Neurons within the dorsomedial hypothalamus (DMH) play a critical role in subserving the cardiovascular and neuroendocrine response to psychological stress. An increase in respiratory activity is also a characteristic feature of the physiological response to psychological stress, but there have been few studies of the role of DMH neurons in regulating respiratory activity. In this study we determined the effects of activation of DMH neurons on respiratory activity (assessed by measuring phrenic nerve activity, PNA) and the relationship between evoked changes in respiratory activity and changes in sympathetic vasomotor activity in spontaneously breathing urethane-anesthetized rats. Microinjections of bicuculline (4–40 pmol in 20 nl) into the DMH evoked dose-dependent increases in PNA burst frequency and amplitude. These were accompanied by dose-dependent decreases in mean tracheal CO2 levels, indicative of hyperventilation. In control experiments, microinjections of bicuculline into sites adjacent to the DMH evoked much smaller or no changes in PNA. In experiments where renal sympathetic nerve activity (RSNA) was also measured, cycle-triggered averaging revealed that RSNA under resting conditions was partly correlated with the PNA, but in response to DMH disinhibition there was no consistent change in the amplitude of the respiratory-related variations in RSNA. The results indicate that DMH neurons can exert a powerful stimulatory effect on respiratory activity, causing hyperventilation. This is not associated with an increase in the degree of coupling between PNA and RSNA, indicating that the DMH-evoked increase in RSNA is not a consequence of increased central respiratory drive.
- arterial pressure
- renal sympathetic nerve activity
- heart rate
- psychological stress
acute stress and anxiety have long been associated with an increase in respiratory drive (for review see Ref. 12). Acute psychological stressors in animals also generate a highly coordinated range of cardiovascular, neuroendocrine, and behavioral responses (6, 26, 32). The dorsomedial hypothalamus (DMH), defined as the region that includes the dorsomedial hypothalamic nucleus (DMN) but also extending dorsally into the dorsal hypothalamic area (DA) (7), has been shown to play a critical role in mediating these responses. For example, inhibition of the DMH with muscimol dramatically reduces stress-induced cardiovascular, neuroendocrine, and behavioral responses (26, 29, 30). In particular, the DMH is believed to play a critical role in mediating panic attacks, because intravenous infusion of sodium lactate, which is known to trigger panic attacks in many humans with panic disorder (10), also triggers a similar paniclike response in conscious rats in which GABA synthesis in the DMH has been chronically disrupted (27). In anesthetized rats, disinhibition of neurons in the DMH by microinjection of bicuculline (a GABAA receptor antagonist) produces cardiovascular and neuroendocrine responses similar to those that occur during naturally evoked stress responses (6).
Although hyperventilation is a characteristic feature of responses to acute stress (12), including panic attacks (28), there have been few studies of the role of DMH neurons in generating changes in central respiratory drive. Shekhar (26) found that microinjection of the GABA receptor antagonist bicuculline into the DMH of pentobarbitone-anesthetized rats resulted in an increase in respiratory frequency, with a maximum increase of ∼50% above the resting level. In that study, however, the amount injected (50 pmol in 250 nl) was rather large, and no observations were made of any indicator of respiratory activity other than frequency.
It has also been shown that injections of GABA receptor antagonists into regions close to the DMH can evoke increases in respiratory rate. In particular, DiMicco and Abshire (5) first showed that injections of bicuculline, albeit in a large volume (1 μl), into the posterior hypothalamic nucleus of anesthetized rats evoked dose-dependent increases in respiratory frequency. Similar observations were also made in anesthetized cats by Waldrop et al. (33). More recently, it also was shown that microinjection of bicuculline (50 pmol) into the paraventricular nucleus (PVN) of conscious rats resulted in a marked increase (>100%) in respiratory frequency and a more moderate increase in tidal volume (24). Consistent with these observations, Yeh et al. (34) found that microinjection of glutamate into the PVN of anesthetized rats also evoked increases in respiratory activity, as indicated by the frequency and amplitude of diaphragm electromyographic activity. Furthermore, there are direct projections from PVN neurons to the phrenic nucleus in the spinal cord as well as to respiratory nuclei in the medulla (16, 34).
The main goal of this study was to characterize more fully the effects of disinhibition of neurons within the DMH on respiratory activity, which was assessed by measuring phrenic nerve activity (PNA). Second, we tested whether the respiratory effects of bicuculline microinjections into the DMH were due to disinhibition of neurons within surrounding regions, rather than the DMH itself. Finally, it is known that increases in central respiratory activity can powerfully influence sympathetic vasomotor activity (8, 13) and also that disinhibition of DMH increases sympathetic activity (1, 9, 15). Therefore, we also determined the extent to which sympathetic activity is modulated by respiratory activity when DMH neurons are activated.
MATERIALS AND METHODS
Experiments were performed using male Sprague-Dawley rats (447 ± 15 g, 10–15 wk old) supplied by University of Sydney Laboratory Animal Services. All experimental procedures were approved by the Animal Ethics Committee of the University of Sydney and were carried out in accordance with the Guidelines for Animal Experimentation of the National Health and Medical Research Council of Australia. Anesthesia was initially induced by inhalation of isoflurane (2–3% in oxygen-enriched air). A thermoregulated heating pad was used to maintain body temperature at 37–38°C as measured via a rectal probe. The femoral artery and vein were cannulated to facilitate measurement of cardiovascular parameters and administration of drugs, respectively. Isoflurane was withdrawn while being replaced by urethane (1.3g/kg iv with supplementary doses of 0.1 g/kg iv, if required). The adequacy of anesthesia was verified by the absence of the corneal reflex and a withdrawal response to nociceptive stimulation of a hind paw. A tracheotomy was performed to maintain an unobstructed airway, and the animals were allowed to breathe freely. The head was placed in a stereotaxic frame with the incisor bar fixed 19 mm below the interaural line, and a small area on the dorsal surface of the cortex was exposed to allow for later insertion of micropipettes into the hypothalamus.
2 in the tracheal tube was sampled continuously throughout the experiment using a CO2 meter (Datex, Engstrom). Because the response time of the CO2 meter was not rapid enough to measure end-tidal CO2, particularly when respiratory rate increased greatly in response to DMH disinhibition, the mean level of CO2 in the tracheal tube was measured instead, because changes in this variable would reflect changes in end-tidal CO2. The PNA and, in some experiments, renal sympathetic nerve activity (RSNA) were also recorded, as described below. At the end of the experiment, the rat was euthanized with an overdose of pentobarbital sodium, and the brain was removed and then fixed in 4% paraformaldehyde. Coronal sections (50 μm) were cut on a freezing microtome, and injection sites were determined using a fluorescence microscope.
The renal sympathetic nerve was exposed retroperitoneally using a method described previously (9). The phrenic nerve recordings were performed using a similar protocol with some adaptations. The left phrenic nerve was exposed and isolated from a dorsal approach, just ventral to the brachial plexus. Following exposure and isolation of the nerve, the distal end of the nerve was cut to eliminate afferent signals in the recording, and the proximal end of the nerve was then placed on a bipolar stainless steel recording electrode. The signal from the electrode was amplified and filtered (band pass 50–5,000 Hz), then rectified and passed through a low-pass filter (5 Hz). The raw and rectified nerve signals were recorded continuously on a computer using Chart software. All nerve recordings were sampled at a rate of 1,000 samples/s. The Chart software was also used to compute the rate and amplitude of the bursts of PNA.
Bicuculline methochloride microinjections (Tocris; 20 nl of 0.2, 0.5, or 2.0 mM solution) were made into the hypothalamus, using a glass micropipette held in a micromanipulator at an angle of 28° (tip caudal). The vehicle solution was artificial cerebrospinal fluid (aCSF) adjusted to pH 7.4. In all cases the injectate also contained green fluorescent latex microspheres (0.5%; Lumafluor) to facilitate the later histological verification of microinjection sites. Microinjections were made by pressure, using a previously described method (9). The tip of the micropipette was positioned stereotaxically into the DMH (3.1 mm caudal to bregma, 0.5 mm lateral to the midline, and 8.6 ventral to the surface of the cortex), posterior hypothalamic nucleus (PH; 3.8 mm caudal to bregma, 0.5 lateral to the midline, and 8.6 ventral to the surface of the cortex), ventromedial hypothalamus (VMH; 3.1 caudal to bregma, 0.5 lateral to the midline, and 9.6 ventral to the surface of the cortex), or an intermediate area between the DMH and PVN (2.8 caudal to bregma, 0.5 lateral to the midline, and 8.6 ventral to the surface of the cortex) as determined using a standard rat brain atlas (22).
For the dose-response experiments, four microinjections (1 each of 4, 10, and 40 pmol of bicuculline and 1 of aCSF vehicle solution) were made in each animal (n = 6) into the same site in the DMH. In each experiment all of the injections were performed on the same side of the brain (in 3 experiments on the right and in 3 experiments on the left). The order of the microinjections was randomized between experiments. After each injection, the pipette was withdrawn and either replaced with a new pipette containing the vehicle or a lower dose of bicuculline, or, in cases where a higher dose of bicuculline was subsequently injected, the pipette was washed out and the injectate replaced with a new solution containing the higher concentration of bicuculline.
In a series of control experiments, two microinjections (1 of 4 pmol of bicuculline and 1 of aCSF vehicle solution) were made into each of four sites in each animal (the DMH, VMH, PH, and an intermediate area between the DMH and PVN). The order of microinjections in these experiments was randomized, with half of the experiments performed on the left side and half on the right side of the brain. After each microinjection of bicuculline, the next microinjection was not performed until all cardiovascular variables had returned to their baseline levels and were stable. A waiting period of at least 15 min was observed in all cases.
In seven experiments, PNA and RSNA were recorded simultaneously. In these experiments, the correlation between bursts of PNA and changes in RSNA was determined before and at the peak of responses induced by microinjection of 10 pmol of bicuculline into the DMH, using cycle-triggered averaging as described in detail below.
To analyze the data for the dose-response experiments, 10-s samples of MAP, HR, and the rate and amplitude of PNA were taken every minute. In each experiment for each dose, the average values of MAP, HR, and PNA burst rate and amplitude during the 10-s sampling period were determined and then combined with the average values from other experiments to calculate the means ± SE for each time point before and after the time of injection. For experiments in which the responses evoked from the DMH were compared with responses evoked from surrounding sites, the maximum change in MAP, HR, and PNA burst rate and amplitude compared with the preinjection baseline level was determined following each microinjection of bicuculline or aCSF.
In seven experiments, cycle-triggered averaging was used to determine the relationship between bursts of PNA and changes in RSNA, using a procedure similar to that described previously (4
ANOVA was used to compare the maximum changes in MAP, HR, or PNA burst rate or amplitude after microinjections of different doses of bicuculline or aCSF into the DMH, followed by paired comparisons using the t-test with application of the Helm step-down procedure for multiple comparisons where appropriate (25). A P value <0.05 was regarded as statistically significant. All values are means ± SE.
Baseline levels of MAP, HR, PNA burst rate and amplitude, and mean tracheal CO2 level, measured just before microinjections of the vehicle (aCSF) solution or bicuculline (4, 10, or 40 pmol) into the DMH, were not significantly different among different groups (see Table 1). The centers of the injection sites for these experiments are shown in Fig. 1. Microinjections of bicuculline into the DMH evoked dose-dependent increases in PNA burst rate (respiratory rate) and PNA burst amplitude (Table 1 and Fig. 2). The increases in respiratory rate, however, were proportionately much greater than the increases in PNA burst amplitude. For example, microinjection of 10 pmol resulted in a peak increase in respiratory rate of 89 ± 15%, compared with 21 ± 5% in PNA burst amplitude. There was also a dose-dependent reduction in the mean tracheal CO2 level (Table 1), indicative of hyperventilation. In confirmation of previous findings (15), these respiratory effects also were accompanied by dose-dependent increases in both MAP and HR (Table 1 and Fig. 2). Microinjections of the vehicle aCSF solution had no significant effect on any of the variables (Table 1 and Fig. 2).
Dependence of effects on the site of injection.
In another series of four experiments, the respiratory responses evoked by bicuculline microinjections into sites surrounding the DMH were compared with those evoked from the DMH itself. As explained in materials and methods, in each of these experiments two microinjections (1 of aCSF and 1 of 4 pmol of bicuculline) were made into each of four different sites within the DMH, VMH, PH, and an intermediate area between the DMH and PVN. The centers of these injection sites (plus the centers of 6 injections of 4 pmol of bicuculline from the first series of dose-response experiments) in the DMH and surrounding regions are depicted on sagittal sections reconstructed from the original coronal sections (Fig. 3). As shown, the largest increases in respiratory rate (>50 breaths/min) were all evoked from sites in the DMH, particularly its more medial portion. Moderate (25–50 breaths/min) or small increases (10–24 breaths/min) in respiratory rate were evoked from sites within the more lateral part of the DMH and in the PH located slightly more dorsally and caudally to the DMH, whereas bicuculline injections into the VMH or the more rostral region between the DMH and the PVN and on the boundary of the PVN had no significant effect on respiratory rate (Fig. 3). Microinjection of aCSF into each of these sites had no significant effect on respiratory rate (data not shown).
Correlation between PNA bursts and cardiovascular variables.
In seven other experiments, an example of which is shown in Fig. 4, RSNA also was recorded, in addition to the other cardiorespiratory variables. The centers of the sites of the microinjections of bicuculline (10 pmol) in these experiments are also shown in Fig. 1. As previously reported (9, 15), bicuculline microinjections into the DMH resulted in a very marked increase in RSNA (Fig. 4A). As shown in Fig. 4B, RSNA occurred in bursts. Cycle-triggered averaging revealed that under baseline conditions, there was a component of the RSNA correlated to the bursts of PNA such that the peak RSNA occurred during the late inspiratory or early expiratory phase (Fig. 5). In addition, there also were, in most experiments, cyclic variations in arterial pressure that also correlated with the bursts of PNA (Fig. 5), although these occurred at a higher frequency (5.59 ± 0.17 Hz), which corresponded to the heart rate (5.59 ± 0.17 Hz, or 336 ± 10 beats/min). Thus these respiratory-related variations in arterial pressure presumably reflect the fact that the heart rate was, at least in part, synchronized with the PNA burst rate.
After microinjection of bicuculline (10 pmol) into the DMH, respiratory-related variations in arterial pressure were preserved in five of seven cases, but they now occurred at a higher frequency (7.12 ± 0.23 Hz, n = 5), which was very similar to that of the heart rate after microinjection of bicuculline (10 pmol) into the DMH in the same experiments (7.09 ± 0.12 Hz, or 427 ± 7 beats/min). After microinjection of bicuculline (10 pmol) into the DMH, the overall level of RSNA also was increased in all experiments (maximum increase of 117 ± 25% baseline), but the respiratory-related component of the RSNA varied. In one group of four experiments (group A), of which an example is shown in Fig. 5A, the RSNA still showed respiratory-related cyclic variations, although the magnitude of these were similar to those observed under preinjection baseline conditions. In the remaining three experiments (group B), of which an example is shown in Fig. 5B, the RSNA showed marked cyclic variations after microinjection of bicuculline into the DMH that were greatly enhanced in magnitude compared with those observed under baseline conditions, but these variations were now at the same frequency as the variations in arterial pressure, rather than the bursts of PNA.
A possible explanation for the appearance in some experiments of marked cyclic variations in RSNA at the frequency of the heart rate rather than the respiratory rate is that these variations were a direct consequence of the pulsatile changes in arterial pressure, which in turn are partly correlated to the bursts of PNA. To test this, cycle-triggered averaging was performed in which the trigger was now the arterial pressure pulse. As shown for example in Fig. 6, this analysis revealed that the cardiac-related cyclic variations in RSNA were greatly enhanced in all seven experiments. Under baseline conditions, the amplitude of these cardiac-related cyclic variations were 73 ± 10% of baseline, but after microinjection of bicuculline (10 pmol) into the DMH, this increased greatly to 309 ± 59% of baseline (P < 0.01).
The results of this study clearly demonstrate that disinhibition of neurons (by microinjection of bicuculline) in the DMH results in very marked increases in central respiratory drive, as indicated by large increases in PNA burst rate and more moderate increases in PNA burst amplitude. Second, we have demonstrated that although, as previously reported (1, 9, 15), disinhibition of the DMH also results in a marked increase in sympathetic activity, this is not associated with an increase in the degree of coupling between PNA and RSNA, as evidenced by the lack of increase in the amplitude of the PNA burst-related cyclic changes in RSNA. This observation indicates that the DMH-evoked increase in RSNA cannot be explained as simply a consequence of increased central respiratory drive. The broader implications of our findings are discussed in more detail below, after first considering some methodological issues.
Previous studies have reported that microinjections of bicuculline into the PVN (24) or PH also evoke increases in respiratory frequency (5, 33). Our results show, however, that the largest increases in respiratory frequency evoked by microinjections of small doses of bicuculline (4 pmol in 20 nl) were evoked by microinjections centered within the DMH (particularly its medial portion), whereas much smaller responses were evoked by the same dose injected into more caudal sites centered in the PH, in more rostral sites within the region between the PVN and the DMH and on the boundary of the PVN, or in more ventral sites within the VMH. Thus the marked increase in respiratory activity evoked by bicuculline microinjections within the DMH cannot be attributed to spread of bicuculline to these other nuclei. Similarly, because as already noted the largest responses were more frequently evoked by microinjections centered in the medial rather than the lateral part of the DMH, it is also unlikely that the effects could be due to spread of bicuculline to the perifornical region located more laterally.
In the studies referred to above in which bicuculline was injected into the PH, either the amount [>380 pmol in the study by Waldrop et al. (33)] or volume [1 μl in the study by DiMicco and Abshire (5)] was much greater than the amount or volume (4–40 pmol in 20 nl) injected into the DMH in the present study. It is therefore possible that in these previous studies (5, 33), the respiratory effects evoked by bicuculline injections into the PH were due, at least in part, to diffusion to neurons in the adjacent DMH. Similarly, in the study by Schlenker et al. (24) in which microinjections of 50 pmol of bicuculline in 50 nl were made into the PVN, it is possible that the evoked respiratory response was partly due to diffusion to neurons within the DMH. There are, however, direct projections from neurons in the PVN to the phrenic nucleus in the spinal cord as well as to respiratory nuclei in the medulla (16, 34). Thus it seems more likely that respiratory effects can be evoked by disinhibition of separate populations of neurons in the PVN and DMH.
In the present study, we did not perform a detailed systematic mapping of sites within the entire DMH and all surrounding regions. Thus, although the results of our study indicate that the effects of bicuculline microinjection into the DMH are due to disinhibition primarily of neurons within the DMH itself, rather than surrounding regions, further studies are required to define the precise extent of the responsive region.
In our experiments, the vagi were left intact, and the rats breathed spontaneously, to ensure that the respiratory pattern was as close to normal as possible. Thus the fact that the PNA burst amplitude increased to a lesser degree than the PNA burst frequency may reflect the fact that the increase in burst amplitude was limited by inhibitory feedback from pulmonary stretch receptors. Nevertheless, the results clearly show that central respiratory drive is markedly increased by disinhibition of DMH neurons, although afferent feedback from pulmonary stretch receptors may have modulated the pattern of respiratory activity, as would occur under natural conditions where central respiratory drive is increased.
It has been shown previously that activation of the DMH in anesthetized rats causes an increase in the level of expired CO2 (1). In the present study, we were not able for technical reasons to measure expired CO2 accurately. We did observe, however, that activation of the DMH resulted in a large decrease in the mean level of tracheal CO2, indicating that the expired CO2 decreased rather than increased. In our study, the rats were allowed to breathe spontaneously, whereas in the study by Cao et al. (1) the rats were paralyzed and artificially ventilated. Together, therefore, the results of our study and that of Cao et al. (1) indicate that although activation of the DMH leads to an increase in CO2 production (presumably as a result of thermogenesis in brown adipose tissue), in spontaneously breathing animals the level of expired CO2 decreases as a result of the very marked hyperventilation that occurs.
Correlation between respiratory activity and sympathetic activity.
It is well known that central respiratory activity can influence sympathetic activity (8, 13). Furthermore, the respiratory-related modulation of sympathetic activity can increase greatly in some situations where respiratory activity increases, such as in response to activation of peripheral chemoreceptors (4, 17). Consistent with previous studies (4, 17, 20), cycle-triggered averaging in our experiments revealed that under resting conditions, there was a respiratory-related modulation of RSNA such that the peak RSNA occurred during the late inspiratory or early expiratory phase. In contrast to chemoreceptor stimulation, however (4, 17), there was no consistent change in the amplitude of the respiratory-related variations in RSNA in response to disinhibition of the DMH, despite the marked increase in PNA burst amplitude and frequency that occurred under those conditions. This indicates that the marked increase in RSNA in response to disinhibition of the DMH is independent of the increased central respiratory drive that is also generated as part of the response.
In contrast to respiratory modulation, cycle-triggered averaging revealed that variations in RSNA related to the cardiac pulsatile changes in arterial pressure were greatly enhanced in response to DMH disinhibition. This increase in the cardiac pulse-related modulation of RSNA is consistent with the previous finding in our laboratory (18) that DMH disinhibition causes a marked increase in the gain of the sympathetic component of the baroreflex.
Cycle-triggered averaging also revealed that the arterial pressure was partly synchronized with the respiratory cycle, although at higher frequencies than the respiratory frequency (Fig. 5). It has been shown in humans that synchronization can occur between the heart beat and the respiratory cycle such that under resting conditions, there are periods during which there are three heart beats to one respiratory cycle, or sometimes five heart beats to two respiratory cycles (23). This phenomenon, which is thought to be due to coupling between oscillators in the brain that determine the respiratory rate and heart rate (23), would account for the partial synchronization that we observed between the PNA burst rate and the arterial pressure changes.
Apart from its role in mediating responses to acute psychological stress, the DMH also has been identified recently as a key component of the central pathways mediating the thermoregulatory response to exposure to a cold environment (7, 19). The sympathetic and respiratory responses to acute psychological stress and exposure to a cold environment are similar, since both types of stress evoke tachycardia, cutaneous vasoconstriction, sympathetically mediated thermogenesis, and increased ventilation and respiratory rate (7, 11, 19). On the other hand, a marked decrease in body core temperature (hypothermia) is associated with a bradycardia and a decrease in respiratory rate and ventilation (3, 21). Thus the sympathetic and respiratory effects evoked by disinhibition of the DMH are consistent with the hypothesis that the DMH neurons subserving these effects are part of the neural circuitry subserving the physiological response to acute psychological and cold stress rather than to hypothermia.
Descending pathways mediating respiratory effects.
There do not appear to be any descending projections from the DMH to the spinal cord (31). There is, however, a projection from the DMH to the dorsomedial medulla, which includes the dorsal respiratory group nucleus (9, 31), so it is possible that this pathway may mediate respiratory effects. There also is a major descending projection from the DMH to the midbrain periaqueductal gray (31), which can regulate respiratory activity via projections to the parabrachial/Kölliker-Fuse region in the dorsolateral pons (14), which in turn contains neurons known to have a major role in regulating respiratory rate (2). The precise organization of the descending pathways from the DMH that mediate its effects on respiratory activity requires further study.
The present study has highlighted the importance of the DMH in regulating respiratory activity. As pointed out in the Introduction, however, this is just one component of a coordinated pattern of physiological responses that closely resemble those responses that are naturally evoked by psychological stress. A very important underlying question is, what is the neuronal organization within the DMH that permits such a coordinated pattern of responses? Are there, for example, individual “command neurons” within the DMH that have collateral outputs to nuclei controlling the sympathetic outflow to the cardiovascular system, and to nuclei controlling respiratory rate and depth? Alternatively, are there separate subgroups of DMH neurons, each of which projects to brain stem nuclei regulating either sympathetic or respiratory activity? In the latter case, it is conceivable that coordinated cardiorespiratory patterns could result from interconnections between such subgroups of neurons within the DMH. Such questions need to be addressed by future studies.
This work was supported by the National Health and Medical Research Council of Australia and the Australian Research Council.
We thank Dr. Qi-Jian Sun and Dr. Simon McMullan for advice regarding the procedures for recording phrenic nerve activity, and Assoc. Prof. Robin McAllen for helpful discussions.
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