Adelson, David W., Jen Yu Wei, and Lawrence Kruger. Warm-sensitive afferent splanchnic C-fiber units in vitro. J. Neurophysiol. 77: 2989–3002, 1997. Receptive fields of 41 slowly conducting sensory fibers were located using a thermal (warm) search stimulus in an in vitro splanchnic nerve-mesentery preparation. Warm-sensitive receptive fields were punctate and were densest in the region surrounding the prevertebral ganglia, an area with prominent deposits of brown adipose tissue, where the abdominal aorta branches into the major trunks supplying the abdominal viscera. Impulse activity was recorded while applying a warm stimulus to identified receptive fields (RFs). The warm stimulus consisted of a warming ramp (10–15°C in 1–2 s to a 42–49°C peak temperature) followed by a 10- to 30-s period during which the RF was maintained at this peak temperature (plateau phase). Eighty percent (33/41) of warm-sensitive units responded to warming with discharge comprising both a phasic and a tonic component (slowly adapting warm-sensitive, or SA-W, units). The remainder (8/41) responded with only phasic discharge (rapidly adapting warm-sensitive, or RA-W, units). Units' adaptation characteristics were consistent from trial to trial and when applying stimuli from different positions. Fifty percent of SA-W units (8/16) and 17% of RA-W units (1/6) were activated by transient exposure to 9–90 nM bradykinin (BK). Twenty-seven percent (9/33) of SA-W units and 12%(1/8) of RA-W units were activated by probing their RF with von Frey hairs with bending forces <10 mN (∼1 g equivalent mass). An additional five SA-W units tested were activated by strong mechanical stimuli (compression with a metal probe or firm stretching). No BK-responsive warm-sensitive units were activated by von Frey probing <10 mN, but two (both SA-W) responded to strong mechanical stimuli. In six SA-W units and one RA-W unit, the number of impulses evoked by warming ∼5 min after exposure to BK was >2 SD greater than the mean pre-BK response, indicating sensitization. This sensitization was transient, the response to warming returning to within one standard deviation of the pretrial mean or less over the course of the next 5–10 min. Changes in background activity, mechanical sensitivity, BK sensitivity, and BK-induced sensitization were noted in various splanchnic units over the course of prolonged observations, suggesting that these indices may not reliably distinguish unit type, but instead may indicate the functional state of the sense organ. Splanchnic neurons responsive to the intense warming used in the present in vitro experiments may participate in the cardiovascular responses observed in vivo in heat-stressed rats. The dense distribution of warm-receptive fields in the vicinity of the celiac-superior mesenteric ganglionic complex is consistent with the localization of splanchnic thermosensitive units previously noted in vivo in the rabbit.
Understanding of visceral sensory innervation is limited relative to the innervation of more accessible tissues, such as the skin, for a number of reasons. In addition to technical difficulties in recording single-unit activity in visceral nerves, investigations into visceral sensory mechanisms are impeded by the ill-defined nature, both in quality and in spatial distribution, of visceral sensations, a condition that makes it difficult to determine the proximate causes and initiating sites of centripetal impulse activity. Further, a great deal of afferent activity in visceral nerves may not evoke conscious sensation, rendering the function of such afferent activity obscure.
Much neurophysiological work on abdominal primary afferent neurons has focused on the responses to mechanical stimuli of sensory units in hollow organs (Berkley et al. 1988; Blumberg et al. 1983; Cervero and Sann 1989; Häbler et al. 1990) and/or serosal tissues (Bessou and Perl 1966; Cervero and Sharkey 1988; Floyd and Morrison 1974; Morrison 1973). Even in studies focusing on ischemia- or chemosensitivity of visceral afferent neurons, mechanical search stimuli (e.g., poke, pinch, stretch) have been used to identify receptive fields (Haupt et al. 1983; Jänig and Koltzenburg 1990; Longhurst and Dittman 1987; Longhurst et al. 1984).
In addition to mechano- and chemosensitive units, visceral innervation includes afferent neurons responsive to warming. Rawson and Quick (1972) demonstrated that behavioral responses (panting, sweating) to intraabdominal warming in the sheep were abolished by sectioning the splanchnic nerves, and Reidel (1976) recorded warming-induced discharge in 13 splanchnic units in the rabbit in vivo. These units were divided into two groups, one with static and dynamic maxima centered around 40°C, the other around 46°C, and all were claimed to lack mechanosensitivity. More recently it has been shown that 45°C saline applied to abdominal mucosal or serosal tissues elicits a viscerocirculatory reflex dependent on intact splanchnic primary afferent C fibers (Rozsa et al. 1988), apparently via both centripetal impulse activity in and peripheral neuropeptide release from these neurons. This circulatory thermoreflex (Rozsa et al. 1988) resembles the changes seen in heat-stressed rats when core temperatures approach or exceed 42°C (Kregel et al. 1988). Similar cardiovascular pseudoaffective responses previously have been used to define noxious stimulus intensities in the viscera (Cervero 1982).
Wei (1991) has developed an in vitro splanchnic nerve-mesentery preparation that allows investigation of the properties of visceral sensory neurons in a controlled environment. To locate receptive fields of units unresponsive to moderate mechanical perturbations (e.g., chemosensitive units of undetermined specificity or units that require prior sensitization to be excited by mechanical stimuli) in this relatively delicate tissue without repeatedly applying potentially damaging strong mechanical stimuli, another type of search stimulus was desired. Search stimuli capable of activating both efferent and afferent fibers, such as KCl or electrical stimulation were deemed inefficient because an estimated 80% of the axons in the splanchnic nerve are sympathetic efferents (Cervero 1994; Jänig and Morrison 1986). Chemical stimuli (e.g., capsaicin, bradykinin) capable of altering the sensitivity of units to subsequent stimuli were also of limited value. Given the paucity of data on thermosensitive splanchnic neurons and the common occurrence of chemosensitivity in C-fiber units activated by heating in other tissues (Kessler et al. 1992; Kumazawa and Mizumura 1980a; Szolcsányi et al. 1988), we have used a warm stimulus to locate mesenteric receptive fields. Warming the receptive fields of polymodal units in other tissues at intensities that are not overtly damaging does not cause pronounced changes in the sensitivity of these units, although responses to repeated “identical” warm stimuli may vary widely(Cohen and Perl 1988; Kumazawa and Mizumura 1980b; Reeh et al. 1986).
The present report deals with the properties of slowly conducting splanchnic afferent neurons responsive to warming up to maximum temperatures of 42–49°C. Mechanosensitive units insensitive to either warm, cool, or transient chemical stimuli (e.g., 90 nM BK, 880 mM H2O2, 1 mM histamine) also were encountered in this preparation. Mesenteric mechanoreceptors have been described in detail by other authors (Bessou and Perl 1966; Cervero and Sharkey 1988; Morrison 1973) and were not studied in the current work. Warm-sensitive units were characterized in terms of their response to rapid focal warming, ongoing activity, sensitivity to mechanical stimulation, responsiveness to bradykinin (BK) and sensitization to warming induced by BK. Sensitization phenomena observed in some units suggested features in common with polymodal receptors in other tissues, many of which may serve a nociceptive function. A nociceptive function for splanchnic afferent units responsive to rapid warming is consistent with their role in mediating cardiovascular reflex responses to intense warming (Rozsa et al. 1988) and their likely participation in the similar responses to severe core temperature increases during heat-stress (Kregel et al. 1988). However, given the behavioral responses to moderate intraabdominal warming in other species (Rawson and Quick 1972; Reidel et al. 1973), the dense distribution of splanchnic receptive fields in the vicinity of the celiac brown adipose tissue (BAT) deposits suggest the possibility that some warm-sensitive splanchnic neurons may play a role in normal thermoregulation.
Surgery and preparation
Experiments were performed on male Sprague-Dawley rats weighing 50–200 g. Rats were anesthetized by 0.5–0.8 ml 25% (wt/vol) urethan (im) per 100 g body wt. The mesentery and the subdiaphragmatic splanchnic nerves were removed as described elsewhere (Adelson 1995; Adelson et al. 1996). The tissue was placed in the main chamber of a two-chambered acrylic perfusion bath (18-ml capacity) into which oxygenated 31–33°C rat physiological saline plus glucose [RPSG containing (in mM) 135 NaCl, 5 KCl, 1 MgCl2, 1.3 Na2HPO4, 5 N-(2-hydroxyethyl)piperazine-N′-(2-ethanesulfonic acid) (HEPES), 2 CaCl2, 11.2 glucose, pH 7.4] (Wiley-Livingston and Ellisman 1982) was circulated (2ml/min). The lower temperature (relative to core body temperature) of the superfusate promoted long-term (>8 h) tissue viability. The splanchnic nerve was threaded into the recording chamber, a “picking window” (∼3 mm in length) was opened in the epineural and perineural coverings using fine forceps (Dumont No. 5), and dissected (“picked”) nerve twigs (length: ∼2 mm; diameter:5–15 μm) were placed on a platinum recording electrode in a layer of mineral oil over the nerve.
The final preparation comprised three regions: base, stem, and branch. The base region consisted of the segment of aorta extending from the celiac artery to the superior mesenteric artery and supported by a small amount of underlying muscle tissue; the severed stumps of the renal, pancreatico-duodenal, and suprarenal (superior and inferior) arteries and veins; the celiac-superior mesenteric ganglionic complex; large deposits of BAT; a number of lymph nodes and lymphatics; fascia; and the intact superior mesenteric artery. The stem region consisted of the trunks of the mesenteric artery, vein, lymphatics and nerves and the cisterna chyli and intestinal lymph nodes underlying them. The branch region comprised the web of bifurcating tributaries of these trunks, the transparent, avascular supporting tissue between branches, and the border of the mesentery, which was the former site of intestinal attachment.
Neural impulse activity was amplified and monitored using standard electrophysiological equipment and was recorded on VHS videotapes using a digital data recorder (Vetter 200 PCM Recorder). Stimulus data (time marks or temperature data) were recorded simultaneously on an auxiliary channel. Record lengths ≤6 h were obtained. The condition of the preparation was judged by its color and appearance and by the quality and stability of impulse waveforms. In deteriorating preparations, waveforms decayed substantially, becoming broader with lower peak amplitudes and greater variability in shape.
Locating, mapping, and stimulation of receptive fields
Before searching for receptive fields, ongoing activity (or silence) was recorded for several minutes. The tissue then wasexplored using a camel's hair brush. The mesenteric stem was explored first, followed by the branch region, and then the base. If no mechanoreceptive fields were located, the tissue was searched again using a stream of warm (42–49°C) rat physiological saline without glucose (RPS) delivered from a hand-held prewarmed (in a 55°C water bath) hypodermic syringe through a 24-gauge hypodermic needle (0.3 mm ID, Becton Dickinson). Receptive fields were found much more commonly using a warm search stimulus than with a mechanical search stimulus. Glucose was omitted from the test solution to eliminate the possibility of bacterial growth in either the syringe used for searching or the fixed-position stimulus apparatus (described below). No obvious differences were noted between the responses to warming using RPS and those to RPS + glucose, which was used in preliminary experiments. Warm-sensitive units were not responsive to application of either comparable, or even more vigorous, streams of bath-temperature RPS, indicating that neither the mechanical component of the stimulus nor the difference in osmolarity between the superfusate and the test solution (<4% hypotonic relative to superfusate) was adequate to excite the units tested.
Warm-sensitive receptive fields were explored in detail undera dissecting microscope, using warm RPS ejected at a rate of∼0.02–0.2 ml/s. Warm-receptive fields were spotlike; displacements of the stimulus of as little as 0.2 mm usually resulted in substantially reduced discharge. After accurate localization of the receptive field, the sensitivity of the unit to probing with von Frey hairs with bending forces up to ∼10 mN (1 g equivalent mass) or with a camel's hair brush was tested. In some cases, strong stimuli (grasping with forceps, firm stretch in various directions, pressing with the bristle base of the brush) were applied if the unit did not discharge to these moderate mechanical stimuli.
To measure the temperature at the receptive spot, a fine K-type thermocouple (0.001-in wire, Omega Engineering) connected to a digital thermometer (Omega DP205-TC) then was positioned against the tissue surface, immediately adjacent to the receptive spot, using a ball-joint manipulator. For experiments requiring repeated comparable warming trials applied to the same point, i.e., those investigating sensitization to warming after BK application, a fixed-position stimulus apparatus—consisting of a 24-gauge hypodermic needle (0.3 mm ID) held in place with a heavy-duty ball-joint manipulator and connected to a length of polyethylene tubing terminating with a Luer-lock connector—was positioned above the tissue, adjacent to the receptive field and thermocouple, at an angle typically 45–75° with respect to the plane of the tissue. The total void volume of tubing plus syringe was 0.07 ml. To warm the tissue, a preheated RPS-containing syringe was connected to the stimulus apparatus, and a stream of warm saline was ejected. The ejection of the void volume RPS resulted in a transient cooling of the receptive field before warming. The number and frequency of evoked impulses was highly sensitive to the positioning of the fixed-position stimulus apparatus, and as many as 10 trials commonly were required to achieve optimal placement of the thermocouple and the stimulus apparatus. Because repeated mechanical or warm stimulation at short interstimulus intervals was prone to cause erratic behavior and/or temporary inactivation, units were allowed to recover for ∼4 min or more after each stimulus that evoked substantial discharge (i.e., more than ∼5 impulses/s). Thus all records presented here that show responses to repeated warming were obtained after an extended period of repeated warm stimulation.
The voltage output of the thermometer was displayed on a digital oscilloscope (Tektronix 2211). A typical temperature trace, along with a corresponding plot of the rate of temperature change, and a neurogram of activity evoked during this stimulus is illustrated in Fig. 1. The entire period shown, comprising an initial cooling phase (due to evacuation of the void volume RPS), a warming ramp/transient, a plateau/static phase, and the cessation of the stimulus, is referred to as a single trial. Once the stimulus apparatus had been positioned accurately, the temperature profile of the first trial in a series was stored on the digital oscilloscope as a reference. In subsequent trials, saline was ejected from the same, prewarmed syringe at a rate allowing reproduction (tracing) of the reference temperature profile. In this way, it was possible to produce a series of comparable warming trials at a defined position. Departures from the original warming profile were registered by the thermocouple and used to qualitatively assess the effects of small temperature excursions on discharge. Manual control of the syringe was superior to use of a syringe pump (Sage Instruments, Model 341), because the pump produced large erratic and irreproducible surges in the initial delivery before stabilizing at a given flow rate.
BK (Sigma)—usually 1 μl of a 9–90 nM solution, although in preliminary experiments concentrations from 0.9 to 450 nM were tested—was delivered using a pipetman (Rainin, P20). The pipet tip (0.5 mm ID) was positioned so that the volume of solution to be delivered exceeded the volume intervening between the tip and the identified receptive field (i.e., within 0.5 mm). In several experiments, units were stimulated with 0.5 ml of 9 nM BK delivered through the fixed stimulus apparatus, which was flushed immediately with bath-temperature RPS.
Conduction velocities were determined after observations on natural stimulation had been completed, using a concentric electrode (0.45 mm OD) placed either at the receptive field or, in cases where the terminal had become unresponsive, on the nerve trunk. Measurements were made at the end of the period of observation because the firm pressure of the electrode against the tissue that was required to obtain satisfactory electrical contact often caused lasting deformation of the tissue. Conduction distances were taken as the shortest arterial path (because the nerves travel alongside the arteries) from the concentric electrode to the recording site (typically 5–30 mm). Conduction velocities for units in >70 splanchnic nerve twigs were measured, and at no time were conduction velocities in excess of 1.4 m/s observed in agreement with the histological observation that the fibers of the mesenteric nerves of the rat are exclusively unmyelinated (Cervero and Sharkey 1988). Although mesenteric fibers from the splanchnic nerve may lose their myelin beyond the prevertebral ganglia, studies on the splanchnic nerve of the rat (Kashiba et al. 1991), although more equivocal, indicate <10% of the splanchnic afferent fibers are myelinated at the level of the diaphragm.
Activity records were analyzed using the WAVEFORM acquisition, sorting, and graphing software, which we developed using the Borland C++ 3.0 compiler (Borland International) forMS-DOS (Adelson 1995). Data were acquired from tape onto an IBM-compatible computer (80386 or 80486 CPU) equipped with an A/D-D/A board (DT2831, Data Translation). Impulse activity was acquired in an event-driven mode (voltage threshold triggering, capturing 1.5 ms pretrigger and 4.5 ms posttrigger data) at a 25-kHz sampling rate on one channel while stimulus data were acquired continuously on a separate channel at 500 Hz. Impulses in multiunit recordings were sorted initially by clustering based on maximum peak height and dominant descending slope. Final sorting was performed by comparing sequentially each individual impulse waveform in the record with reference waveforms representing each waveform type present in that record (Adelson 1995; Adelson et al. 1996). Processed data files containing the time of occurrence of each impulse waveform type (unit) were generated. Discharge frequency versus time (frequency plots), and impulses per unit time (impulse histograms) were plotted above temperature traces using the WAVEFORM graphing module. For analyzing the relationship of impulse activity to temperature and rate of change of temperature, temperature and impulse frequency data were exported from WAVEFORM into a commercially available mathematics software package (MATLAB, The Math Works). Rate of change of temperature, ΔT/Δt, for each 100-ms time bin was calculated as the difference between the mean temperature in that time bin and that in the preceding bin.
BK-induced sensitization to warming was evaluated by comparing the number of impulses evoked by each post-BK warming trial with the mean (n = 3–6) of the pre-BK responses to comparable warming. A unit was considered sensitized during a post-BK warming trial if the number of impulses elicited exceeded the pretrial mean by >2 SD of that mean. This definition was used because it allowed evaluation of the state of single units at particular time points. A parametric statistical test of the significance of the change for each unit is not possible because the variance of the response of a single unit at a single time point cannot be determined. However, a response 2 SD greater than the pretrial mean corresponds to P < 0.05 for a t-test comparing a single post-BK response with the pretrial mean assuming equivalent variances.
Responses to warm stimuli
The responses of most units (33/41) to a warming trial consisted of a phasic component during the warming transient and tonic discharge at stable or slowly decreasing suprathreshold temperatures (Fig. 2). These units were designated slowly adapting (SA-W), to contrast their behavior with that of units (8/41—rapidly adapting, RA-W) whose discharge in response to warming was purely phasic (Fig. 3). The adaptation characteristics of a given unit's response to warming were reproducible and consistent from a variety of stimulus positions, although some variation in the time to adapt/fatigue in SA-W units occasionally was noted.
SA-W units differed in their relative sensitivity to the transient and sustained components of the warm stimulus, and, in some cases, discharge during the plateau phase of the warming trial decayed before the end of the plateau. Additional activity often could be elicited by a further temperature increase of a degree or more, although this discharge was typically not sustained for more than a second or two even at the higher temperature. Small increases in temperature (e.g., <0.5°C) occurring after a suprathreshold temperature plateau had been reached often caused a single high-frequency volley of discharge, which usually was followed by several seconds of reduced discharge or silence. In units with ongoing discharge, impulse activity was depressed by the brief initial cooling preceding a warming trial, and a period of silence (recovery) lasting 20–120 s typically followed the termination of the warm stimulus. The duration of recovery was generally consistent for consecutive stimulus trials, but it took distinctly longer after responses that included a high-frequency burst of discharge or when stimuli were repeated in quick succession (i.e., interstimulus interval <4 min).
Up to a point, tonic discharge in the 33 SA-W units increased with increasing temperature. However, in some units, it was found that raising the temperature above a “critical” value caused the pattern of discharge to become erratic and less clearly associated with the stimulus than at more moderate temperatures (Fig. 4). Often, the total discharge was lower, although the peak discharge frequency was typically higher, for intense warming than for more moderate warming trials of the same duration. This behavior is consistent with the bell-shaped thermal response curves (i.e., maximal discharge at a particular temperature and fewer impulses/second at either lower or higher temperatures) of both polymodal and specific warm fibers in other tissues (Cloze et al. 1976; Duclaux and Kenshalo 1980; LaMotte and Campbell 1978; Spray 1986).
RA-W units (8/41) discharged only in response to warming transients and were silent at any stable suprathreshold temperatures (Fig. 3). Because the existence of units with purely phasic responses to warming was unexpected, the possibility that this behavior could be an artifact of stimulus positioning was tested by delivering the stimulus from a number of positions and also by directing the stimulus to a number of sites surrounding (∼1 mm displaced from) the maximally sensitive spot. RA responsiveness was consistent for all stimulus-target combinations tested. Also, units with RA responses to warming ceased discharging at each plateau temperature tested (range 34–42°C). A relationship of RA discharge to the rate of change of temperature was apparent at temperatures sufficiently (≥1°C) higher than the threshold temperature (Fig. 5). The number of impulses and peak discharge frequency evoked by a warming step was greater at higher temperatures. The temperature at which units with little or no ongoing activity first discharged during warming, measured on three or more consecutive trials for each unit, spanned an equally wide range (30–47°C) for both RA-W (n = 6) and SA-W (n = 5) units.
Receptive field distribution and ongoing activity
Seven units had receptive fields located in the mesentery branch region, either along the border of the mesentery(n = 2) or along one of the radial branches (n = 5). One unit had a receptive field located on the mesenteric stem, ∼1 cm distal to the celiac ganglion. The majority of the units (33/41, 80%), including all units with RA responses to warming, had spotlike receptive fields in the base region, usually on or adjacent to blood vessels of the splanchnic vascular distribution, with the remainder located on the fascia, brown adipose tissue deposits or lymph nodes overlying much of the base area. This predominance of units encountered in the base region occurred even though the branch and stem regions always were explored carefully before searching the base. Units with receptive fields overlying the prevertebral ganglia were discarded to avoid the complicating possibility that warming-evoked activity might result from ectopic discharge initiated within a ganglion.
Background activity (0.02–2.0 impulses/s) initially was present in 21 of 33 units with SA responses to warming (SA-W). In 10 of the remaining 12 units, it developed during the course of the observations. The two SA-W units in which ongoing activity was never observed were studied for 30 min and 4 h, respectively. One of eight RA-W units exhibited background activity, which was present from the beginning of the recording. In most units, background activity was either stable or increased steadily (e.g., a steady increase from 0.3 to 0.8 impulses/s in 40 min). In some units, however, abrupt increases in background activity occurred immediately after stimulation (e.g., from 0.05 to 0.66 impulses/s), whereas in a few units, steady decreases in discharge occurred (e.g., from 0.32 to 0.22 impulses/s over 60 min). In several units, ongoing activity ceased before the end of observations, although applied stimuli continued to excite impulse activity.
The manner in which background activity developed in two units was unusual: it appeared as a series of clusters, or crescendos, of impulses initially lasting ∼20 s/cluster, the discharge rising to a peak discharge frequency of 6 Hz after 10 s and then decaying back to 0 (Fig. 9). In the unit shown, these crescendos began several minutes after a warming trial that elicited a response and 15 min after the unit responded to BK (1 μl 90 nM; not shown). During a 30-min period,the time between crescendos decreased from ∼90 s one of the two such units and 60 s in the other, until the activity became a steady ongoing discharge with a mean frequency of 0.6 Hz in both cases.
In several units, the period of recovery was followed consistently by a cluster of discharges clearly distinct from the background. This increased activity (relative to background activity preceding the stimulus) was termed postrecovery afterdischarge (Fig. 6). Postrecovery afterdischarge was itself succeeded by a brief recovery (silence) in some cases (Fig. 6, A and B).
Mechanical sensitivity of warm-sensitive units
One RA-W unit (1/8, 12%) and nine SA-W units (9/33, 27%) responded to probing with von Frey “hairs” with bending forces <10 mN (∼1 g equivalent mass). A further five SA-W units, which were tested by compression with a metal probe or firm stretching of the tissue, were activated by these strong mechanical stimuli. Units activated by von Frey hair probing had spotlike mechanoreceptive fields, i.e., stimuli applied 0.5 mm from the center of a receptive spot (in any direction) elicited little or no discharge provided the tissue was not appreciably stretched. Firm stretch usually elicited greater discharge than did punctate stimulation. Responses to stretch stimuli were slowly adapting, whereas punctate stimulation yielded both rapidly and slowly adapting responses, depending apparently on the shear stress applied and the position of the probe tip. With one exception, in which the mechano- and warm-receptive spots appeared separated by 0.8 mm, the mechano- and warm-receptive spots coincided. As with thermal stimulation, mechanically evoked discharge was followed by a recovery period during which discharge was absent.
One SA-W unit initially unresponsive to mechanical stimuli (using a camel's hair brush and von Frey hairs ≤10 mN) became responsive to stimulation with a 2 mN (0.2 g) von Frey hair for several minutes after brief warming (Fig. 7). This mechanical sensitivity decayed over a period of ∼5 min, but could be restored by repeating the warming stimulus.
Because the thermal stimulus, a stream of warm saline, visibly deformed (“dimpled”) the tissue during its application, the possibility that mechanical forces could contribute to warming-evoked discharge was investigated. Although no unit responded to application of a similar stream of bath-temperature RPS (including the units activated by von Frey hair probing), it was thought that the sensitivity to the mechanical component of the stimulus might increase with increasing temperature, a possibility that found support in the observation of a unit whose mechanical sensitivity was enhanced transiently by prior warming (described above). To examine the extent to which the mechanical component of the stimulus might influence the discharge at elevated temperature, two approaches were used. In the first, the syringe containing the saline to be ejected was warmed to either 50, 55, or 60°C. The temperature to which the tissue is warmed by the saline depends on both the initial temperature of the saline and the velocity at which it is delivered. If mechanical sensitivity at warmer temperatures contributed substantially to evoked discharge, then initially cooler, higher velocity streams would be expected to cause greater impulse activity than initially warmer, slower streams, that warmed the tissue to the same plateau temperature. Two units were tested using this method, and no obvious difference in discharge at a plateau temperature of ∼45°C was observed. However, the warmer, more slowly delivered streams caused a steeper warming ramp than did the less warm, higher velocity streams, and the more rapid warming transient associated with a lower mechanical force in the former case caused a greater discharge than did the slower ramp in the latter case, suggesting that temperature and not mechanical perturbation was the critical activating stimulus. The second approach to investigating whether the mechanical component of the warming stimulus had a substantial effect on impulse activity was to vary the angle at which the saline was delivered. Although streams delivered perpendicular to the tissue or nearly parallel to the tissue cause the same degree of surface warming, the force applied on the tissue by the perpendicular stream is much greater than that for the nearly parallel stream. One unit was tested several times using this method, and no obvious differences in impulse activity were noted using perpendicular versus nearly parallel approaches producing the same final temperature.
BK sensitivity of warm-sensitive units
SA-W units and RA-W units were tested for responses to focal, transient application of 9–90 nM BK; concentrations that have been shown to elicit responses from testicular polymodal receptors (Kumazawa et al. 1991). If no response was observed after an initial application of BK, the stimulus was repeated after a pause of ≥10 min. BK elicited responses (Fig. 8) in 8 of 16 SA-W units tested, including twoSA-W units lacking ongoing activity. Two BK-responsive SA-W units were demonstrably mechanosensitive, and in both cases, strong mechanical stimuli were required to activate them. Three of the eight BK-unresponsive SA-W units responded to probing with von Frey hairs of equivalent mass <1 g, whereas one was activated only by strong mechanical stimuli. One of six RA-W units tested responded to BK application. This BK-sensitive RA-W unit had ongoing discharge but was not mechanoresponsive. Thus mechanosensitivity did not predict BK sensitivity nor vice versa. Two units did not respond to the first application of BK but did to the second.
Warm-sensitive splanchnic units displayed several types of sensitization, which typically is recognized by increased number and frequency of impulses evoked by a particular stimulus, reduced threshold for response, and/or increased background activity. As already noted, the background activity of a number of warm-sensitive units increased during the course of observations. In addition, transient sensitization induced by warm stimuli or by BK also were observed. In the following discussion, it is useful to distinguish between the stimulus that induces the sensitization (here warming or BK) and the means by which the sensitization is recognized (e.g., increased background activity or increased responsiveness to a subsequent mechanical or warm stimulus).
SENSITIZATION INDUCED BY WARMING.
As described earlier, one unit was encountered that was insensitive to probing by von Frey hairs with bending forces <10 mN before warming but that could be activated by a von Frey hair with a bending force of 2 mN for several minutes after its receptive field had been briefly warmed to temperatures >42°C (Fig. 7). Units that were both mechano- and warm-sensitive appeared to give more vigorous responses to probing with von Frey hairs shortly after a warm stimulus, but the effect was not dramatic.
Another form of warming-induced sensitization could be noted in both of the units described above in which ongoing activity developed as a series of clusters. During the period in which the background discharge was organized into clusters, warming both evoked discharge and led to an increase in the maximum frequency and number of impulses in the following one or two clusters (Fig. 9). Thus warming may be considered to have transiently sensitized the unit to whatever factor was responsible for periodically exciting the unit. The phenomenon of postrecovery afterdischarge (Fig. 6) may represent a related process.
SENSITIZATION INDUCED BY BK.
Sensitization to warming after application of BK was investigated by stimulating a unit with a series of comparable warming trials at regular intervals, followed by transient focal application of BK and then another series of warming trials (Fig. 10). Impulses evoked by each warming trial after BK application were compared with the responses obtained in the series of trials preceding BK application. The number of impulses evoked by repeated warming trials before BK application varied from the mean by less than ±25% for all units analyzed. BK transiently sensitized one RA-W and six SA-W units to subsequent warming, as judged by an increase in the amount of impulse activity evoked by a warm stimulus of >2 SD above the pre-BK mean (range of increase: 40–180%). The response to warming of the RA-W unit sensitized by BK remained rapidly adapting.
The BK-induced sensitization to warming decayed over the course of approximately 10 min (Fig. 10). In only one unit (Fig. 10 C) was the number of impulses per trial more than one standard deviation above the pre-BK mean on the second post-BK warming trial. In this same unit, BK was applied again following three further warming trials for which the number of evoked impulses had returned to pre-BK levels. The second BK application evoked 8% more impulses than did the first (144 in 36 s vs. 133 in 34 s), but no augmentation of the response to warming was observed in the following warming trial (on the contrary, a slight depression occurred). The presence of comparable levels of BK-evoked activity for the two BK trials, in contrast with a pronounced difference in the degree of BK-induced sensitization to subsequent warming, indicates distinct pathways mediating BK-evoked impulse generation and BK-induced sensitization to warming.
BK sensitized the response to warming without directly evoking discharge in two units. In both of these units BK did not evoke a response on the first application, augmented the discharge to the next warm stimulus, and elicited impulse activity the second time it was applied. In one of these units, the second BK application also augmented the response to the following warming trial; in the other, it did not, again indicating the separability of BK-evoked discharge and BK-induced sensitization. The former of these two units had not been activated by any stimulus prior to and including its first exposure to BK, despite thorough searching of the area during placement of the stationary stimulus apparatus (to test the responsiveness of another unit in the same nerve twig). Thus this unit might be considered a silent or “sleeping” unit. After “awakening,” the unit became responsive to both warm stimuli and BK. It later developed ongoing activity and remained responsive to repeated warming for >2 h. This unit was not the sole example of a unit that awakened during the course of an experiment (often after a BK stimulus). However, it was the only such unit whose receptive field overlapped that of another unit being studied in detail, ensuring that it had been repeatedly stimulated before awakening.
In these experiments, a thermal (warm) search stimulus was used to locate the receptive fields of splanchnic afferent neurons. Seventy-six percent (31/41) of these units did not respond to moderate mechanical perturbation and escaped detection using a mechanical search stimulus. Warm search stimuli enabled identification of receptive fields much more commonly than did mechanical search stimuli in this in vitro preparation, even though the tissue was searched carefully with a mechanical search stimulus before using a thermal search stimulus. Thus units responsive to rapid warming of their receptive fields appear to constitute a sizable subset of splanchnic afferent neurons.
Apart from testicular units (Kumazawa and Mizumura 1980b; Kumazawa et al. 1987, 1991), warm stimuli rarely have been used to characterize visceral afferent neurons. Information on the warm-sensitivity of nontesticular visceral afferent units is, according to comprehensive reviews of the literature (Cervero 1994; Mei 1983), limited to one in vivo study of 13 splanchnic fibers in the rabbit (Reidel 1976), in which it was found that warm-sensitive splanchnic receptors were localized to the dorsal wall of the upper abdomen in the vicinity of the aorta. In the present in vitro experiments, warm-sensitive units were encountered frequently. Warm-sensitive receptive fields were punctate and were most dense in the region surrounding the prevertebral (sympathetic) ganglia, consistent with the in vivo finding in the rabbit (Reidel 1976). Reidel (1976) reported that the splanchnic units that he studied were not mechanosensitive, although no details were provided concerning the manner in which mechanosensitivity was tested.
Responses to warming
The splanchnic units in the present study were divided into two groups, RA-W and SA-W, on the basis of whether their responses to warming trials included tonic discharge during the plateau phase of the stimulus. This feature was reproducible on repeated warming trials. Although the thermal coding properties of the neurons in this study were not systematically investigated, the responses of the SA-W units are consistent with those described by Reidel (1976), who distinguished two populations of units among 13 splanchnic fibers: one with static and dynamic response maxima centered around 40°C, the other around 46°C. The threshold of one unit from the latter group exhibited activity at temperatures as low as 30°C (Reidel 1976), indicating that the position of the response maximum is likely to provide a better clue to physiological function than does the threshold temperature. The activity of the former group became burstlike at temperatures >40°C and ceased as the temperature reached ∼45°C (Reidel 1976), a phenomenon that also was observed in several SA-W units in the present study.
The behavior of the RA-W units described here, in contrast, do not correspond to any of the units described by Reidel (1976), perhaps because the warming transient tested in that study (∼7°C/min) was slow relative to the stimulus used here. An illustration of rapidly adapting responses of specific warm fibers to rapid, intense warming (12°C in 300 ms) has been published in a report comparing activity evoked in warm and polymodal fibers of monkey skin (LaMotte and Campbell 1978). In that study, more moderate warming of the receptive field of specific warm fibers elicited slowly adapting responses. In contrast, activity in RA-W units in the present study ceased even when tested at low (<38°C) plateau temperatures. Of greatest concern during the experiments was the possibility that the RA responses observed might be an artifact of the positioning of the stimulus. Polymodal receptors of the dog testis give rapidly adapting responses when mechanical stimuli are displaced slightly from the optimal receptive spot, but slowly adapting responses when stimulated directly on the receptive spot (Kumazawa and Mizumura 1980b). We have observed the same phenomenon in mechanosensitive units in the splanchnic nerve-mesentery preparation. However, this was not the case for warm stimuli applied to RA-W units: despite repeated application of warm stimuli delivered from a number of positions and orientations, tonic discharge during the plateau phase of the stimulus was consistently absent in RA-W units. Although positioning influences cannot formally be ruled out, several observations suggest that the presence or absence of a tonic component in the activity evoked by warming may reflect genuine differences between units: 1) phasic discharge evoked from RA-W units was in many cases greater, not less, than that of SA-W units, both in number of impulses and peak discharge frequency; if inaccurate stimulus positioning were responsible for rapid adaptation, then the reverse would be expected; 2) the temperature during warming at which units began discharging covered the same (wide) range for both types of units; and 3) in the RA-W unit in which BK augmented the number of impulses evoked by the subsequent warming trial, the response to that warming trial was still rapidly adapting.
The converse notion, that RA-W units might mimicSA-W responses to warming if the stimulus were applied to a site displaced from the receptive field, could be argued on the grounds that a site distant from the thermocouple would warm more slowly than indicated by the thermocouple and therefore could continue to experience warming when a plateau temperature at the thermocouple had been reached. This arrangement would explain the typically lower discharge frequencies and the prolonged responses of SA-W units. However, if SA-W responses resulted from “off-target” delivery of stimuli, then the much greater proportion of chemically and mechanically sensitive SA-W units would be difficult to explain, as receptive fields would be expected tobe less accessible, not more, than those of RA-W units. Although no sensitivity or combination of sensitivities was coexpressed exclusively with the RA-W or SA-W phenotype, ongoing activity, BK sensitivity, and mechanical sensitivity all occurred with lower frequency in RA-W than in SA-W units.
For these reasons, distinctions were made between units based on degree of static and dynamic sensitivity to warming. The physiological implications of such differences are not clear. Differences in adaptation characteristics may be meaningful in the context of a thermosensitive system, although temperature stimuli more likely to occur in vivo must be tested to investigate this possibility. Alternatively, differences in adaptation rates to warming might distinguish between classes of units with distinct physiological functions yet to be determined. Responses to warming do not prove that temperature is the signalled variable in vivo. Raising temperature increases the thermodynamic activity (i.e., the “effective concentration”) of chemical compounds in solution, and so responses to rapid warming may result from changes in the effective concentrations of particular endogenous chemicals. For example, the discharge evoked from testicular units by a particular concentration of BK is greater at higher temperatures, as are responses to hypertonic saline (Kumazawa et al. 1987). Thermal stimuli therefore may provide an effective means of locating receptive fields of chemosensitive units, whose particular chemosensitivity is not known, in nerves in which a predominance of efferent fibers (as in the splanchnic nerve) renders electrical stimulation inefficient as a search stimulus for afferent units. Finally, the possibility that the RA-W phenotype results from damage to either axons or terminals cannot be ruled out. It is doubtful, however, that RA discharge results from ectopic activation of normal axons, because the RA-W units' receptive fields clearly were punctate.
Possible physiological relevance of splanchnic warm sensitivity
The warm-sensitive splanchnic afferent neurons identified in this work may play a role in the profound changes in heart rate and vascular resistance that accompany severe heat stress and/or heat stroke. In a model of heat stroke developed by Kregel et al. (1988), exposure of conscious rats to an ambient temperature (T a) of 46°C causes core temperature (T c) to rise to 44°C during the course of 1–2.5 h. Accompanying this rise in core temperature are first, increases in superior mesenteric arterial (SMA) resistance, followed by a precipitous drop as T c approaches 42°C. Systemic mean arterial pressure (MAP) also falls rapidly at elevated core temperatures, the drop in MAP occurring 10–15 min after the drop in SMA resistance. These cardiovascular responses are similar in anesthetized animals (Kregel et al. 1988). In this heat-stress model, celiac ganglionectomy blocks the initial increases in SMA and attenuates the increases in MAP, but both indices still show pronounced drops as T c approaches 42°C (Kregel and Gisolfi 1989).
Thermally evoked cardiovascular reflexes sharing features with those seen in heat-stressed rats have been observed in response to transient, local application of prewarmed saline to exposed serosal and/or mucosal abdominal tissues of the rat in vivo (Rozsa et al. 1988). This stimulus, similar to that used in the present experiments, causes a transient increase in splanchnic afferent nerve activity, along with a complex pattern of responses dependent on the temperature of the applied fluid (Rozsa et al. 1988). Using 45°C saline, tachycardia and systemic hypotension are found consistently, along with a mesenteric vasodilator response when sympathetic activity is blocked by pretreatment with reserpine (Rozsa et al. 1988). All cardiovascular responses to a warm saline stimulus are absent in rats treated neonatally with capsaicin and are attenuated substantially in normal animals by perineural application of capsaicin to the mesenteric trunk. Bilateral cervical vagotomy has no effect, but splanchnic ganglionectomy abolishes blood pressure and heart rate changes, without preventing local increases in mesenteric blood flow (Rozsa et al. 1988). Taken together, these data indicate that rapid local warming to 45°C results in centripetal impulse activity in capsaicin-sensitive splanchnic afferent neurons that mediates changes in heart rate and systemic blood pressure, whereas local vasodilator responses require the presence of intact peripheral processes of these afferent neurons.
A number of studies have demonstrated that peripheral release of calcitonin-gene-related peptide, contained in 88% of rat splanchnic afferent neurons (Kashiba et al. 1991), is responsible for mesenteric arterial vasodilation in response to splanchnic nerve stimulation (Claing et al. 1992; Nuki et al. 1993; Takenaga et al. 1995). By analogy to the situation in the skin (Kenins 1981), afferent fibers mediating vascular effects and responding to intense warming might be expected to be polymodal. Indeed, a number of warm-sensitive units studied here were responsive to either moderate mechanical stimuli or BK, although none were found that were responsive to both. Both SA-W and RA-W units described here would be expected to be activated by the warm stimulus used by Rozsa et al. (1988) to elicit a cardiovascular reflex indicative of noxious stimulation (Cervero 1982).
In addition to units mediating responses to noxious temperature increases, some warm-sensitive splanchnic units may play a role in modulating abdominal temperature. This possibility was suggested by earlier experiments demonstrating behavioral thermoregulatory responses in sheep that were dependent on an intact splanchnic, but not vagal, nerve supply (Rawson and Quick 1972). It is supported by the present observation of a dense distribution of warm-sensitive splanchnic receptive fields in the vicinity of the prominent brown fat deposits overlying the aorta and the celiac-superior mesenteric ganglionic complex. The function of BAT is heat generation (thermogenesis) (Smith and Roberts 1964), and BAT thermogenesis is under sympathetic control (Kawate et al. 1993, 1994). The primary function of interscapular BAT deposits in the mouse appears to be control of the temperature of blood supplying the heart (Smith and Roberts 1964). The celiac BAT deposits, one of the major BAT stores in the rat, may play an analogous role with regard to the abdominal arterial supply, possibly via a peripheral autonomic reflex arc operating through the celiac-superior mesenteric (sympathetic) ganglia. In the rat, the temperature of the upper abdomen has been shown to fluctuate in a regulated manner: in unanesthetized, unrestrained rats a circadian temperature rhythm (CTR) with an amplitude of ∼3°C has been observed (Satinoff et al. 1991), over which is superimposed a bihourly thermal oscillation with amplitudes reaching 1.5°C (Closa et al. 1993). The amplitude of the CTR is correlated with the degree of sympathetic motor activity and appears to be centrally regulated, as lesioning of the medial preoptic area of the hypothalamus increases the amplitude of the CTR to as much as 12°C (Satinoff et al. 1982).
Although the stimulus used in the present experiments was ill-suited to a systematic characterization of the thermal stimulus-response properties and coding potential of the units studied (owing to the difficulty of presenting a controlled series of small temperature steps), Reidel (1976) has demonstrated previously the presence of a minority of splanchnic warm-sensitive units with bell-shaped thermal stimulus-response curves with response maxima centered around 40°C. Such units could well participate in the normal modulation of abdominal temperature.
Whatever the physiological implication of the warm-sensitivity exhibited by many splanchnic fibers, it proved useful in demonstrating sensitization effects, because repeated comparable warming trials produced sufficiently reproducible responses in single units.
Transient sensitization resulting from warming
Several observations demonstrated that warming in the range tested may alter transiently the excitability of the neuronal sense organ. The first of these observations is that a brief warming stimulus can reveal reproducibly a responsiveness to subsequent punctate mechanical stimulation (Fig. 8) that decays over the course of several minutes after the warming trial. A similar effect has been observed by others: Bessou and Perl (1969) found that sensitization induced by heat stimuli was associated with decreased thresholds for mechanical stimulation in some cutaneous polymodal units they tested and that the mechanical threshold of the units that were sensitized by heating returned to preheating levels during the course of 15–30 min. The second observation of the transient sensitizing effect of warming on splanchnic fibers was noted in two units in which background discharge was initiated as a series of crescendos of activity, which decayed gradually into unpatterned, irregular activity over the course of about half an hour. Warming potentiated the number of impulses and the peak frequency attained in crescendos beginning ≥1 min after cessation of the warming trial (Fig. 9). Finally, several units exhibited postrecovery afterdischarge, in which elevated ongoing activity immediately followed the recovery period. This could result from the accumulation of some factor capable of causing increased discharge (relative to the subsequent background) during the recovery period after warming. Such a factor might be produced constantly and “consumed” as a result of impulse activity, so that after a period of recovery, it was present in a higher concentrations, or it could simply be released acutely as a result of warming and either diffuse away or be broken down within several minutes. Alternatively, these behaviors may result from warming-induced changes in excitability intrinsic to the sense organ itself, unassociated with changes in the terminal microenvironment.
Both temporary enhancement of sensitivity to a mechanical stimulus and temporary augmentation of background activity after warming suggest transient production or release of a factor or factors that sensitize(s) splanchnic afferent terminals. The reproducibility of the enhancing effect of prior warming on discharge for repeated trials in each case indicates that the transience of the effect after each trial was not due to a desensitization of the terminal but rather to a progressive reduction in the quantity or effectiveness of the putative actuating substance or process. Among the possibilities for such a substance are the prostaglandins and/or leukotrienes, which are produced at inflammatory sites and sensitize nociceptors (Martin et al. 1987, 1988). Exogenous prostaglandins sensitize polymodal cutaneous units to warm stimuli, but single doses are ineffective for this purpose; instead, continuous infusion of prostaglandins must be employed (Handwerker 1976). The splanchnic nerve-mesentery preparation provides an attractive system for investigating sensitization of visceral afferent units by prostaglandins, leukotrienes, and other compounds due to the simplicity of the final preparation, the direct accessibility of receptive fields, and the ability to control the physical and chemical environment of the tissue.
Transient sensitization to warming after BK application
Transient sensitization to warming after BK application was observed in a number of units as an increase in the number of impulses evoked by warming >2 SD above the pre-BK mean response. A transient increase in the number of impulses elicited per warming trial after BK application was observed in one RA-W and six SA-W units. In two cases, BK stimuli that did not evoke a response were capable of sensitizing units to subsequent warming, consistent with the finding in testicular afferent neurons that concentrations of BK inadequate to evoke discharge can augment the response to warming (Kumazawa et al. 1991). Sensitization to warming by BK also has been observed in an in vitro rat skin preparation (Koltzenburg et al. 1992), but in that system the number of impulses evoked by warming was found to vary by as much as a factor of two on repeated pre-BK trials, and so sensitization was determined using the pooled activity of a population of eight afferent units, rather than in individual units.
A novel finding of the current work was the observation that a BK stimulus that both evoked a substantial response and augmented the subsequent response to warming, when repeated half an hour later, elicited approximately the same amount of impulse activity but did not augment the response to subsequent warming at all (in fact, the response to the next warming trial was depressed). This indicates separate mechanisms mediating BK-induced impulse generation and BK-induced sensitization to warming and suggests the possibility that those units in which sensitization to warming after BK treatment was not observed already may have accommodated to the sensitizing effects of BK by factors generated in situ. Exogenous BK causes the release of prostaglandins in a dose-dependent manner (Lembeck and Juan 1976; McGiff et al. 1972), and sensitization to thermal stimuli by BK is believed to occur via prostaglandin production in other systems (Mizumura et al. 1987, 1991; Taiwo et al. 1990).
In the present study, BK-induced sensitization to warming decayed over the course of ∼10 min. Because concentrations of BK two orders of magnitude lower than the minimum capable of evoking discharge can sensitize testicular afferent neurons to warming (Kumazawa et al. 1991), it is possible that the BK-induced sensitization to warming observed in the present study resulted from the presence of dilute BK in the perfusion bath during post-BK warming trials. Progressive dilution of the bath concentration of BK might explain the time course in the decay of BK-induced sensitization. However, the time course of decay was approximately the same after application of either 1 μl of 90 nM BK or 0.5 ml of 9 nM BK, a 50-fold difference in the final bath concentration achieved after mixing. For the tests using 1 μl of 90 nM BK, the concentration of BK in the bath immediately after BK application, assuming adequate mixing, would be 0.05 nM, already half the minimum concentration found to have a sensitizing effect in the in vitro study of canine testicular afferent neurons (Kumazawa et al. 1991). Approximately the same time course of decay of sensitizing effects as that observed in the present study have been reported in in vitro studies of rat skin and dog testis (Koltzenburg et al. 1992; Kumazawa et al. 1991).
In summary, the use of a warm search stimulus provides a rapid means of locating receptive fields of splanchnic units, a number of which are not activated by moderate mechanical stimuli. These units may function physiologically as thermoreceptors, nociceptors, or chemoreceptive units whose activity scales with temperature. A particularly dense distribution of units around the upper abdominal prevertebral ganglia, the site of one of the five major stores of brown adipose tissue in the rat, suggests a possible role in sympathetic reflexes involved in thermoregulation. Although it is possible that some splanchnic warm-sensitive units mediate abdominal thermoregulation, a number of splanchnic units share features in common with nociceptors in other systems, i.e., they respond to more than one stimulus modality and they exhibit sensitization. These units may be responsible for a number of the cardiovascular effects, including eventual circulatory shock, observed in heat-stroke.
The authors thank Drs. T. H. Bullock, J. T. Enright, J. Graham, H. T. Hammel, and T. L. Yaksh for helpful discussions and comments and Dr. V. L. W. Go for providing facilities and equipment.
This work was supported by National Institute of Neurological Disorders and Stroke Grants NS-5685 and NS-28433.
Address for reprint requests: D. W. Adelson, West L.A. VAMC/CURE, Building 115, Room 325, 11301 Wilshire Blvd., Los Angeles, CA 90073.