediated motor responses to luminal distension applied at different levels of the tubular gullet, 2) to determine the underlying neural activity patterns at the level of the medulla oblongata, and 3) to assess the role of reafferent feedback. Parts of the present study have been reported in preliminary form (19).

	METHODS

Top Abstract Introduction Methods Results Discussion References

Surgical preparation. Experiments were performed in 94 adult male Sprague-Dawley rats weighing between 350 and 450 g. After anesthesia with intraperitoneal urethan (1.0-1.2 g/kg body wt), a tracheal tube was inserted and the right external jugular vein cannulated for intravenous infusion of saline and drugs. Rectal temperature was maintained between 37.5 and 38°C by means of radiant heat. Spontaneous respiratory activity derived from intratracheal tidal pressure fluctuations was continuously recorded. Before and during neuromuscular blockade, animals were artificially ventilated with oxygen-enriched air by means of a respirator (Harvard). Ventilation was maintained at a rate of 100 cycles/min and a tidal volume of 2.5-3 ml.

Esophageal manometry and distension. Miniature balloon-tipped catheters constructed from PE-90 polyethylene tubing were filled with water and connected via a three-way stopcock to pressure transducers. The length of the catheter from the center of the balloon was marked in centimeters. Pressure signals from each transducer were filtered at a frequency of 0.5-3 Hz (to reduce interference from cardiac and respiratory movements) and displayed on a (Grass) polygraph. Balloons used for recording intraluminal pressure had an outer diameter of 4 mm; balloons used for esophageal distension had a diameter of 6 mm and a volume of 200-215 µl when fully inflated. The balloon for distension was filled with water or room air and connected in parallel to a 500-µl syringe. Esophageal distension was performed manually or by means of a variable-speed infusion pump (Sage Instruments, model 355). The pressure-volume relationship and the diameter-volume relationship of the inflation balloon were determined before insertion into the esophagus. As shown in Fig. 1, a small pressure step occurred at the start of infusion, reflecting resistance to flow in the catheter. Between 5 and 200 µl, the maximal transverse diameter of the balloon expanded from 2.7 to 5.8 mm without further change in pressure, indicating a high compliance of the balloon. When positioned in the upper alimentary tract, the recording balloons were filled with 20-25 µl water and equilibrated with atmospheric pressure for 5-10 s.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Pressure-volume and diameter-volume relationships of inflation balloon as determined outside animal. A: balloon is filled with water by a syringe pump at a fixed rate of 9 µl/s during period indicated by dashed lines; volume of balloon is held constant during interval indicated by solid lines (~160 µl in A1). As volume infused approaches capacity of balloon (~210 µl), pressure rises abruptly (A2). Step increase at start of infusion represents pressure required to overcome resistance to flow through catheter into balloon. B: between 5 and 210 µl, transverse diameter at center of the balloon increases from 2.5 to 6.0 mm, with a negligible change in internal pressure, indicating the high compliance of the balloon.

To characterize the local response patterns, a single inflation/recording balloon was successively positioned at 1-cm intervals along the length of the esophageal body (Fig. 2) and secured at each position with adhesive tape. In some experiments the balloon catheter was tethered to a force transducer (Fig. 3), enabling any aboral movement of the inflated balloon to be recorded. To reduce sensory adaptation, the balloon volume was kept <40 µl during the inter stimulus period. For distension of the esophagus, the balloon was filled by means of the infusion pump at a predetermined rate of 4.5 or 9.0 µl/s. When an active response was elicited, the pump was stopped immediately and inflation was maintained for 20-50 s before the balloon was manually emptied and the pump was reset.

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Segmental esophageal responses to intraluminal balloon inflation. Schematic diagram of rat esophagus (A) indicates position of a single balloon catheter in esophageal body from which responses shown in B are recorded. Positions represent distance from upper incisors to the center of balloon. Balloon inflation is applied at pump rate of 9 µl/s, as indicated by dashed lines under each pressure trace and continued until threshold volume (VT) for active response (marked in µl at arrowheads) or maximum test volume (90-100 µl) is reached. During interval marked by solid lines, inflation is maintained at constant volume until the balloon is manually deflated. A monophasic slow-wave (type I) response is evident at positions 7, 8, and 9, whereas sites 10-12 display repetitive wave (type II) activity. Superimposed small waves seen at levels 8-10 cm are synchronous with respiratory cycle. Note absence of reflex response at both 5- and 12.5-cm sites, even with increased filling of balloon, as well as similarity of VT at levels showing active response. These traces should be compared with those shown in Figs. 1, 10, and 12.

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Propulsive nature of local esophageal reflex elicited by inflation. A: diagram showing experimental setup for simultaneous measurement of intraballoon pressure and aboral movement of catheter. B: when tethered balloon is positioned in the midthoracic portion of the esophagus, rapid step inflation () evokes a monophasic pressure wave followed by a monophasic force wave, indicating that the inflated balloon is propelled distally. Balloon does not return to starting position as indicated by persistence of force after balloon deflation (). C: when balloon is located in the intercrural portion, rapid step inflation produces rhythmic force waves that coincide with rhythmic pressure waves in a phase-locked manner; note, however, initial small lag between force and pressure wave. Pi, intraluminal pressure; Fm, force mechanogram. Superimposed rapid wave activity is of respiratory origin.

The smallest inflation volume that elicited a reproducible esophageal response was defined as the threshold volume (V_T). After a reproducible V_T was established by a ramp inflation at 4.5 µl/s, subsequent esophageal distensions were manually performed as a one-step injection. To examine the inflation volume-response relationship, serial esophageal distensions were performed at 5- or 10-µl incremental steps.

To examine responses of the entire esophageal body to local inflation, the placement of manometric catheters was arranged in two ways. For observation of responses occurring proximal and distal to the distended segment, the inflation balloon was positioned in the thoracic segment (9 cm from the upper incisors), with two additional recording balloons placed 7-8 and 10-11 cm, respectively, from the upper incisors. For investigating proximal responses to distension of the distal esophagus, the inflation balloon was positioned distally (11 cm from upper incisors) and two proximal recording balloons were located in the cervical and thoracic esophagus, respectively. In most experiments, a fourth balloon was placed in the pharynx to monitor pharyngeal responses, if any, during the esophageal distension.

The intraluminal position of the manometric catheters was visually confirmed in 23 rats after an intravenous overdose of urethan and a thoracotomy. In rats between 375-400 g body weight, the total length of esophagus was ~9 cm. The distances from the upper incisors to the pharyngoesophageal junction, first rib, diaphragm, and cardiac were ~4, 7, 11, and 13 cm, respectively. Accordingly, the cervical, thoracic, and distal portions lay between 4 and 7, 7 and 11, and 11 and 13 cm, respectively, from the upper incisors. The rostral two-thirds of the latter segment were defined as the intercrural portion.

Effects of uni- and bilateral vagotomy were examined in animals breathing spontaneously. The upper cervical vagal trunk was gently dissected free and severed with microscissors. In medullary unit recording experiments, the cervical vagi were snared with loose silk threads for subsequent vagotomy at the level of the cricoid cartilage. Postvagotomy reflex esophageal responses were tested up to 40 min.

Extracellular recording and drug application. The animals were mounted in a stereotaxic frame after manometric catheters were secured in their appropriate positions. The caudal roof of the fourth ventricle and surrounding structures of the dorsal medulla were surgically exposed under a dissection microscope. Cerebrospinal fluid was drained continuously with a wick. Extracellular microelectrode recordings were made to determine the locations of rhombencephalic neurons responsive to esophageal distension and to study the esophagomotor reflex responses of these neurons. The electrode consisted of a single-barrel glass micropipette containing a fine carbon fiber (8 µm in diameter). The carbon fiber protruding from the pipette tip was electrically etched to a length of 4-6 µm with chromic acid after the pipette was filled with 4 M NaCl. To stimulate neurons at the recording site, three-barrel micropipettes were used. These consisted of one recording barrel containing a carbon fiber and 4 M NaCl and two barrels filled with glutamate (0.5 M) and ACh (0.2-0.5 M). Both substances were pressure ejected in pulses of low amplitude (10-15 lb/in.²) and duration (5-10 ms) so as to minimize displacement of the recording barrel. Under microscopic control, the micropipette was stereotaxically inserted into the medullary tissue with the rostral margin of the area postrema (AP) as a reference landmark. Based on previous work (1, 4, 6, 13), explorations of esophagomotor-related neurons concentrated on two medullary regions: namely the medial portion of the intermediate NTS and the rostral portion of the ambiguus (AMB) complex. Electrode penetrations were made 100-150 µm apart in both the mediolateral and rostrocaudal planes. The electrode was advanced in 50-µm dorsoventral steps with esophageal distension applied between each step. Unit discharges were monitored on an oscilloscope and discriminated by means of a spike trigger (Neurolog, NL200 Digitimer). Discharge frequency was displayed on a Grass polygraph through a ratemeter along with intraesophageal pressure signals. In nine experiments, 4% lucifer yellow (Sigma) in 4 M NaCl was electrophoretically ejected at the recording site by means of negative direct-current pulses (500 ms, 2 Hz, 5-10 nA, 5-10 min) after recordings were completed.

Histological procedures. Fifteen minutes after the dye ejection, the animals were perfused with 150-250 ml saline, followed by 250 ml of fixative (pH 7.2 phosphate-buffered 4% paraformaldehyde). The medulla oblongata was removed, stored overnight in the same fixative at room temperature, and sectioned on a vibratome. Transverse sections (75 or 100 µm thick) containing the recording site were mounted on slides in 0.1 M phosphate buffer and viewed under a fluorescence microscope. The recording sites were identified by their yellow fluorescence under dark-field ultraviolet illumination and by the electrolytic lesion. Maps of recording sites were prepared based on previous neuroanatomic studies (1, 6, 13).

Data analysis. To ensure reproducibility of esophageal reflex responses, the interval between stimulations in each protocol was kept between 2 and 3 min, and vagotomy or drug applications were performed after at least four consistent control responses were obtained in each esophageal segment. The V_T values were determined by recording both syringe plunger displacement and duration of the pump on cycle. The mean amplitude of rhythmic responses was calculated by averaging the amplitudes of pressure waves during the first 10 s of inflation. Because the multiunit discharges of medullary neurons could not be reliably discriminated in terms of single units, peak evoked firing rates were used as an index of neuronal responsiveness. Graphs were constructed by means of software packages (SigmaStat and SigmaPlot, Jandel Scientific). Data were expressed as means ± SE; the significance of differences was examined with Student's paired t-test at P < 0.05.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Local response patterns. Motor responses occurring at the level of balloon inflation showed consistent regional variations in both pattern and relative threshold. Figure 2 shows a representative experiment that was replicated with similar results in seven rats. In the upper to midcervical portion, contractions were not reliably evoked at balloon volumes of up to 150 µl. An active monophasic pressure wave first appeared in the lower cervical portion at inflation volumes of 60 to 80 µl (n = 4 cases) and became more robust and reproducible as the balloon entered the upper thoracic portion. At mid- to low-thoracic levels, either single monophasic (n = 5) or repetitive (n = 2) pressure waves were observed, whereas the intercrural portion typically responded with robust repetitive activity (n = 7). For convenience, inflation-induced single wave activity is hereafter referred to as type I and repetitive rhythmic activity as type II. A local response to distension was absent when the inflation balloon was positioned at or near the gastroesophageal junction (12.5 or 13 cm level) (Fig. 2), except in one case where a variable type I response was obtained (not illustrated). The V_T, as determined for type II responses, varied among animals (62 ± 6.5 µl; range 40-70 µl; n = 14); however, in a given preparation, V_T and amplitude of the evoked pressure wave(s) remained steady for 2-3 h. V_T values for type I responses fell in the same range (see Fig. 2), but were not analyzed statistically.

The cervical and thoracic type I response typically showed a faster rate of rise than rate of decay. Total wave duration ranged from 4 to 9 s (5.3 ± 0.3 s; n = 21 trials in 7 animals). The intercrural type II response had a highly regular wave shape and rhythm in the range of 0.5-0.8 Hz (0.65 ± 0.1 Hz; n = 55). Individual waves lasted 0.99 ± 0.09 s (n = 50 waves from 10 cases), and their mean amplitude increased with the inflation volume from 3.0 ± 0.32 kPa at V_T to a maximum of 6.6 ± 0.5 kPa at 75-80 µl (n = 14). When the inflation balloon was positioned in the thoracic esophagus and left mobile, inflation resulted in a visible catheter movement, recorded as an aborad pulling force trailing the reflex-evoked pressure wave (Fig. 3). With the balloon held in the intercrural portion, distension-evoked rhythmic force and pressure waves were synchronous, except for the initial wave (Fig. 3C).

Coordinated responses. As recorded by a three-balloon assembly (Fig. 4), the type I response in the distended segment was accompanied by a small sustained rise in pressure in the adjacent oral segment (n = 10 rats). In the segment aboral to the distension, intraluminal pressure showed a small transient drop in two (Fig. 4), a small phasic rise in five, or remained unchanged in three of ten rats tested. Except for one case where the monophasic pressure wave propagated from the inflated (thoracic) segment to the distal esophagus (not shown), the "inflation responses" recorded above, in or below the stimulated segment, occurred in a synchronous (nonpropagated) manner. In most cases (n = 7 of 10), the pressure wave in the oral segment outlasted that in the inflated segment. On rapid, but not slow, deflation, a phasic pressure wave invariably occurred in the aboral segment (Fig. 4). In the thoracic portion, this "deflation response" was highly reproducible in virtually all preparations tested (10-15 trials per animal, n = 10). In the cervical portion, however, a deflation response could not be evoked in a reproducible manner. The deflation response typically consisted of a monophasic pressure wave, but in the distal esophagus it occasionally showed a rhythmic pattern.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Coordinated esophageal reflex response. Three intraesophageal balloons are positioned at the depths indicated. Sustained inflation () of the middle balloon evokes a high-amplitude pressure wave in the distended segment coinciding with a low-amplitude pressure rise in the proximal segment and a phasic pressure drop in the distal segment. On deflation (), a monophasic pressure wave is elicited in the distal segment. Rate and depth of respiration (R) remain unaltered during inflation. Esoph, esophagus; Diaph, diaphragm.

When inflation was applied via the distal balloon in the intercrural portion (Fig. 5), an attenuated type II response was recorded 2 cm proximal to the inflation balloon, with the cervical esophagus remaining quiescent. As seen in high-speed recordings (Fig. 5B), pressure waves in the thoracic segment led those in the intercrural segment in a phase-locked manner and had a sinusoidal shape, unlike the saw-tooth-like configuration of intercrural pressure waves. Depth and rate of respiration were unchanged during the type II response (Fig. 5).

View larger version (17K): [in this window] [in a new window]   Fig. 5.   Coordinated responses to distal esophageal inflation. Distal balloon is filled at a pump rate of 4.5 µl/s (dashed lines) and kept inflated for a period indicated by the solid lines. A: evoked rhythmic pressure waves are present in the lower thoracic (9 cm) and in the distended distal segment, but not at the upper thoracic (7 cm) level. B: trace recorded at a faster speed shows that each pressure wave in the lower thoracic segment precedes the pressure wave in the distal segment in a phase-locked manner. Note different shape of the pressure waves recorded at the 2 levels.

Localization of brain stem neurons responding to esophageal distension. Extracellular recordings were obtained from esophageal distension-responsive neurons in the dorsal and ventral medulla oblongata (Fig. 6A) The dorsal region extended between 100 µm caudal and 400 µm rostral to the cranial edge of the AP, 600 and 800 µm lateral to the midline, and 350 and 550 µm below the dorsal medullary surface. The ventral region was located between 1,000 and 1,800 µm rostral to the rostral edge of the AP, 1,900 and 2,200 µm lateral to the midline, and 2,000 and 2,300 µm below the dorsal medullary surface. Fluorescent dye marks and electrolytic lesions at the recording sites for 14 type I and II recordings were histologically verified in nine animals (Fig. 6B). The dorsal sites (n = 7) were located in the medial region of the intermediate NTS, coextensive with the NTS_C. In the ventral medulla, four sites were located within and one site 50-80 µm dorsal to the boundary of the AMB_C; in addition, two recording tracks ended within 100 µm dorsal to the AMB_C.

View larger version (27K): [in this window] [in a new window]   Fig. 6.   Distribution of medullary loci where unit discharges were evoked by esophageal distension. A: dorsal view of electrode penetrations yielding responsive () and nonresponsive (small dots) loci. B: as projected onto the transverse planes spanning rostrocaudal levels B1 and B2, responsive loci (n = 14) are clustered in a dorsocaudal region coextensive with the subnucleus centralis of the nucleus of the solitary tract (NTSC) and a rostroventral region comprising the compact formation of the nucleus ambiguus (AMBC). Sites marked by fluorescent dye are represented by ;  denote response loci that correspond to ventralmost point of pipette track. Distances on x- and y-axes are shown with reference to midline and cranial edge of the area postrema. R, rostral; M, medial; IV, fourth ventricle; VII, facial nucleus; ST, solitary tract; XII, hypoglossal nucleus; DMV, dorsal motor nucleus of vagus.

Neural activity patterns. At rest, tonic unit discharges (0.5-3 Hz) were observed in the NTS_C region (n = 19), whereas the majority of units in the AMB_C region (n = 16 of 22 recordings) were silent. Esophageal distension evoked burst discharges in both regions. In the NTS_C, these could be elicited by small-volume inflations at which esophagomotor reflex responses remained undetectable. In contrast, AMB_C activity occurred only in conjunction with overt motor responses. Although neither NTS_C- nor AMB_C-evoked activity showed a phase relationship with breathing, its pattern and timing correlated with that of evoked pressure wave activity. Thus inflation of the midthoracic esophagus produced single burst discharges in both the NTS_C (n = 9; Fig. 7A) and AMB_C (n = 8; Fig. 7B) that typically started with a high-frequency barrage and decayed gradually (n = 12 of 17 recordings). The five remaining units fired for a 2- to 3-s period and then ceased abruptly. In contrast, distension of the intercrural esophagus produced burst discharges in both the NTS_C (Fig. 7C; n = 10) and the AMB_C (Fig. 7D; n = 14) that were rhythmic and in phase with type II activity. Type II responses (Fig. 8) showed a highly regular (0.5-0.8 Hz) burst-pause pattern in both the NTS_C and the AMB_C. Peak frequencies (F_Peak) of the burst discharge in the NTS_C ranged from 80 to 165 Hz (mean 125 ± 17 Hz, n = 19). In the AMB_C, F_Peak values ranged from 50 to 130 Hz (mean 92 ± 16 Hz, n = 22 samples from 16 animals).

View larger version (24K): [in this window] [in a new window]   Fig. 7.   Neural discharge patterns in brain stem recorded during distension-evoked reflex response of esophagus. Each sample is from a different rat. A nonrhythmic discharge occurs in the NTSC (A) and AMBC (B) on inflation via balloon positioned 9 cm from upper incisors (indicated by thick bar under pressure traces). Nonrhythmic discharges lead monophasic pressure waves recorded about 2 cm proximal to level of inflation in the upper thoracic esophagus (UTE). Vertical calibration bars in A and B are identical. Rhythmic discharges seen in the NTSC (C) and AMBC (D) during distension of the distal esophagus (DE) lead phase-locked pressure waves in distended segment. Pressure waves in C and D are attenuated due to recording by an air-filled balloon. , Inflation; , Deflation.

View larger version (31K): [in this window] [in a new window]   Fig. 8.   Analysis of reflex-evoked type II response. A: representative example of burst activity in the AMBC evoked by distal esophageal distension. Horizontal dotted line indicates level (15-20 Hz) at which burst-pause phase length is determined (n = 52 burst-pause cycles from 9 animals). B: timing of repetitive burst activity demonstrates stereotyped pattern as evidenced by burst duration (TB, 0.90 ± 0.06 s), mean interval between bursts (TI, 0.65 ± 0.10 s), period of bursting cycle (TC, 1.55 ± 0.11 s), or burst frequency (1/TC, 0.64 ± 0.04 Hz).

Although our recording and stimulation techniques did not permit exact determination of the phase relationship between unit discharges and the onset of esophageal activity, the data clearly showed that burst discharges in the NTS_C or AMB_C preceded type I and II activity at a fixed latency. AMB_C burst discharges started between 150 and 200 ms before the onset of type II waves in the thoracic segment rostral to the level of inflation, whereas the corresponding latency in the inflated distal segment itself was at least 300 ms (Fig. 9A). However, glutamate (0.2 M, 20-50 pl) ejected at the recording site in the AMB_C elicited unit discharges that were followed after a latent period 50 ms (Fig. 9B) by a synchronized nonpropulsive esophageal pressure wave.

View larger version (19K): [in this window] [in a new window]   Fig. 9.   Phase relationship between type II esophagomotor burst discharges and esophageal pressure wave activity. A: rapid distension ( ) of the DE evokes AMBC rhythmic discharges that lead coupled pressure waves propagating from lower thoracic esophagus (LTE) to DE; note lag (~0.2 s) between onset of each burst and thoracic pressure wave. B: glutamate (0.2 M, ~20 pl; ) pulse ejected at the same recording site in the AMBC elicits a volley of unit discharges that leads contraction in thoracic and distal esophagus with a latency 50 ms. Pressure waves in DE are attenuated because of recording via air-inflated balloon.

Effect of vagotomy. After unilateral (left or right) vagotomy, the type I response in the thoracic esophagus disappeared (n = 2; Fig. 10A). Unilateral vagotomy impaired the type II response in the distal esophagus, as evidenced by an increase in the V_T and a reduction in both the number and amplitude of the rhythmic pressure waves (n = 7; Fig. 10B). Bilateral vagotomy abolished the type II response (n = 7, data not shown).

View larger version (14K): [in this window] [in a new window]   Fig. 10.   Effect of acute unilateral vagotomy on the reflex esophageal response. Traces shown are taken from 2 separate experiments in which balloon distension was applied to the thoracic esophagus (18 cm from incisors; A) and the intercrural esophagus (12 cm from incisors; B). Type I reflex response is abolished in the thoracic segment and not restored by increasing the inflation volume (indicated below each pressure trace). Rhythmic reflex activity in the distal segment is markedly attenuated. Dashed portion of line marks period during which balloon is being inflated, solid portion represents sustained inflation at volume indicated.

The laterality of vagal afferent input to medullary esophageal neurons was determined in six preparations. After contralateral vagotomy, the distension-evoked unit discharges in the NTS_C region persisted (n = 4; Fig. 11A). Ipsilateral vagotomy eliminated the evoked unit discharges in both the NTS_C (n = 6; Fig. 11B) and AMB_C regions, (n = 2; Fig. 11C).

View larger version (23K): [in this window] [in a new window]   Fig. 11.   Effects of uni- and bilateral vagotomy on esophageal distension-evoked neural discharges in medulla oblongata. Unit discharges in the NTSC region evoked by distal esophageal distension (indicated by thick bar) persist after contralateral cervical vagotomy (A) but disappear after ipsilateral cervical vagotomy (B). Midthoracic distension-evoked unit discharges recorded in the AMBC of another rat are also abolished by ipsilateral vagotomy (C).

Effect of curarization. Intravenous D-tubocurarine abolished distension-evoked pressure waves, including the proximal type I (n = 4; Fig. 12A1), the distal type II (n = 6; Fig. 12A2), and the "off" response (n = 2; Fig. 12B). During paralysis, a change in pressure-volume relationship was evident in the upper thoracic esophagus (n = 4), suggestive of a decreased compliance of the esophageal wall.

View larger version (18K): [in this window] [in a new window]   Fig. 12.   Inhibition of reflex esophageal responses by curarization. A: local reflex responses evoked by ramp inflation in upper thoracic (A1) and intercrural level (A2) are reversibly abolished by D-tubocurarine (dTc, 0.15 µmol/kg iv) and not restored by increasing inflation volume (indicated under each pressure trace). Biphasic and steeper rise in intraballoon pressure in the upper thoracic esophagus during motor paralysis (with inflation volume <120 µl) indicates altered resistance of esophageal wall to distension. B: rapid inflation and deflation (marked by solid bar) in the thoracic segment produce a coordinated reflex response in segment 2-3 cm aboral to the inflation balloon. Neuromuscular block with intravenous D-tubocurarine (0.15 µmol/kg) reversibly eliminates all reflex activity.

The effects of curarization on NTS_C (n = 5) or AMB_C (n = 6) unit discharges and esophageal contractions were observed in 11 animals. Within 30 s after intravenous administration of D-tubocurarine (0.075 µmol/kg), the amplitude of reflex esophageal pressure waves declined to undetectable levels. At the same time, the F_Peak of type I unit discharges in the NTS_C (n = 3, not shown) and AMB_C (n = 3; Fig. 13A) was reduced. Likewise, type II unit discharges in the NTS_C (n = 3; Fig. 13B) and AMB_C (n = 5; Fig. 13C) were markedly altered. In the former region, the F_Peak was reduced by ~10%; in the latter, it decreased to 70.8 ± 9.7% of the precurare control (n = 6 trials in 3 rats; Fig. 13D). Furthermore, the burst-pause pattern of the discharges appeared less distinct or was obliterated. During recovery of the rhythmic reflex pressure waves, the burst-pause pattern reappeared and firing rates regained control levels.

View larger version (20K): [in this window] [in a new window]   Fig. 13.   Effects of curarization on esophageal distension-evoked activity at premotor and motoneuronal levels. After intravenous D-tubocurarine (0.075-0.15 µmol/kg iv), unit discharge in the AMBC evoked by inflation of the thoracic esophagus (TE) is reduced and delayed (A). Rhythmic (type II) burst activity in the NTSC (B) and AMBC (C) evoked by distension of the DE in 2 other rats shows altered periodicity and decrease in peak frequency during curarization. Note greater reduction of type II discharge peak frequency in AMBC compared with NTSC (D). *P  0.05 vs. control (Cont).

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The results presented here place into functional context an extensive body of neuroanatomical literature describing vagovagal reflex circuits of the rat esophagus (1, 3, 5, 6, 9, 22). The basic elements of this circuitry comprise motoneurons of the rostral AMB innervating the striated muscle tunica propria (3, 5, 6), premotor (internuncial) neurons of the NTS_C (1, 3, 4, 9), and primary vagal afferents from intramural mechanoreceptors of the tunica propria (1, 22). Clearly, the rat esophagus possesses reflex capabilities no less elaborate than those described in other mammals possessing a mixed striated/smooth muscle gullet. In particular, our findings provide further insights into 1) the segmental organization of esophageal motility patterns, 2) its neural correlates at the brain stem level, 3) the half-center connectivity of vagal afferent input, and 4) the role of reafferent feedback in peristaltic rhythmogenesis.

Segmental organization. The proximodistal variation in distension-induced reflex responses was an unexpected finding, given the histological uniformity of the rat esophagus (5, 12). Although both nonrhythmic and rhythmic reflex patterns have been described in various other laboratory animals (7, 8, 10, 24) and in the human (25, 27), their segmental organization has received little attention. In the rat, type II rhythmic reflex activity was more or less confined to the intercrural (diaphragmatic) portion of the esophagus. Although its rhythm lay in a narrow frequency range potentially overlapping with that of breathing, it had no relationship to diaphragmatic activity. Distension of the esophagus in the cat (17) and dog (15, 23) has been reported to selectively inhibit phasic inspiratory activity in the crural portion of the diaphragm. If the same mechanism existed in the rat, it would preclude diaphragmatic contractions as the source of the rhythmic esophageal pressure waves. More to the point, the type II rhythmic response was markedly reduced by unilateral vagotomy with only minor changes in respiratory activity.

Both type I and type II local reflex responses were of a propulsive nature and, in the latter case, were observed to start above the distended portion of the esophagus (see below). Because of the greater degree of stretch, pressure waves generated in the inflated segment were larger than those in the nondistended proximal segment. In view of the robustness of type I and type II responses seen in all animals at thoracic and distal levels, the apparent absence of responses in the upper cervical esophagus in most cases is puzzling. A possible, but not convincing, explanation would be surgical damage to esophageal nerve fibers running in the ramus externus of the superior laryngeal nerve, which supply the rostralmost esophagus (6).

Esophageal reflex activity also showed a segmental variation in terms of the different pressure wave forms seen in the thoracic and the intercrural segments. These regional differences in motility pattern suggest that the central esophagomotor network consists of segmentally organized subunits. This concept agrees with previous studies showing that 1) both vagal afferent input from the esophagus to the NTS_C and output from AMB_C motoneurons to the esophageal tunica muscularis propria are viscerotopically organized (1, 3, 6) and 2) microphoresis of glutamate receptor or muscarinic cholinoceptor agonists in the NTS_C evokes esophagomotor responses that involve discrete segments and show rhythmic or nonrhythmic patterns (4, 13).

The coordinated "inflation" and "deflation" responses of the rat esophagus invite comparison with "on" and "off" responses of the smooth muscle esophagus of the human (25), the opossum (8, 24), and the cat (7). As defined in the present study, the inflation response in the rat esophagus was dependent on an intact vagal innervation. The in vivo on response of the opossum smooth muscle esophagus likewise appears to be vagally mediated (24); however, it is reportedly nonpropulsive like that of the human (25) and has an atropine-sensitivity that decreases in the aboral direction (24). At variance with these two species, the rat exhibited a propulsive inflation response, although this was evident only when the inflation balloon was left free to move. Arguably, the technique used here may produce a similar result in the other species.

The vagally mediated deflation response in the rat resembled the off response of other species, particularly that described in the feline striated muscle esophagus (7). The present study did not determine whether the deflation response in the rat was propulsive; however, off responses in the smooth muscle esophagus in other species are typically propulsive (7, 26, 28). Because the off response of the opossum esophagus can be elicited in vitro (8, 26), it is thought to be effected by intramural neurons. The underlying mechanism may involve a rebound excitation of the muscle after inhibition by nitrergic myenteric neurons (26). Whereas some workers equate the off response with secondary peristalsis (7, 24, 25), the analogous deflation response described here may represent a component of centrally coordinated reflex peristalsis that operates at high velocities of bolus transport.

Because curarization effectively abolished all contractile responses to luminal distension, a common vagal efferent pathway terminating at motor endplates of the tunica muscularis propria can be inferred. However, further studies are required to determine if vagal (parasympathetic) efferents to nitrergic intramural ganglia mediating inhibition of smooth muscle in the tunica muscularis mucosae may be responsible for the relaxation observed at levels distal to the inflated segment. Inhibition of the same pathway could conceivably account for the decrease in esophageal compliance observed during curarization (cf. Fig. 12A).

Brain stem neural correlates. Previous electrophysiological work in the rat (18) has mapped the location in the medulla oblongata of deglutitive unit discharges occurring during reflex swallows elicited by stimulation of the superior laryngeal nerve. The dorsal and ventral groups of extracellularly recorded unit activity described in that report probably related mostly to the buccopharyngeal stage. Units active during the esophageal stage of swallowing were not explicitly described, except for a few elements that discharged up to 140 ms after the onset of buccopharyngeal activity.

The present observations bear on the functional organization of medullary interneurons controlling reflex esophageal peristalsis in the rat. Unit burst discharges induced by distension of different esophageal segments were localized to two circumscribed medullary regions coextensive with the NTS_C and AMB_C. In light of neuroanatomic evidence (1, 3, 6, 9), these two structures can be considered to form the medullary throughput of the esophageal reflex arc. As expected, the neuronal discharges recorded in both regions and the esophagomotor output elicited showed a fixed phase relationship. Because of technical limitations, the latency between the evoked neuronal discharges and the esophageal contractions could be determined only by approximation. Given the neuroanatomical evidence, however, it seems reasonable to regard NTS_C neuronal burst discharges as premotor and AMB_C bursts as motor output. Esophageal premotor and motoneurons within the NTS_C and AMB_C, respectively, have a crude organotopic distribution with considerable rostrocaudal overlap (1, 3, 6). It is thus noteworthy that distension of different segments of the esophagus produced distinct response patterns.

Distension of the intercrural esophagus evoked rhythmic (type II) contractions in the lower thoracic segment that propagated to the inflated segment itself. Indeed, evoked AMB_C rhythmic discharges were observed to lead the pressure wave activity in the inflated portion by ~0.3 s; however, in the segment just proximal to the level of inflation the lead time between motoneuronal burst discharges and pressure waves was reduced by a factor of at least two. This phenomenon cannot be a recording artifact, because a pulse of glutamate delivered at the same recording site in the AMB_C motoneuronal pool produced volleys of unit discharges that preceded an esophageal contraction at a latency of 50 ms. One may therefore conclude that segmental distension activates NTS_C subcircuits that control motoneurons innervating esophageal segments rostral to the level of inflation. This functional arrangement would facilitate aborad bolus propulsion toward the stomach.

Vagal sensory fibers innervating the esophagus in the rat (2), cat (20), and sheep (10) display spontaneous activity, but electromyographic activity is absent in the resting esophagus (14, 21). As shown here, the premotoneurons in the NTS_C fired tonically at rest, whereas the esophageal premotoneurons in the AMB_C were silent. One may therefore surmise that 1) esophageal premotoneurons receive tonic excitatory inputs originating from peripheral sensory receptors at rest and 2) esophageal motoneurons integrate premotor inputs and fire only when the discharge frequency of the premotoneurons exceeds a threshold level.

Half-center organization of afferent input. The esophagus is controlled by bilateral inter- and motoneuronal networks that are located in each one-half of the medulla oblongata (1, 3, 5). In sheep, a small electrolytic lesion in the medial region of the NTS selectively eliminates the esophageal stage of swallowing induced by ipsilateral, but not contralateral, stimulation of the superior laryngeal nerve (SLN) (16). This observation suggests that excitatory input to esophageal interneurons derives chiefly from ipsilateral buccopharyngeal stage (deglutitive) interneurons activated by SLN afferents or from ipsilateral esophageal afferents in the SLN. In the present study, NTS_C interneuronal discharges induced by distension of the esophagus were abolished by ipsilateral, but not contralateral, vagotomy. This observation not only suggests that afferents from the esophagus in each vagal trunk project mainly to the ipsilateral NTS_C but also agrees with anterograde tracing experiments (unpublished data) showing that the central projection from the nodose ganglion to the NTS_C region is uncrossed. Furthermore, consistent with neuroanatomical and functional data (5, 9), the esophageal motoneurons likewise would appear to receive excitatory input only from ipsilateral interneurons, because unilateral vagotomy abolished distension-evoked unit discharges in the ipsilateral AMB_C.

Role of reafferent input. In the rat (2), opossum (29), cat (20), and sheep (10), vagal afferent fibers fire during both distension of the esophagus and the esophageal stage of swallowing. Rhythmic vagal afferent discharges associated with rhythmic esophageal contractions in the sheep are elicited by sustained distension of the esophagus (10), suggesting that esophageal mechanoreceptive impulses in response to stretch (feedforward) and contraction (feedback) are conveyed centrally by two different sets of vagal afferent fibers. In the present study, sustained distension of the intercrural portion of the rat esophagus produced rhythmic burst discharges in the NTS_C neurons. As NTS_C neurons may be considered both premotor interneurons and second-order sensory neurons (5), their rhythmic activity could represent either an endogenous oscillator response to tonic input conveyed by feedforward peripheral afferent fibers or a response to phasic afferent feedback.

Because motor paralysis does not disrupt the sequential discharge of the cranial nerve motoneurons during reflex-evoked swallowing, it has been hypothesized that a medullary interneuronal network can program the entire motor sequence of esophageal peristalsis without peripheral feedback (16, 28). As regards rhythmic reflex esophageal activity in the rat, peripheral feedback appears to play a critical part, because marked alteration of reflex activity of both NTS_C and AMB_C neurons was evident during curarization. As curarization did not reduce the apparent resistance of the esophageal wall to stretch, the observed attenuation of type II neuronal discharges cannot be attributed to slackening of the muscle tunic, but instead implies a role of reafferent signals in the reinforcement of esophagomotor output. Moreover, the disruption of the regular burst-pause pattern of type II activity strongly suggests that reafferent feedback, although not absolutely required for esophagomotor rhythmogenesis, is essential for shaping the final motor pattern and for motoneuronal recruitment. The limitations inherent in extracellular multiunit recording preclude a definitive answer to the question of whether the rhythmicity of the evoked type II burst activity was abolished altogether or no longer detectable due to loss of synchronization. Single cell studies in NTS_C slice preparations (Lu and Bieger, unpublished observations) show that NTS_C neurons produce oscillations with a pattern similar to that of type II reflex activity during stimulation of glutamate or acetylcholine receptors, suggesting that NTS_C neurons can generate esophagomotor rhythm in the absence of phasic afferent input.

In summary, the rat esophagus exhibits segmentally organized reflex responses to local distension with different segments showing distinct response patterns, suggesting the existence of segmental pattern generator subcircuits. Neurons of the NTS_C and AMB_C exhibit activity patterns correlated with distension-evoked reflex contractions. Esophagomotor reflex responsiveness and rhythm generation depend on vagal reafferent feedback.

Perspectives

Our findings suggest conceptual parallels to lower vertebrate locomotor pattern generation, which is thought to be governed by a central pattern generator made up of multiple neuronal circuits. For instance, in the lamprey, each segmental circuit, comprising about three spinal segments, is capable of producing fictive locomotor activity with intersegmental propulsion (11). Similarly, in the rat gullet, different segments seem to be controlled by "unit burst generators," however, the latter exhibit regionally distinct functional properties. Intersegmental coordination during swallow-, as opposed to bolus-, induced peristalsis would thus require reconfiguration of individual subcircuits, rather than simple serial coupling as envisaged by a chain-reflex model.

Teleologically, the regionally localized rhythmic reflex activity in the distal esophagus may represent a safety mechanism designed to facilitate peristaltic bolus transport through a mechanically restricted diaphragmatic hiatus. Our curarization experiments did not provide a clear-cut answer as to whether this esophagomotor rhythm was centrally generated or entrained by phasic feedback from vagal mechanoreceptors. Nonetheless, they underline the importance of reafferent input in central esophagomotor rhythmogenesis.