INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

INCREASED COLONIC TEMPERATURE (T_c) produced by acute heating profoundly influences sympathetic nerve regulation. First, hyperthermia provides a potent stimulus to the sympathetic nervous system as demonstrated by heating-induced increases in the level of activity in sympathetic nerves (renal, splanchnic, splenic, and lumbar) that innervate different regional arterial beds (10, 18, 19). Second, hyperthermia changes the pattern of efferent sympathetic nerve discharge (SND) bursts. For example, during progressive increases in T_c from 38 to 41°C, the percentage of total power in SND at noncardiac-related frequencies is reduced with subsequent increases in total power at the frequency of the heart rate (HR) (18). This indicates that SND bursts can become more pulse synchronous during acute heat stress. Because the cardiac-related component in SND is dependent on intact arterial baroreceptors (8, 19), these results suggest that the arterial baroreceptors can influence the SND bursting pattern during hyperthermia. If T_c is increased to almost 41.5°C, the cardiac-related SND bursting pattern is transformed to a pattern that contains low-frequency, noncardiac-related bursts (19). Third, hyperthermia enhances the correlation between phrenic and SND bursts (19), demonstrating prominent respiratory modulation of efferent sympathetic nerve outflow during acute heat stress. Fourth, the discharge bursts in selected sympathetic nerve pairs (renal-splenic, renal-splanchnic, renal-lumbar) remain prominently coupled at frequencies between 0 and 15 Hz (including cardiac-related and low-frequency bursts) during acute heating (19), indicating that there is little regional selectivity in the frequency characteristics of SND bursts during hyperthermia. Pharmacological blockade of ganglionic transmission during increased T_c (41.5°C) reduces mean arterial pressure (MAP) to values less than those produced by ganglionic blockade at control (38°C) (19), suggesting that the sympathetic nervous system plays a key role in offsetting vasodilatory influences during progressive increases in T_c.

An important unresolved issue concerns what level(s) of the neuraxis are involved in mediating heating-induced changes in SND frequency components in chloralose-anesthetized rats. At least three possibilities can be considered. First, because the preoptic anterior hypothalamic area is considered an important thermointegrative center of the central nervous system (3, 11, 13, 14), forebrain systems, along with brain stem and spinal neural circuits, may be required for mediating SND responses to hyperthermia. If this is the case, then decerebrate (midbrain-transected) rats would be unable to generate the full spectrum of SND responses to acute heating. Second, because the brain stem contains thermosensitive neurons (11, 15, 20, 24, 25) and is known to play an important role in sympathetic nerve regulation (1), arterial baroreflex function (32, 33), and cardiorespiratory coupling (12), it may be that brain stem and spinal neural circuits, in the absence of forebrain systems, are capable of mediating changes in SND frequency components in response to hyperthermia. If this is the case, then heating-induced SND responses would be similar in decerebrate (midbrain-transected) and sham-decerebrate rats (forebrain and brain stem neural connections left intact). Third, because the role of the spinal cord as a temperature sensor in thermoregulation is well established (31) and because spinal sympathetic neural circuits are capable of generating sympathetic nerve activity (in anesthetized rats) after cervical spinal cord transection (26, 34), spinal systems, in the absence of supraspinal neural circuits, may be capable of influencing the frequency characteristics of SND bursts to acute heat stress. Because supraspinal neural circuits play a crucial role in the arterial baroreceptor reflex and cardiorespiratory coupling (12, 32, 33), it would be anticipated that changes in SND frequency components to heating in spinal cord-transected rats would not be identical to those observed in decerebrate or sham-decerebrate rats. However, this does not exclude the possibility that activation of spinal thermal systems during heat stress in spinal cord-transected rats profoundly influences noncardiac- and nonrespiratory-related frequency characteristics of efferent SND.

In the present study, we investigated the contributions of forebrain, brain stem, and spinal neural circuits to heating-induced SND responses. We determined the effect of increased T_c (38 to 41.6-41.7°C) on the level of activity and the frequency characteristics of renal and splenic SND bursts in midbrain-transected (superior colliculus), spinal cord-transected [first cervical vertebra (C1)], and sham-transected (midbrain and spinal cord) chloralose-anesthetized rats.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

General procedures. The surgical procedures and experimental protocols used were approved by the Institutional Animal Use and Care Committee. Experiments were performed on male Sprague-Dawley rats (300 g). Anesthesia was initially induced with methohexital sodium (Brevital, 50-60 mg/kg ip). Two catheters (PE-10 and PE-50) were placed in the femoral vein. The PE-10 catheter was used during the surgical preparation for administration of maintenance doses of methohexital sodium (10-20 mg/kg). The PE-50 catheter was used for the administration of an initial dose of -chloralose (50-60 mg/kg) and for maintenance doses (35 mg · kg¹ · h¹) throughout the surgical preparation and experiment. The trachea was cannulated with a PE-240 catheter, and rats were paralyzed with gallamine triethiodide (5-10 mg/kg iv, initial dose; 10-15 mg · kg¹ · h¹, maintenance dose) and artificially ventilated. Femoral arterial pressure and HR were recorded using standard procedures. T_c was measured with a thermistor probe inserted ~5-6 cm into the colon and was kept at 38.0°C during surgery by a temperature-controlled table. End-tidal CO₂ was kept between 4.5 and 5% during all surgical and experimental interventions.

Neural recordings. Activity was recorded biphasically with a platinum bipolar electrode after capacity-coupled preamplification (band pass, 30-3,000 Hz) from the central end of cut or distally crushed renal and splenic sympathetic nerves. The left renal and splenic nerves were isolated retroperitoneally. The nerve-electrode preparations were covered with silicone gel to prevent exposure to room air. The sympathetic nerve potentials were full-wave rectified and integrated (time constant, 10 ms), which produced a smooth tracing of the synchronized discharges. Activity in SND recordings was quantified as volts × seconds (V · s) (4, 19). The sympathetic nerve recordings were corrected for background noise after administration of the ganglionic blocker trimethaphan camsylate (10-15 mg/kg iv).

Midbrain transections. The rat was placed in a stereotaxic apparatus. After removal of a portion of the skull, midbrain transections were completed by performing sequential left and right hemisections at the level of the superior colliculus. Transections were completed through the rostral portion of the superior colliculus. The level of transection was verified by gross examination of the brain stem and by evaluation of sagittal sections (40-µm thickness) stained with cresyl violet. Data are reported only for those experiments in which midbrain transections were complete and verified to be through the rostral level of the superior colliculus. Sham transections were completed by removing similar portions of the skull followed by gently placing a scalpel blade on the exposed neural tissue without performing the surgical hemisections.

Cervical spinal cord transections. The rat was placed in a stereotaxic apparatus, and a laminectomy was performed to expose the spinal cord. After removal of the dura, mineral oil was applied to the exposed spinal cord to keep it moist. Using a scalpel blade, the spinal cord was completely transected at C1. Sham transections involved completing a laminectomy to expose the spinal cord followed by gently placing a scalpel blade on the exposed neural tissue without performing surgical transection of the spinal cord. Spinal cord transection was verified at the end of each experiment by dissection at the lesion site. Data are reported only for those experiments in which spinal cord transections were completed and verified to be at the level of the first cervical vertebra.

Experimental protocol. After completion of the initial surgical procedures (e.g., arterial and venous cannulations, isolation of sympathetic nerves, removal of portions of the skull or spinal cord laminectomies), the anesthetized rats were allowed to stabilize for 30 min (referred to as pretransection control) before completion of surgical transections (midbrain or spinal cord) or sham transections (midbrain or spinal cord). At the end of the pretransection control period, midbrain and spinal cord transections (or the related sham transections) were completed. The animals were allowed to recover for 30-40 min after the surgical and sham transections (posttransection recovery period). T_c was maintained at 38°C during the pretransection control and posttransection recovery periods. At the end of the posttransection recovery period, T_c was increased at a rate of ~0.1°C/min from 38 to 41.6-41.7°C using a heat lamp positioned ~40 cm above the animal. The heating protocol was completed in both surgical-transected (midbrain and spinal cord) and sham-transected (midbrain and spinal cord) rats.

Data analysis. Autospectra and coherence analyses of the arterial pulse and SND bursts were computed using the methods and programs described earlier (4, 17, 22). Fast Fourier transform was performed on 12-24 contiguous windows of data that were 5 s in duration. The signals were sampled at 200 Hz. Autospectra and coherence functions were computed over a frequency band of 0-15 Hz. The amplitude of the autospectra were autoscaled to the highest peak (22). The frequency resolution was 0.2 Hz/bin. Spectral analyses provide the following information (17, 21, 22). The autospectrum of a signal shows the relative power present at each frequency. The coherence function (normalized cross spectrum) provides a measure of the strength of linear correlation of two signals as a function of frequency. The squared coherence value (referred to as coherence value) is 1.0 in the case of a linear system undisturbed by noise and 0 if the two signals are completely unrelated. The coherence value is >0 but <1 when 1) the two signals arise from common and uncommon sources, 2) noise is present in the system, and/or 3) the system relating the two signals is nonlinear.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Effect of midbrain and C1 transections on MAP, HR, and SND responses to heating. Twelve midbrain-transection (renal-splenic nerve pairs, n = 7; single renal nerves, n = 5) and twelve sham midbrain-transection (renal-splenic nerve pairs, n = 8; single renal nerves, n = 4) experiments were completed. C1 transections were completed in 14 experiments (renal-splenic nerve pairs, n = 6; single renal nerves, n = 7; single splenic nerve, n = 1), whereas sham C1 transections were completed in 9 experiments (renal-splenic nerve pairs, n = 6; single renal nerves, n = 3).

View larger version (11K): [in this window] [in a new window]   Fig. 1.   Mean arterial pressure (MAP) (A), heart rate (HR) (B), and renal (C) and splenic sympathetic nerve discharge (SND) (D) during pretransection control (Pre), posttransection recovery (Post and 38°C), and hyperthermia (39, 40, 41, and 41.7°C) in midbrain-transected and sham midbrain-transected rats. bpm, Beats/min; Tc, colonic temperature. * Significantly different from pretransection control (P < 0.05).  Significantly different from sham midbrain transection (P < 0.05).  Significantly different from posttransection recovery (P < 0.05).

View larger version (10K): [in this window] [in a new window]   Fig. 2.   MAP (A), HR (B), and renal (C) and splenic SND (D) during pretransection control (Pre), posttransection recovery (Post and 38°C), and hyperthermia (39, 40, 41, and 41.6°C) in C1-transected and sham C1-transected rats. * Significantly different from pretransection control (P < 0.05).  Significantly different from sham midbrain transection (P < 0.05).  Significantly different from posttransection recovery (P < 0.05).

Effect of midbrain and C1 transections on heating-induced changes in SND frequency components. Figure 3 shows SND (renal-splenic) and pulsatile arterial pressure before transection (left), during posttransection recovery (middle), and after heating to 41.6-41.7°C (right) from three representative experiments (A, midbrain transection; B, sham midbrain transection; C, C1 transection). MAP values recorded during each period are shown beneath the pulsatile arterial pressure traces. In all three examples shown in Fig. 3, renal and splenic SND contained bursts coupled to the arterial pulse (cardiac-related SND) in control. This pattern of activity persisted after midbrain and sham midbrain transection but was eliminated by C1 transection. Cardiac-related SND was converted to low-frequency bursts after heating to 41.7°C in both midbrain-transected and sham midbrain-transected rats. In contrast, heating in C1-transected rats did not change the basic pattern of SND, although the amplitude increased.

View larger version (32K): [in this window] [in a new window]   Fig. 3.   Traces of integrated SND bursts and pulsatile arterial pressure (AP) from 3 experiments in which simultaneous recordings of 2 sympathetic nerves (renal and splenic) were completed before (Pre) and after (Post) midbrain transection (A), sham midbrain transection (B), and C1 transection (C) and after heating to 41.6-41.7°C. Top 2 traces in each experiment show SND bursts, and bottom trace is pulsatile AP. MAP values for each period are shown below pulsatile AP traces. Horizontal calibration is 400 ms. Amplifier settings were the same for individual nerves in the renal-splenic pairs.

View larger version (30K): [in this window] [in a new window]   Fig. 4.   Renal (top) and splenic SND autospectra (middle) and renal-splenic coherence functions (bottom) constructed during pretransection control (Pre), posttransection recovery (Post), and after heating to 41.0 and 41.7°C in a midbrain-transected rat. Amplitudes of autospectra are autoscaled to highest peak. Frequency band is displayed from 0 to 15 Hz.

View larger version (30K): [in this window] [in a new window]   Fig. 5.   Renal (top) and splenic SND autospectra (middle) and renal-splenic coherence functions (bottom) constructed during pretransection control (Pre), posttransection recovery (Post), and after heating to 41.0 and 41.7°C in a sham midbrain-transected rat. Amplitudes of autospectra are autoscaled to highest peak. Frequency band is displayed from 0 to 15 Hz.

View larger version (30K): [in this window] [in a new window]   Fig. 6.   Renal (top) and splenic SND autospectra (middle) and renal-splenic coherence functions (bottom) constructed during pretransection control (Pre), posttransection recovery (Post), and after heating to 41.0 and 41.6°C in a C1-transected rat. Amplitudes of autospectra are autoscaled to highest peak. Frequency band is displayed from 0 to 15 Hz.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study determined the effect of progressive increases in T_c on renal and splenic SND bursts in midbrain-, C1-, and sham-transected (midbrain and spinal cord) rats. Our results provide experimental support for two findings that are central to understanding the organization of neural circuits responsible for changing SND frequency characteristics in response to acute heat stress. First, forebrain neural structures are not required for mediating changes in the level of activity or the frequency components of renal and splenic SND bursts in response to progressive hyperthermia. Second, acute heating in spinal cord-transected rats increased the level of renal and splenic sympathetic nerve activity but did not affect the pattern of SND bursts or the coupling between the discharges in these nerves, demonstrating that supraspinal neural circuits are required for mediating changes in SND frequency components to acute heat stress.

Although the preoptic area of the anterior hypothalamus is considered an important thermointegrative center of the brain (3, 11, 13, 14), thermoregulatory responses can occur in the absence of hypothalamic thermoregulatory integration. For example, spinal cord heating in decerebrate rabbits increases cardiac and splanchnic sympathetic nerve activity and reduces vasoconstrictor tone in the skin (16), and whole-body heating in chloralose-anesthetized, midbrain-transected rats increases renal and splenic nerve activity, HR, and MAP (current results). Consistent with these sympathetic nerve responses, skin blood flow increases, whereas intestinal blood flow is reduced during spinal cord heating in anesthetized dogs (23). A recent study by Shibata and Hashimoto (30) reported that neurons that tonically inhibit nonshivering thermogenesis are located in the midbrain and that the ability of these neurons to inhibit nonshivering thermogenesis does not require an intact hypothalamus. Important relative to the current results, brown adipose tissue is an important effector for nonshivering thermogenesis and is prominently regulated by the sympathetic nervous system (6, 7, 9, 28). Thermoregulatory responses (i.e., sweating and vasodilation) can be elicited in paraplegic humans in areas innervated by the spinal cord below the spinal lesion (5, 27). In addition, Berner and Heller (2) reported finding no cells in the preoptic anterior hypothalamic area that responded to changes in skin temperature without corresponding electroencephalographic changes. Taken together, these results support the view of Satinoff (29) who suggested that it is unlikely that the hypothalamus is the sole integrator of body temperature. Consistent with this concept, the results of the current study demonstrate that heating-induced changes in the frequency characteristics of SND do not require the participation of forebrain neural structures.

Although sympathoexcitatory responses to heating were reduced (renal SND) or tended to be reduced (splenic SND) in spinal cord-transected rats compared with sham spinal cord-transected rats, it is important to note that sympathetic nerve activity (renal and splenic) was increased from control levels during heating in C1-transected rats. This supports the concept that, in response to heating, spinal neural systems can activate the sympathetic nervous system in the absence of supraspinal neural circuits. On the other hand, the pattern of renal and splenic SND bursts remained unchanged and the renal and splenic SND bursts were only modestly coupled during heating in C1-transected rats, demonstrating that heating-induced changes in SND frequency components require the participation of supraspinal neural structures. The lack of change in the SND bursting pattern during heating in the C1-transected rats likely contributed to the attenuated sympathoexcitatory responses because we have previously reported (19) that SND pattern transformation is an important central neural strategy for increasing the level of sympathetic nerve activity in response to acute heat stress.

Results from previous studies (26, 34) demonstrate that in rats acute spinal transection decreases arterial pressure and lumbar SND but significantly increases renal and splenic nerve activity, indicating that spinal systems are capable of maintaining the activity in regionally selective nerves. In the current study, acute spinal transection did not significantly change the level of activity in the renal and splenic nerves but did reduce arterial pressure. Factors that may account for these disparate sympathetic nerve results remain to be determined, although differences in the anesthetic regimen can be discounted as -chloralose has been used in studies demonstrating either no change (current results) or increases in SND after acute spinal transection (34). In the present study, acute spinal transection significantly reduced peak and mean renal-splenic coherence values, supporting the previous findings of Taylor and Schramm (34) who reported that brain stem systems play a key role in synchronizing the discharges in regionally selective sympathetic nerves. It should be noted that in the current study acute midbrain transection significantly increased splenic SND and tended to increase renal SND from pretransection control values, suggesting the release of inhibitory forebrain influences on efferent sympathetic nerve outflow. However, because baroreceptor-innervated rats were used in the present experiments and because midbrain transection tended to reduce arterial pressure, the renal and splenic sympathoexcitation during the posttransection recovery period may have been reflexly mediated secondary to unloading of the arterial baroreceptors.

We have previously established (19) that pharmacological blockade of ganglionic transmission (produced by trimethaphan administration) during hyperthermia reduces arterial pressure to values less than those produced by ganglionic blockade during the preheating control period, suggesting that the sympathetic nervous system plays a key role in offsetting vasodilatory influences during progressive increases in T_c. In addition and as stated previously (19), hyperthermia-induced changes in the SND bursting pattern directly contribute to increasing sympathetic nerve activity. These results suggest that the sympathetic nervous system is a prominent effector in arterial blood pressure regulation during heating by providing increased blood flow distribution for heat dissipation while maintaining arterial blood pressure (i.e., vital organ perfusion pressure). In this regard, the results of the current study suggest that brain stem neurons, in the absence of forebrain neural structures, are capable of mediating thermoregulatory effector responses to acute heat stress.

Perspectives