We recorded changes in right inferior cardiac and either left inferior cardiac or left vertebral sympathetic nerve discharge (SND) produced by unilateral microinjections of GABA-A and excitatory amino acid (EAA) receptor antagonists into the ventrolateral medulla (VLM) of urethane-anesthetized, baroreceptor-denervated cats. Unilateral microinjections of GABA-A receptor antagonists, SR-95531 or bicuculline, into single tracks in VLM anywhere between 1 and 5 mm rostral to the obex eliminated or markedly reduced 10-Hz power in SND on both sides of the body. Low-frequency components (<6 Hz) of SND were unaffected. Complete blockade of the 10-Hz rhythm occurred with a dose of SR-95531 as low as 6.25 pmol in a 50-nl volume. Unilateral microinjections of the nonselective EAA receptor antagonist, kynurenate (KYN; 7.5 nmol), into the caudal or rostral VLM significantly reduced, but did not eliminate, 10-Hz SND ipsilateral to the injection sites, while 10-Hz SND contralateral to the injection sites was not significantly changed. These observations suggest that 1) GABAergic transmission in VLM is critical for generation of the 10-Hz rhythm, 2) the caudal and rostral portions of VLM act together to generate the 10-Hz rhythm, and 3) 10-Hz rhythm generation depends, at least in part, on tonic or phasic excitatory drive to GABAergic interneurons in caudal VLM and presympathetic neurons in rostral VLM. The data also suggest that pathways interconnecting the two halves of the brain stem play an important role in promoting 10-Hz rhythm generation.
- GABAergic transmission
- sympathetic rhythms
a 10-hz rhythm is commonly seen in the discharges of sympathetic nerves with cardiovascular targets in urethane-anesthetized, decerebrate, or conscious cats (3–9, 14, 15, 21–24, 32, 35, 47–49). Although the 10-Hz rhythm in sympathetic nerve discharge (SND) is most evident after baroreceptor denervation, it can occur in combination with the cardiac-related rhythm in baroreceptor-innervated cats (4, 7, 14, 32, 35). The 10-Hz rhythm is functionally relevant because its sudden appearance in SND is accompanied by an abrupt increase in arterial pressure (5). Because the 10-Hz discharges of sympathetic nerve pairs are much more strongly correlated than their aperiodic discharges (6, 21, 48), the rhythm also serves the purpose of coordinating the discharges of sympathetic nerves with different targets. Moreover, the 10-Hz rhythm has been implicated in the differential pattern of spinal sympathetic outflow that occurs during the defense reaction in baroreceptor-denervated cats (22, 27).
The 10-Hz rhythm in SND persists after midcollicular decerebration (6, 14, 24) but is eliminated by high spinal cord transection (24, 48). Moreover, medullary 10-Hz rhythmic field potentials that are strongly correlated to the 10-Hz rhythm in SND in decerebrate cats persist after cervical spinal cord transection (21). These observations support the view that the 10-Hz rhythm reflects the fundamental organization of a brain stem network responsible for a significant proportion of the basal discharges of sympathetic nerves.
Single neurons with naturally occurring discharges correlated to the 10-Hz rhythm in SND have been identified in three lower brain stem regions: caudal and rostral portions of the ventrolateral medulla (VLM) and medullary raphe nuclei (3, 4, 7, 8, 15). As demonstrated with antidromic activation, many of these neurons in the rostral VLM and medullary raphe complex send their axons to the intermediolateral nucleus (IML) of the thoracic spinal cord (4). Those in the caudal VLM do not have spinal projections (8). Microinjection of the GABA-A receptor agonist, muscimol, into any one of these regions reduces or eliminates the 10-Hz rhythm in SND (7, 49). Nonetheless, information on the extent and nature of the interconnections among these cell groups remains precursory, as does the issue of which group(s) comprise the rhythm generator and which comprise the follower circuits.
The current study was designed to shed light on the source and synaptic transmitters involved in 10-Hz rhythm generation. As such, we studied the effects on the 10-Hz rhythm in SND produced by unilateral microinjections of GABA-A receptor and excitatory amino acid (EAA) receptor antagonists into the VLM while recording simultaneously from postganglionic nerves on both sides of the body. The bilateral recordings allowed us to estimate the degree to which rhythm-generating networks located in the left and right halves of the brain stem are interdependent. Regarding this issue, as demonstrated by antidromic mapping, bulbospinal inputs to preganglionic neurons in the thoracic IML are primarily uncrossed in the cat (1, 2).
The protocols used in these studies on 41 adult cats (3.33 ± 0.86 kg) were approved by Michigan State University's Institutional Animal Care and Use Committee. Cats were initially anesthetized with 2.5% isoflurane mixed with 100% O2. The right femoral artery and left and right femoral veins were cannulated to measure arterial pressure and to administer drugs, respectively. Urethane (1.0–1.8 g/kg iv) was then administered, and isoflurane inhalation was stopped. This dose of urethane has been shown to maintain a surgical level of anesthesia for 8–10 h in cats (19). This time period exceeded the duration of the experiments.
Cats were placed in a stereotaxic apparatus, paralyzed (gallamine triethiodide, 4 mg/kg iv, initial dose), pneumothoracotomized, and artificially respired with room air. The parameters of artificial ventilation (40.5 ± 2.8 cc and 21.0 ± 3.9 cycles/min) were within the physiological range (17) so that end-tidal CO2 was held near 4.6%. Rectal temperature was kept near 38°C with a heat lamp. Before neuromuscular blockade, the adequacy of anesthesia was indicated by the absence of a palpebral reflex. When cats were paralyzed, an adequate level of anesthesia was indicated by the inability of noxious stimuli (pinch, heat, surgery) to increase arterial pressure or change the pattern of SND.
The carotid sinus, aortic depressor, and vagus nerves were sectioned bilaterally. Two observations verified the completeness of baroreceptor denervation in these experiments. First, there was no sharp peak in the autospectrum of SND at the frequency of the heart beat, and the coherence value relating SND to the arterial pulse wave was <0.1 at this frequency. Second, SND was not reflexly inhibited during the pressor response produced by an injection of norepinephrine bitartrate (1–2 μg/kg iv).
As described in other reports from this laboratory (6, 21, 48), the inferior cardiac and vertebral branches of the stellate ganglia were exposed retropleurally by removing the head of the first rib. These nerves project to the heart and vasculature of the forelimb, respectively. In most experiments, recordings were made from two nerves (right inferior cardiac nerve and either the left inferior cardiac or left vertebral nerve). Potentials were recorded monophasically from the central ends of the cut nerves placed on platinum bipolar electrodes. The capacity-coupled preamplifier bandpass was set at 1–1,000 Hz so that the synchronized discharges of sympathetic fibers appeared as slow waves or envelopes of spikes (20).
All chemicals used for microinjection were diluted in PBS. Solutions were adjusted to a pH of 6–8 (litmus paper test) and placed in a glass micropipette (∼40-μm tip diameter) that was glued (cyanoacrylate) to the needle of a 5-μl Hamilton syringe and mounted on a microinjection unit (David Kopf Instruments, model 5000). A 50-nl injection was made slowly (∼10 s) at each medullary site (see below) by turning the calibrated micrometer on the microinjection unit. The following drugs were injected into the medulla: the GABA-A receptor antagonist bicuculline (BIC; 1.0 mM) or SR-95531 (0.125–1.0 mM) and the nonselective EAA receptor antagonist kynurenate (KYN; 150 mM). All drugs were purchased from RBI Sigma (St. Louis, MO). These drug concentrations are within the range used in past studies from this and other laboratories (30, 31, 33, 38, 40–42).
The dorsal surface of the brain stem was exposed by removing portions of the occipital bone and cerebellum. The midline, obex, and dorsal medullary surface were used as landmarks for placement of the micropipette. The micropipette was positioned into the VLM in tracks located between 0 and 7 mm rostral to the obex and 3.8–4.2 mm lateral to the midline. Microinjections of BIC (50 pmol) or SR-95531 (6.25–50 pmol) were made at two depths (4.0 and 5.5 mm below the dorsal surface) in a single track through the VLM on the left (22 cats) or right (3 cats) side of the neuraxis. Microinjections of KYN (7.5 nmol) were made at two depths (4.0 and 5.5 mm from the dorsal surface) in two tracks through either the rostral portion of the VLM (4.5 and 5.5 mm rostral to the obex) or the caudal portion of the VLM (1.5 and 2.5 mm rostral to the obex) on the left side. In three cats, BIC (50 pmol) was microinjected into the midline medullary raphe nuclei at three levels (2, 3, and 4 mm rostral to the obex) at depths of 3.5 and 5.0 mm below the dorsal surface. Injection sites were in the regions of VLM and raphe, from which we have recorded from neurons with activity correlated to the 10-Hz rhythm in SND (3, 4, 7, 8).
Data were continuously collected for a minimum of 2 min before microinjection of a drug and for at least 15 min after completing the injections. Data were then collected again at 30-min intervals until partial or full recovery of SND. Some cats were used for more than one set of injections. For example, if one set of injections did not affect SND, a second set was placed at another level of the VLM after waiting at least 20 min. If a drug produced a change in SND, we waited for recovery from the first set of injections before injecting the same drug at another level of the VLM. As a control, vehicle (PBS) was injected into the VLM of three cats at sites where SR-95531 microinjections eliminated the 10-Hz rhythm in SND. Injections of PBS did not change SND or arterial pressure.
Data (5-ms sampling period) were acquired with a Digidata1322A digitizer (Axon Instruments; Union City, CA). Fast Fourier transform was performed on 47 5-s windows of data with 50% overlap (2-min data block) to construct autospectra and coherence functions. Spectral analyses were done over a frequency band of 0 to 100 Hz with a resolution of 0.2 Hz/bin but displayed on a scale of 0 to 20 Hz, which contains essentially all of the power in SND (6). The autospectrum of a signal shows how much power (voltage squared) is present at each frequency. The coherence function (normalized cross spectrum) is a measure of the strength of linear correlation of two signals at each frequency. The squared coherence value (referred to as coherence value) is 1.0 in the case of a perfect linear relationship and 0 if two signals are unrelated. A coherence value ≥0.1 reflects a statistically significant relationship when ≥32 windows are averaged (11).
ASCII files of the spectra were saved for transfer to spreadsheet, graphics, and statistical programs (GraphPad Prism version 4.03 for Windows and GraphPad Instat, GraphPad Software, San Diego CA). The autospectra of SND before and after microinjection of a drug were displayed on the same power scale. The 10-Hz band of SND is defined as the range of frequencies surrounding the sharp peak in the autospectrum of SND near 10 Hz. As described by Orer et al. (36), a macro written in Microsoft Excel version 7.0 was used to measure 10-Hz power. Briefly, 10-Hz power was calculated as the area above a line that connected the left and right limits of the 10-Hz band in the autospectrum of SND. Low frequency (<6 Hz) power was calculated by arithmetically summing the values for the bins in the 0- to 6-Hz range. This is the component of SND that can become entrained to the arterial pulse when baroreceptor afferent nerves are intact (5, 14, 20, 36). Total power in SND refers to the arithmetic sum of the values for the bins in the 0- to 20-Hz frequency range.
Values in the text and figures are means ± SE. A paired t-test was used to evaluate the effects of microinjection of a drug on absolute values of 10 Hz, <6 Hz, and total power in SND, coherence value relating the discharges of two sympathetic nerves, and mean arterial pressure (MAP). An unpaired t-test was used to compare changes in SND ipsilateral and contralateral to the injection sites. P < 0.05 indicated statistical significance. Raw values of power were used for statistical analyses, but changes in SND are expressed as a percentage of control in the text and figures.
The brain stem was removed and fixed in 10% buffered formalin. Frontal sections of 30-μm thickness were cut and stained with cresyl violet. Sites of microinjection were identified with reference to the bottom of the tracks made with the micropipette and the stereotaxic planes of Berman (12). The actual injection sites were within 200 μm of the target sites in the VLM.
Effects of unilateral blockade of GABA-A receptors in the VLM.
Figures 1 and 2 show representative results from an experiment in which we microinjected 25 pmol of SR-95531 at depths of 4.0 and 5.5 mm below the dorsal surface in the VLM at a level 2 mm rostral to the obex on the left side of the neuraxis. Figure 1A shows that arterial pressure and SND began to decrease shortly after the injections. The changes in the frequency composition of left vertebral nerve (lVN) and right inferior cardiac nerve (rCN) activities are shown by the oscillographic traces in Fig. 1, B and C, respectively, and the autospectra in Fig. 2, top and middle. SND contained a prominent 10-Hz rhythm before microinjection of the GABA-A receptor antagonist into the VLM (Fig. 1, B1 and C1, and the solid black traces in Fig. 2). The 10-Hz component of SND was reduced almost immediately after completing the injections of SR-95531 as shown by both the oscillographic traces of lVN and rCN discharges recorded 60 s after the injections (Fig. 1, B2 and C2) and the marked decrease in power in the 10-Hz band of the lVN and rCN autospectra for data collected 30–150 s after the injections (dotted traces in Fig. 2). The residual 10-Hz activity in these two nerves remained strongly correlated as shown by the similarity of the lVN-rCN coherence functions (Fig. 2, bottom) before (solid black trace) and after (dotted trace) the injections. By 3 min after the microinjections of SR-95531, the 10-Hz rhythm was eliminated (Fig. 1, B3 and C3 and gray traces in Fig. 2). The loss of the 10-Hz rhythm is also evidenced by the absence of a peak in the lVN-rCN coherence value at this frequency (gray trace in Fig. 2, bottom). In this experiment, low-frequency (<6 Hz) power was reduced, but only in the lVN (ipsilateral to the injection site).
Figure 3 shows that the effects of unilateral microinjections of a GABA-A receptor antagonist on the 10-Hz rhythm in SND were reversible. In this example, the 10-Hz rhythm in lCN and rCN activity was eliminated by microinjection of 6.25 pmol SR-95531 at 4.0 and 5.5 mm below the dorsal surface in the left VLM at 3.5 mm rostral to the obex. Note that the peak near 10 Hz in the autospectra of lCN and rCN activity seen during control (solid black trace) was absent 4 min after (gray trace) the injection (Fig. 3, top and middle). Also, after GABA-A receptor blockade in the VLM, there was no longer a sharp peak near 10 Hz in the lCN-rCN coherence function (Fig. 3, bottom). SND power at frequencies <6 Hz was essentially unchanged. Importantly, the 10-Hz rhythm was again prominent in the autospectra of SND 45 min later, and the coherence function again showed a strong relationship between lCN and rCN activities at this time (dotted traces in Fig. 3). In this experiment, the peak frequency of the rhythm was somewhat lower than in control after recovery from the effects of GABA-A receptor blockade.
By making two injections in single tracks from the obex to 7 mm rostral to the obex (near the pontomedullary border), we were able to identify the levels of the VLM that are critical for maintenance of the 10-Hz rhythm in SND. Figure 4A shows that microinjections of 50 pmol of SR-95531 into the left VLM at the level of the obex had little effect on the 10-Hz rhythm in lVN and rCN activities. Specifically, the peaks near 10 Hz in the autospectra of lVN and rCN activities (Fig. 4A, top and middle) ∼5 min after microinjection of SR-95531 (dotted traces) were only modestly smaller than those in the control autospectra (solid black traces). However, as shown by the gray traces in Fig. 4A, top and middle, microinjection of the same dose of SR-95531 into the VLM at 1.0 mm rostral to the obex markedly reduced the 10-Hz power in both lVN and rCN discharges. The residual 10-Hz activity in these two nerves remained strongly correlated (gray trace in Fig. 4A, bottom). SND at frequencies <6 Hz was little affected by the injections of SR-95531.
In the example shown in Fig. 4B, microinjections of 50 pmol of SR-95531 into the left VLM at 7.0 mm rostral to the obex caused little or no change in the 10-Hz rhythmic discharges of the lCN and rCN. Note the similarity in the autospectra before (solid black traces) and 5 min after (dotted traces) microinjection of the GABA-A receptor antagonist in Fig. 4B, top and middle. However, the 10-Hz rhythm was essentially abolished when the same dose of SR-95531 was injected into the VLM at 5.0 mm rostral to the obex (gray traces). Not surprisingly, there was no longer a peak in the 10-Hz band of the lVN-rCN coherence function (gray trace in Fig. 4B, bottom). Note that the low-frequency power in SND was little affected. Although not shown, a 10-Hz rhythm reappeared in SND about 2 h later.
The histological sections on the left side of Fig. 5, top and bottom, show the tracks made by the micropipette whose tip was in the left VLM at 5.0 and 1.0 mm rostral to the obex in the experiments illustrated in Fig. 4, B and A, respectively. Figure 5, middle, is from another cat in which microinjections of SR-95531 into the VLM at 3.5 mm rostral to the obex eliminated the 10-Hz rhythm in lCN and rCN activities.
We quantified the effects of microinjection of 6.25–50 pmol of SR-95531 (n = 20 cats) or 50 pmol of BIC (n = 5 cats) at two depths (4.0 and 5.5 mm below the dorsal surface) in a single track through the VLM in 25 baroreceptor-denervated cats in which there was a prominent 10-Hz rhythm in SND. The higher doses of SR-95531 were used in the initial series of experiments in which we identified the levels of the VLM that are critical for the maintenance of the 10-Hz rhythm in SND. We then lowered the dose to 25 and then to 6.25 pmol in later experiments. The main difference was that recovery occurred more rapidly with the lower doses. Whereas recovery from the effects produced by a dose of 6.25 pmol occurred within 1 h, recovery after microinjection of a 50-pmol dose took 2 or 3 h. Figure 6 summarizes the changes in SND ipsilateral (A) and contralateral (B) to the microinjections of GABA-A receptor antagonists at different levels of the VLM. The results obtained by microinjection of SR-95531 and BIC were comparable, and thus the data were pooled. Also, the changes in lCN and lVN were comparable; thus these data were also pooled. Power in the 10-Hz band of SND was significantly attenuated after blockade of GABA-A receptors in the VLM anywhere between 1.0 and 5.0 mm rostral to the obex. Importantly, the reductions in 10-Hz power were comparable in nerves located ipsilateral and contralateral to the injection sites. The injection sites at these levels of the VLM are depicted by the dots in the schematics on the right side of Fig. 5. The 10-Hz rhythm in SND was eliminated in 17 of 27 cases ipsilateral and in 17 of 26 cases contralateral to the injection sites between 1.5 and 5.0 mm rostral to the obex. Figure 6 also shows that SR-95531 and BIC selectively affected the 10-Hz band of SND, as there was not a significant change in power at frequencies <6 Hz. Total power in SND ipsilateral to the injection sites was either significantly reduced or had a tendency to be reduced after GABA-A receptor blockade between 1.0 and 5.0 mm rostral to the obex. Total power in SND contralateral to the injection sites was significantly reduced after GABA-A receptor blockade between 1.0 and 3.5 mm rostral to the obex.
Figure 7 summarizes the changes in MAP and the coherence values relating the 10-Hz discharges of sympathetic nerve pairs produced by unilateral microinjections of GABA-A receptor antagonists at different levels of the VLM. Before the injections into the VLM, MAP averaged 115 ± 3 mmHg, and the peak coherence value relating the 10-Hz discharges of SND averaged 0.92 ± 0.01. MAP was significantly reduced after microinjections of SR-95531 or BIC at 2.5 or 3.0 mm rostral to the obex (Fig. 7A). Coherence values were significantly reduced when these drugs were injected between 2.5 and 4.5 mm rostral to the obex (Fig. 7B). This reflects the relatively large number of cases of elimination of the 10-Hz rhythm after microinjections of SR-95531 or BIC at these levels of the VLM. It should be noted that even if only a small amount of 10-Hz activity persisted after GABA-A receptor blockade, the coherence value relating this component of the discharges of two nerves remained high (for example, see Fig. 4A). In those cases in which there was some 10-Hz activity remaining, on the average, blockade of GABA-A receptors in the VLM did not affect the peak frequency of this component of SND (10.89 ± 0.25 Hz before and 10.90 ± 0.30 Hz after microinjection of SR-95531 or BIC).
The effects of GABA-A receptor blockade in the rostral portion of the VLM on SND and MAP in these baroreceptor-denervated cats differ considerably from those noted in baroreceptor-innervated cats in which SND contains a cardiac-related rhythm. In agreement with others (16, 39, 45), we have shown that microinjections of BIC or SR-95531 bilaterally into the rostral VLM of baroreceptor-innervated cats leads to an increase in SND and MAP as well as a loss of the cardiac-related rhythm in SND (S. M. Barman, G. L. Gebber, and H. S. Orer, unpublished observations).
Effects of blockade of GABA-A receptors in the medullary raphe.
To determine whether GABA-A receptor-mediated transmission in the medullary raphe plays a role in generating the 10-Hz rhythm in SND, we microinjected BIC (50 pmol) into the medullary raphe complex of three cats (see methods for injection sites). In no case was power in the 10-Hz band of SND decreased. In two of the cats, 10-Hz power in lCN and rCN activity was increased to between 135 and 355% of control. In the other cat, 10-Hz power was essentially unchanged (95 and 102% of control for rCN and lCN, respectively). Thus, GABA-A receptor-mediated transmission in the raphe is not critical for 10-Hz rhythm generation.
Effects of unilateral blockade of EAA receptors in the VLM.
We quantified the changes in 10-Hz SND produced by unilateral microinjections of the nonselective EAA receptor antagonist KYN into the rostral or caudal VLM. As shown in Fig. 8, unilateral microinjections of KYN into either the caudal (A) or rostral (B) portions of the VLM reduced 10-Hz power in lCN and rCN activities; however, the reduction was more pronounced for the lCN (ipsilateral to the injection sites). This is evident by comparing the autospectra of SND before (black traces) and 5 min after (gray traces) KYN (7.5 nmol) was injected into the caudal VLM at 1.5 and 2.5 mm rostral to the obex (Fig. 8A, top and middle) or into the rostral VLM at 4.5 and 5.5 mm rostral to the obex (Fig. 8B, top and middle). lCN-rCN coherence values were essentially unchanged even though 10-Hz power was reduced (Fig. 8, A and B, bottom).
Figure 9 summarizes the results from the experiments in which KYN was microinjected into the caudal (n = 10) and rostral (n = 6) portions of the left VLM. Recordings were made from the lCN in all of these experiments, and rCN activity was recorded in seven cats in which injections were made into the caudal VLM (1.5 and 2.5 mm rostral to the obex) and in three cats in which injections were made into the rostral VLM (4.5 and 5.5 mm rostral to the obex). On a group basis, the reductions in 10-Hz power in lCN activity produced by microinjections of KYN into the left caudal or rostral VLM were statistically significant. Although there was a tendency for 10-Hz power in rCN activity to be reduced, the change was not statistically significant. In no case did microinjections of KYN into the VLM eliminate the 10-Hz rhythm in either lCN or rCN activity. Total power in SND was significantly reduced ipsilateral to the injection sites in the rostral but not caudal VLM. Low-frequency (6-Hz) power was not significantly changed by microinjections of KYN into the caudal or rostral VLM. Table 1 shows that there was not a significant change in the frequency of the peak within the 10-Hz band of SND, the peak coherence value relating the 10-Hz discharges in lCN and rCN, or MAP when KYN was injected unilaterally at two levels within the caudal or rostral VLM.
The current study is the first to demonstrate that the 10-Hz rhythm in SND is dependent on GABA-A receptor-mediated synaptic transmission in the VLM. The fact that microinjections of SR-95531 or BIC placed in single tracks at widely separated levels of the VLM (anywhere between 1.5 and 5.0 mm rostral to the obex) could eliminate 10-Hz SND suggests that the caudal and rostral portions of the VLM act together to generate the 10-Hz rhythm in SND. Strikingly, unilateral blockade of GABA-A receptors in the VLM eliminated the 10-Hz discharges in sympathetic nerves on both sides of the body. This was indicated by the disappearance of the sharp peak near 10 Hz in the autospectra of the discharges of the nerves and in the coherence function relating their discharges. These data support the view that communication across the midline of the brain stem is, in some way, involved in 10-Hz rhythm generation. Regarding this point, we (47) have reported that midsagittal section of the medulla essentially eliminated the 10-Hz rhythm in SND on both sides of the body. In contrast, power in the 10-Hz band of SND was essentially unchanged by midsagittal section of the thoracic or cervical spinal cord (48).
Regarding the contention that the caudal and rostral portions of the VLM act together to generate the 10-Hz rhythm, Lipski et al. (29) and Nicholson (34) have estimated that a 30–50 nl injectate spreads to form a sphere with a radius of <0.5 mm. On this basis, it is unlikely that elimination of the 10-Hz rhythm in SND after unilateral microinjections of small doses (as low as 6.25 pmol in 50 nl) of SR-95531 into either the caudal and rostral VLM at levels separated by as much as 3.5 mm can be explained by widespread diffusion of the GABA-A receptor antagonist.
It is also unlikely that elimination of the 10-Hz rhythm in SND after microinjection of SR-95531 or BIC into the VLM can be explained by physical injury caused by placement of the micropipette or nonspecific effects of these drugs. First, microinjection of vehicle (PBS) into the VLM did not change 10-Hz SND. Second, block of the 10-Hz rhythm by SR-95531 or BIC was reversible, and the time course of recovery after SR-95531 was concentration dependent. It should be stressed that elimination of 10-Hz rhythm in SND bilaterally was observed even with a dose of SR-95531 as low as 6.25 pmol. Third, rostral VLM-spinal neurons transmit not only the 10-Hz rhythm, but also to the lower-frequency aperiodic activity to preganglionic sympathetic neurons (2, 4). Yet, power at frequencies <6 Hz in SND was essentially unaffected by microinjections of GABA-A receptor antagonists into the VLM. Thus, it is unlikely that elimination of the 10-Hz rhythm by SR-95531 or BIC can be explained by nonspecific depressant effects of these drugs.
Using the Golgi staining technique, Valverde (44) described neurons in the VLM that are distinct from long-axoned cell types prevalent in the medial reticular formation or short-axoned Golgi type II cells. The neurons described by Valverde possess axons of intermediate length (1–2 mm) that run rostrally with numerous collateral branches radiating in all directions for distances less than the length of the main axon. Although direct evidence is lacking, we suggest that a network of intermediate-axoned cells of the type identified in the VLM by Valverde is involved in generating the 10-Hz rhythm in SND. Specifically, we propose that 10-Hz rhythm generation involves periodic synchronization of the discharges of such interneurons in the caudal VLM via their GABA-A receptor-mediated synaptic interconnections. Because the axons of the neurons described by Valverde ascend rostrally for relatively short distances, it is not difficult to imagine that interruption of GABA-A receptor-mediated synaptic transmission at any level of the caudal VLM would desynchronize their firing times, thereby eliminating or at least attenuating the 10-Hz discharges appearing in population recordings from sympathetic nerves. We further suggest that the chain of intermediate-axoned GABAergic interneurons in the caudal VLM imposes the 10-Hz rhythm onto rostral VLM-spinal sympathoexcitatory neurons (4), thus effecting the transfer of this activity pattern to peripheral sympathetic nerves. The fact that caudal VLM neurons with 10-Hz related activity located close to the obex cannot be antidromically activated from the rostral VLM (8) is consistent with this scheme. The model proposed here for the 10-Hz rhythm in SND is similar to those used by others to explain γ-rhythm generation in the mammalian hippocampus (13, 46) and locust olfactory system (28), as well as “spindling” in the neocortex (43). For each of these systems, it has been proposed that the rhythm is generated by a network of synchronously discharging GABAergic interneurons that project to long-axoned output neurons (i.e., principal cells).
In contrast to the results obtained with unilateral GABA-A receptor blockade in the VLM, unilateral microinjections of the nonselective EAA receptor antagonist, KYN, into the caudal or rostral VLM significantly reduced power in the 10-Hz band of SND only on the side ipsilateral to the injections sites. In no case was 10-Hz SND eliminated after unilateral microinjections of KYN. These observations suggest that tonic and/or phasic excitatory inputs to GABAergic interneurons in the caudal VLM and presympathetic neurons in the rostral VLM play a permissive role in 10-Hz rhythm generation. This possibility is also supported by the results obtained in another report from our laboratory (9), in which bilateral microinjections of either an N-methyl-d-aspartate (NMDA) or a non-NMDA receptor antagonist into either the rostral or caudal VLM profoundly reduced 10-Hz SND on both sides of the body. The injection sites were the same as those at which KYN was applied unilaterally in the current study. Interestingly, the dependence of γ-rhythm generation by networks of hippocampal GABAergic interneurons on tonic and phasic excitatory inputs has been demonstrated experimentally and in modeling studies (see review by Bartos et al. ).
Over 100 years ago, Langendorff (26) found that breathing movements in two halves of the diaphragm persisted but were no longer synchronized after midsagittal section of the medulla of vagotomized rabbits. Such has been confirmed in other species, including the cat [see review by von Euler (45a)]. These studies indicate that networks located on each side of the medulla are independently capable of producing rhythmic breathing. The situation appears to be different for the 10-Hz rhythm in SND. This is implied by elimination of the 10-Hz rhythm bilaterally after unilateral injections of SR-95531 or BIC into the VLM. Thus, communication across the midline of the brain stem apparently is required for generation and distribution of the rhythm to sympathetic nerves on both sides of the body. The exact role played by these cross-connections remains elusive. Communication across the midline might serve to interconnect elements of a common 10-Hz oscillator that are located bilaterally. Alternatively, 10-Hz rhythm generation may occur in networks self-contained on each side of the brain stem, but only in the presence of phasic or tonic excitatory drive emanating from neurons located across the midline in the brain stem. Such drive might be derived directly or indirectly from the contralateral rhythm generator. Earlier work from our laboratory supports a role for medullary raphe neurons in 10-Hz rhythm generation. Regarding this possibility, single neurons with naturally occurring activity correlated to 10-Hz SND have been located throughout the medullary raphe complex (3, 4), and microinjection of muscimol or 8-hydroxy-2-(di-n-propylamino) tetralin into the medullary raphe significantly reduces power in the 10-Hz band of SND (37, 49). Furthermore, raphe neurons with 10-Hz related activity can be antidromically activated by stimuli applied to sites in the caudal VLM (8). Likewise, caudal VLM neurons with 10-Hz-related activity can be antidromically activated by medullary raphe stimulation (8). Whether raphe inputs to the VLM simply promote 10-Hz rhythm generation by local circuits in the VLM or, in combination with VLM neurons, comprise an anatomically distributed rhythm generator is unclear at the present time.
Rostral VLM-spinal neurons have cardiac-related discharges when systemic arterial pressure is above the threshold needed to induce pulse-synchronous baroreceptor afferent nerve activity (4). The same neurons have activity correlated to the 10-Hz rhythm in SND when arterial pressure is lowered to minimize baroreceptor nerve activity (4). In contrast to these bulbospinal neurons, caudal VLM neurons with activity correlated to the 10-Hz rhythm in SND do not develop cardiac-related activity at high arterial pressure levels (7). This suggests that caudal VLM neurons with 10-Hz-related activity are different from those in the caudal VLM known to relay inhibitory baroreceptor-mediated influences to sympathoexcitatory rostral VLM-spinal neurons via the release of GABA (25).
In summary, the results of the current study are consistent with the view that GABAergic transmission in the VLM is critical for generation of the 10-Hz rhythm in SND and that the caudal and rostral portions of the VLM act together to generate the rhythm. Moreover, the data point to the importance of communication across the midline of the brain stem in promoting rhythm generation. Future investigations are needed to determine whether the 10-Hz rhythm generators are self-contained in the VLM or comprise interactive neuronal populations distributed in more than one brain stem region.
The authors thank Joseph Prinsen for assistance with preparation of histological sections. This study was supported by National Institutes of Health Grant HL33266.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society