Post-hypoxia frequency decline in rats: sensitivity to repeated hypoxia and α2-adrenoreceptor antagonism

Brain Research 817 Ž1999. 25–33

Research report

Post-hypoxia frequency decline in rats: sensitivity to repeated hypoxia and a 2-adrenoreceptor antagonism Karen B. Bach ) , Richard Kinkead, Gordon S. Mitchell Department of ComparatiÕe Biosciences and Center for Neuroscience, School of Veterinary Medicine, UniÕersity of Wisconsin, 2015 Linden DriÕe West, Madison, WI, 53706, USA Accepted 27 October 1998

Abstract We tested the hypothesis that the post-hypoxia frequency decline of phrenic nerve activity following brief, isocapnic hypoxic episodes in rats is diminished by prior hypoxic episodes and a 2-adrenoreceptor antagonism. Anesthetized Žurethane., artificially ventilated Ž FI s 0.50. and vagotomized rats were presented with two or three, 5 min episodes of isocapnic hypoxia Ž FI f 0.11., separated by 30 O2 O2 min of control, hyperoxic conditions. Phrenic nerve discharge, end-tidal CO 2 , and arterial blood gases were measured before during and after hypoxia. The average maximum frequency decline, measured 5 min after the first hypoxic episode, was 26 " 7 burstsrmin below pre-hypoxic baseline values Ža 70 " 16% decrease.. By 30 min post-hypoxia, frequency had returned to baseline. Two groups of rats were then administered either: Ž1. saline Žsham. or Ž2. the a 2-receptor antagonist, RX821002 HCl Ž2-w2-Ž2-Methoxy-1,4-benzodioxanyl.x imidazoline hydrochloride; 0.25 mgrkg, i.v... Isocapnic hypoxia was repeated 10 min later. In sham rats, the post-hypoxia frequency decline ŽPHFD. was significantly attenuated relative to the initial Žcontrol. response. However, PHFD was attenuated significantly more in RX821002-treated vs. sham rats Žy3 " 3 burstsrmin vs. y12 " 4 burstsrmin @ 5 min post hypoxia for RX821002 and sham-treated, respectively; p - 0.05.. We conclude that the magnitude of PHFD is dependent on the prior history of hypoxia and that a 2 adrenoreceptor activation plays a role in its underlying mechanism. q 1999 Elsevier Science B.V. All rights reserved. Keywords: Respiratory control; Plasticity; Hypoxia

1. Introduction Phrenic nerve discharge increases substantially during short term isocapnic hypoxia. However, close examination of this response reveals that it is multifaceted, consisting of more than just a reflex increase in respiratory nerve activity during hypoxia followed by a return to baseline at the end of the hypoxic stimulus w30x. The response to a single, brief episode of hypoxia in rats consists of at least three time dependent mechanisms: Ž1. an acute increase in inspiratory neural activity observable within a single breath w13x; Ž2. a progressive increase in nerve amplitude over 1–2 min, followed by a similar progressive decrease in amplitude when the stimulus is removed Žshort-term potentiation w32x.; and Ž3. a peak in nerve burst frequency followed by a progressive decrease to steady-state during the first minute of hypoxia, followed by a persistent de) Corresponding author. [email protected]

Fax:

q 1-608-263-3926;

E-mail:

crease in frequency below baseline values after the stimulus is removed Žshort-term depression w20,30x.. Short-term depression has been observed in rats during andror after both electrical stimulation of the carotid sinus nerve and systemic hypoxia w20x, and may Žor may not. consist of separate mechanisms. Some studies have shown that short-term depression occurs during but not after carotid sinus nerve stimulation w26x. Whether or not shortterm depression occurs seems to depend on rat strain andror protocol design. The present study focused on short-term depression that follows an episode of isocapnic hypoxia, the post-hypoxia frequency decline ŽPHFD w8x. in a rat strain ŽSasco Sprague–Dawley. which exhibits frequency depression immediately after an hypoxic episode Žpersonal observation.. Although the mechanism underlying PHFD is unknown, several lines of evidence suggest that the noradrenergic nervous system Žand specifically, a 2-adrenergic receptors. may be involved w5,10,21,23x. Direct application of norepinephrine and clonidine Žan a 2 adrenoreceptor agonist. to bulbar respiratory neurons

0006-8993r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved. PII: S 0 0 0 6 - 8 9 9 3 Ž 9 8 . 0 1 1 8 1 - 0

26

K.B. Bach et al.r Brain Research 817 (1999) 25–33

decreases their firing frequency as well as the burst frequency of the phrenic nerve in decerebrate cats w5x. In addition, Hilaire et al. w21x demonstrated that pontine noradrenergic neurons tonically inhibit the medullary respiratory rhythm generator via a 2-adrenergic receptors. Norepinephrine is produced in several discreet regions of the pons, including the area A5 and the locus coeruleus ŽA6. w28x. If the area A5 is stimulated electrically, respiratory nerve burst frequency decreases in anesthetized rats w10,23x. Hypoxia activates these pontine noradrenergic neurons w14,31x, some of which project to medullary respiratory nuclei w11,12,17x. Based on this indirect evidence, we hypothesized that PHFD depends, at least in part, on the activation of a 2-adrenergic receptors. To test this hypothesis, we determined the magnitude of PHFD before and after administration of the a 2-adrenergic receptor antagonist, RX821002 HCl. In sham experiments, we examined the effects of initial hypoxic trials on the response to subsequent hypoxic trials in the same animal, and found unexpected evidence that prior hypoxic exposures can alter PHFD in subsequent trials, even though frequency had returned to baseline values.

2. Materials and methods Experiments were conducted on 16 adult male Sprague–Dawley rats ŽSasco, Madison, WI. weighing 318–481 g. The study involved two series of experiments. The first series compared PHFD before and after administration of the a 2-antagonist RX821002 HCl Ž2-w2-Ž2Methoxy-1,4-benzodioxanyl.x imidazoline hydrochloride, Research Biomedicals, Natick, PA.. RX821002 has been shown to bind with high affinity to the a 2 receptor subtype w18x. The second Žsham. series assessed the persistent effects of prior hypoxic exposure on subsequent responses. All experiments were approved by the University of Wisconsin animal care and use committee.

ŽStatham Pressure Transducer, P23-id.. The lungs were hyperinflated approximately once per hour to prevent alveolar collapse. All rats were vagotomized bilaterally and paralyzed Žpancuronium bromide, 2.5 mgrkg. to prevent spontaneous breathing efforts and entrainment of respiratory motor outflow with the ventilator. End-tidal CO 2 was monitored with a flow-through capnograph ŽNovametrix; Wallingford, CT. with sufficient response time Ž- 75 ms. to measure end-tidal pCO 2 in rats. End-tidal CO 2 values obtained from this capnograph closely approximated arterial pCO 2 Žusually within 1–2 mmHg.. Discrete blood samples were drawn from a catheterized femoral artery to determine blood gases and pH ŽABL-330; Radiometer, Copenhagen, Denmark.. Blood gas and pH values were corrected to the measured rectal temperature of each rat. Blood pressure was monitored at the femoral artery ŽStatham Pressure Transducer, P23-id.. Rectal temperature was maintained between 378 and 388C with a heated table. The phrenic nerve was isolated unilaterally using a dorsal approach, cut distally and desheathed. The nerve was submerged in mineral oil and placed on a bipolar silver recording electrode. Nerve activity was amplified Ž10,000 = ; CWE BMA 831; Ardmore, PA., band-pass filtered Ž100 Hz–5 kHz. and integrated ŽPaynter filter CWE 821; time constant 100 ms.. The integrated signal was digitized ŽScientific Solutions; Lab Master DMA; Solon, OH. and processed with computer software developed in our laboratory. 2.2. Experimental protocols Following completion of the surgical preparation, 60 min were allowed for the nerve signal to stabilize in hyperoxia Ž FI O s 0.50; paO 2 ) 150 mmHg. and normo2 capnia Ž paCO 2 approximately 3 mmHg above the apneic threshold; see Table 1.. Baseline nerve activity was achieved by manipulating inspired CO 2 and respiratory pump rate andror volume while monitoring end-tidal CO 2 levels until phrenic nerve activity attained a low but stable

2.1. Experimental preparation The animals were anesthetized initially with isoflurane Ž2.5–3.0% in 50% O 2 , balance N2 . and slowly converted to urethane anesthesia Ž1.6 grkg, i.v.. over a period of 30–45 min. The adequacy of anesthesia was assessed regularly by testing corneal reflexes and blood pressure responses to toe pinch. Supplemental urethane was administered as needed through a catheter implanted in a femoral vein. In most rats, a slow infusion of sodium bicarbonate Ž5.0%. and lactated Ringer’s solution Ž50:50, 1.7 ml kgy1 hy1 . was initiated 1–2 h after induction of anesthesia to maintain acid–base balance. All rats were prepared with a tracheostomy through which the animals were artificially ventilated ŽHarvard Rodent Respirator. and tracheal pressure was measured

Table 1 paCO 2 values ŽmmHg. in the 2 experimental series Experimental series

Time post-hypoxia Baseline

5 min

30 min

a 2 antagonist series Control Ž ns6. RX821002 Ž ns6.

42.2"1.0 42.5"1.2

42.1"1.2 42.7"1.3

42.0"1.0 42.6"2.2

Sham series Trial 1 Ž ns8. Trial 2 Ž ns8. Trial 3 Ž ns8.

44.2"0.9 44.2"1.0 44.9"1.0

44.3"0.8 44.5"1.1 44.5"1.0

43.7"1.0 44.9"1.0 45.7"0.8

At 5 min post-hypoxia, paCO 2 was actively maintained within 1 mmHg of the baseline value; otherwise respiratory nerve activities were not included in the analysis. Mean values"1 S.E.M.

K.B. Bach et al.r Brain Research 817 (1999) 25–33

level of activity. The protocol began with a control arterial blood sample Ž0.3 ml drawn into a 0.5 ml heparinized glass syringe; unused blood was returned to the animal.. All subsequent blood samples were compared to this initial control value. Baseline nerve activity was recorded, followed by a 5 min episode of isocapnic hypoxia Ž FI O s 2 0.11.. Relative isocapnia was maintained throughout a protocol by monitoring end-tidal CO 2 and adjusting inspired CO 2 accordingly. It was usually necessary to elevate FI CO 2 Ž0.007 to 0.012. to maintain isocapnia during hypoxia, indicating that CO 2 exchange at the lungs had decreased. A blood sample was taken during the hypoxic episode to assess the level of hypoxia used in each experiment. Average paO 2 values for the a 2 antagonist series Ž49 " 4 mmHg; n s 6. and the sham series Ž40 " 3 mmHg; n s 8. were not significantly different Ž p ) 0.05.. In two animals from the sham series, paO 2 values fell below 30 mmHg; data from these animals were not included in further analysis. On return to hyperoxic control conditions, adjustments in inspired CO 2 were made continuously in an attempt to maintain a constant end-tidal CO 2 . Additional blood samples were taken Žand nerve activity was recorded. 5 and 30 min post-hypoxia to assure that paCO 2 was within 1 mmHg of the baseline value during data collection ŽTable 1.. Rats with CO 2 partial pressures at 5 min post-hypoxia that differed from baseline values by more than 1 mmHg were not included in the analysis. Therefore, decreases in paCO 2 are unlikely to be responsible for the post-hypoxic frequency decline observed following episodic hypoxia in these studies. Thirty min after the initial hypoxic episode, RX821002 was administered intravenously Žin saline vehicle. at a dose of 0.25 mgrkg. Ten minutes later, a new baseline recording was made and the protocol was repeated. At the conclusion of experiments, the response to elevated levels of inspired CO 2 was recorded in the phrenic nerve to obtain a measure of maximal Žor at least a standardized hypercapnic ‘control’. nerve activity Ž pet CO 2 s 80–95 mmHg.. Euthanasia by urethane overdose Ži.v.. terminated the experiment. To control for the effects of sequential hypoxic exposures, the entire protocol was repeated in different animals exposed to three episodes of isocapnic hypoxia, separated by 30 min, but without RX821002 administration.

27

arbitrary Žlow. baseline value w15x. Statistical analyses were conducted using a 2-way, repeated measures ANOVA and paired t-tests with the Bonferroni correction for multiple comparisons. Differences were considered significant if p - 0.05; all values are described as means " 1 S.E.M.

3. Results 3.1. a 2 Antagonist series Fig. 1A shows baseline phrenic activity, followed by the hypoxic response Žonset of phrenic nerve response to hypoxia not shown. and the post-hypoxic response following the first hypoxic episode in one rat. This recording also illustrates the marked decrease in blood pressure during hypoxia in anesthetized rats, with the abrupt return toward baseline values following the hypoxic episode. We did not attempt to control these changes in blood pressure during or after hypoxia. Phrenic burst frequency remained depressed, even after blood pressure had returned to baseline, suggesting that changes in baroreceptor activity were not responsible for the observed frequency decline. In a second hypoxic trial, following RX821002 pre-treatment in the same rat ŽFig. 1B., PHFD was attenuated. Phrenic burst

2.3. Data analysis Peak amplitudes and frequency Žburstsrmin. of phrenic nerve activity were averaged over 50 bursts for each steady-state data point, and in 20-s bins for the first 5 min following hypoxic episodes. Averaged amplitude data were then normalized as a % change from baseline Žpre-stimulus control. activity and as a change, expressed as the percentage of the ŽCO 2-stimulated. maximum nerve activity. The latter form of normalization obviates concerns about expressing data in terms of the percentage increase above an

Fig. 1. Integrated phrenic ŽPhr. neurogram showing the decline in respiratory frequency that follows a 5 min episode of isocapnic hypoxia Ž FI s 0.11. in ŽA. control Žuntreated. rat and ŽB. the same rat after O2 pre-treatment with the a 2 adrenoreceptor antagonist RX821002 Ž0.25 mgrkg; i.v... These representative recordings illustrate that post-hypoxia frequency decline is significantly attenuated by the combination of a second hypoxic exposure and RX821002. Only the end of the hypoxic episode is shown ŽH., followed by post-hypoxic recovery under hyperoxic conditions Ž FI O s 0.50; marked by the arrow.. Baseline nerve 2 activity ŽB. is shown for comparison. ABPsarterial blood pressure.

28

K.B. Bach et al.r Brain Research 817 (1999) 25–33

amplitude gradually returned towards baseline after each hypoxic episode. These observations from an individual rat were generally consistent with mean responses. The onset of hypoxia triggered an immediate increase in burst frequency followed by a slight reduction to a steadystate value at the end of 5 min Ži.e., short-term depression.. Phrenic burst frequency decreased from 62 " 4 burstsrmin at 1 min of hypoxia to 56 " 3 burstsrmin after 5 min of hypoxia ŽFig. 2A; p - 0.05.. Following hypoxia, frequency decreased below baseline values and remained significantly depressed for at least 5 min following the

hypoxic episode Ži.e., post-hypoxia frequency decline; PHFD.. Phrenic burst frequency was decreased by 26 " 7 burstsrmin Ž70 " 16%., 5 min after the hypoxic episode Ž p - 0.01.. Five minute post-hypoxia data are reported because blood gases and blood pressure are stable and have returned to baseline values at this time Žsee Table 1, Fig. 5.. Following RX821002 pre-treatment: Ž1. there were no differences in magnitude or time course of frequency changes during hypoxia, and Ž2. phrenic burst frequency was no longer depressed at 5 min post-hypoxia Ž Dfrequencys -3 " 3 burstsrmin; Fig. 2A.. In both cases, burst frequency had returned to baseline values by 30 min post-hypoxia Žbaseline frequency pre-RX821002s 44.6 " .4.5 and baseline frequency post-RX821002s 42.6 " 1.2.. To minimize normalization artifacts caused by variable baseline nerve activities, amplitude data were expressed both as a % change from baseline and as a change from baseline expressed as a percentage of the maximal CO 2stimulated response Ž% maximum.. Because all findings were similar, regardless of the normalization used, only amplitude data expressed as % change from baseline are presented. Pre-treatment with RX821002 had no effect on baseline phrenic activity prior to the application of the second isocapnic hypoxic episode Žchange in burst amplitude from pre-drug value: y17 " 8% at 15–20 min post RX821002; p ) 0.1.. Phrenic burst amplitude was elevated during hypoxia and returned to baseline levels within 2–3 min post-hypoxia, although there was a nonsignificant trend toward a decreased hypoxic phrenic response ŽFig. 2B. Phrenic burst amplitude remained near baseline values 30 min post-hypoxia. 3.2. Sham series

Fig. 2. ŽA. Changes in phrenic burst frequency Žexpressed as change from pre-hypoxic baseline values. illustrating short-term depression during hypoxia Žbetween 0 and 5 min. and post-hypoxic frequency decline ŽPHFD; between 5 and 10 min. in rats before Ž`; ns6. and following RX821002 treatment Žv; 0.25 mgrkg; ns6.. No significant differences were seen in frequency responses during hypoxia treated and untreated rats. Following hypoxia, burst frequency was significantly less than baseline values for at least 5 min. Treatment with RX821002 prior to the second hypoxic episode attenuated PHFD relative to untreated animals. Phrenic burst frequency returned to baseline levels within 30 min of hypoxic exposure in both conditions. ŽB. Changes in phrenic burst amplitude Žexpressed as percent change from pre-hypoxic baseline values. during and after hypoxic exposure. There was no effect of trial sequence or RX821002 on phrenic burst amplitude during or following hypoxia. )Indicates a value significantly different from baseline in all three groups Ž p- 0.05.. aIndicates a response significantly different prevs. post-treatment Ž p- 0.05.. †Indicates a value at 5 min significantly different from the corresponding peak Ž50 s. value.

Sham studies were conducted to determine the effects of repeated hypoxic exposures on PHFD. In one representative rat ŽFig. 3., phrenic nerve burst frequency was depressed following the first of three hypoxic episodes. However, following the second hypoxic episode, PHFD was attenuated and, in this particular rat, diminished even further following the third hypoxic episode ŽFig. 3 sham trial 2 and 3.. These observations from one rat are generally consistent with mean responses, although there were no consistent differences between the second and third trials. When average phrenic burst frequency was expressed as a change from baseline values ŽFig. 4A., burst frequency exhibited a significant increase during hypoxia, followed by a decline to the new steady-state value Ži.e., short-term depression.; in the first trial, phrenic burst frequency decreased by 17 " 2 burstsrmin between 1 and 5 min of hypoxia Ž p - 0.05.. The phrenic burst frequency response during the third hypoxic episode was significantly less robust ŽFig. 4A; p - 0.05. than the responses observed during the first and second hypoxic trials which were the same. Frequency significantly decreased below baseline

K.B. Bach et al.r Brain Research 817 (1999) 25–33

29

quency and blood pressure respectively; baseline frequency for: trial 1 s 38.1 " 2.9; Trial 2 s 37.4 " 1.7; Trial 3 s 43.3 " 1.8.. 3.3. Blood pressure responses During hypoxic episodes, mean arterial blood pressure decreased, and returned abruptly towards baseline values following hypoxia. However, there were no significant differences in the magnitude or time course of changes in mean arterial blood pressure in any experimental condition

Fig. 3. Integrated phrenic ŽPhr. neurograms from a single rat showing three successive measurements of post-hypoxic frequency decline ŽPHFD.; each trial was separated by 30 min of recovery under hyperoxic conditions Ž FI O s 0.50.. These neurograms illustrate that PHFD is decreased 2 in sham trial 2 compared to sham trial 1, although no additional effect was observed in sham trial 3. Only the end of each hypoxic episode is shown ŽH.; post-hypoxic recovery began at the arrow. ABPsarterial blood pressure.

values for at least 5 min following the first hypoxic episode Ž Dfrequency: y22 " 4 burstsrmin at 5 min posthypoxia, p - 0.05.. Phrenic burst frequencies were also significantly depressed following the second and third hypoxic episodes, indicating that PHFD had occurred Žin sham trials 2 and 3, Dfrequency @5 min post-hypoxias 12 " 4 and -8 " 1 burstsrmin, respectively; both p 0.05.; however, PHFD in trials 2 and 3 were significantly less than that of trial 1 Ž p - 0.05., although not different from one another. Phrenic burst amplitude progressively declined towards baseline values after the hypoxic episode, a manifestation of short-term potentiation, but was similar after successive hypoxic episodes ŽFig. 4B.. Changes in PHFD between hypoxic trials occurred even though phrenic burst frequency and arterial blood pressure had nearly returned to baseline levels prior to the next hypoxic episode Ž p s 0.146 and p s 0.747, for burst fre-

Fig. 4. ŽA. Changes in phrenic burst frequency Žexpressed as change from pre-hypoxic baseline values. illustrating short-term depression during hypoxia Žbetween 0 and 5 min. and post-hypoxia frequency decline Žbetween 5 and 10 min. in three successive hypoxic sham trials in 8 rats: sham trial 1 Ž`.; sham trial 2; Ž^.; and sham trial 3 Žv .. The hypoxic frequency response during the first two hypoxic trials was more robust than the third Ž p- 0.05., although in each case, frequency declined from a peak during hypoxia Ži.e., short-term depression.. Post-hypoxic frequency decline was significantly greater after the first hypoxic episode than after the next two episodes Žsham trials 2 and 3., although these changes were not as great as with RX821002 pre-treatment. ŽB. Changes in phrenic burst amplitude Žexpressed as percent change from pre-hypoxic baseline values. during and after exposure to hypoxia. )Indicates a value significantly different from baseline in all three groups Ž p- 0.05.. †Indicates a value at 5 min significantly different from the corresponding peak Ž50 s. value. ‡Indicates a response significantly different from sham trial a1 Ž p- 0.05..

30

K.B. Bach et al.r Brain Research 817 (1999) 25–33

quency decline in anesthetized rats, and that a single hypoxic episode can reduce PHFD in subsequent exposures to hypoxia, even though all measured variables have returned to baseline values. The dependence of PHFD on the history of hypoxic exposures may represent a form of ‘metaplasticity’ w1x in that a respiratory ‘memory’ Ži.e., PHFD. is altered by experience. 4.1. Post-hypoxia frequency decline

Fig. 5. ŽA. Time course of isocapnic hypoxia-induced changes in mean arterial blood pressure in controlŽ`; ns6. vs. RX821002 Žv; 0.25 mgrkg; ns6. treated rats, and ŽB. in rats exposed to three sham trials of hypoxia: sham trial 1 Ž`.; sham trial 2; Ž^.; and sham trial 3 Žv .. Changes in mean arterial pressure are expressed as a change from pre-hypoxic baseline values. After 5 min of hypoxia Žthis data point is labeled as ‘hypoxia’ on figures., mean arterial blood pressure has decreased in all treatment groups. However, there is no difference between the return of mean arterial pressure toward baseline in any of the treatment groups. This figure illustrates that changes in mean arterial blood pressure are probably not responsible for the decrease in post-hypoxic frequency decline in rats pretreated with RX821002 or in rats exposed to multiple trials of hypoxia. )Indicates a value significantly different from baseline in all groups in the series Ž p- 0.05..

Žsham and RX821002-treated rats; Fig. 5A,B.. Blood pressure had returned to Žor exceeded. pre-hypoxic baseline values by 2 min post-hypoxia in all cases despite the persistence of PHFD for more than 5 min. Thus, changes in arterial blood pressure are not correlated with the development of PHFD post-hypoxia, nor with differences in PHFD caused by RX821002 or successive hypoxic episodes.

In rats, single episodes of isocapnic hypoxia are immediately followed by a post-hypoxic depression of phrenic burst frequency known as post-hypoxic frequency decline w8x. This phenomenon was first described by Hayashi et al. w20x following direct electrical stimulation of the carotid sinus nerve or isocapnic hypoxia. Hypoxic ventilatory decline is a similar phenomenon which also results in decreased ventilation during and sometimes following sustained hypoxia. It has been observed in adult humans, cats, and a variety of neonatal mammals w4x. The mechanism underlying PHFD, as well as the frequency depression observed during the second minute of an hypoxic episode Žshort-term depression., is most likely distinct from that mediating hypoxic ventilatory decline since PHFD and short-term depression result in changes in timing and not the amplitude of respiratory motor output. During sustained isocapnic hypoxia, hypoxic ventilatory decline does not occur in adult rats w19x. The role of adrenergic receptors in hypoxic ventilatory decline has not, to our knowledge, been investigated in any species. Acute, 10-min exposures to severe hypoxia in carotid denervated cats results in a prolonged Ž60 min. depression of phrenic nerve activity Žboth burst amplitude and frequency. upon return to hyperoxic conditions w16x. This inhibition was attenuated by the adenosine receptor antagonist, theophylline. Again, since this phenomenon affected burst amplitude and frequency, it is unlikely that the mechanism is the same as that which underlies PHFD. To our knowledge, no one has attempted to block PHFD with theophylline, nor has anyone attempted to block the longlasting depression reported by Gallman and Millhorn w16x with a 2 receptor antagonists. We have recently shown that pre-treatment with the 5-HT1 receptor agonist 5-carboxamidotryptamine Ž5-CT. or the 5-HT2 receptor antagonist ketanserin accentuates PHFD in rats w24x. These results are consistent with the interpretation that PHFD results from complex interactions between a number of modulatory inputs Ži.e., noradrenergic and serotonergic. to brainstem respiratory neurons. 4.2. RX821002 attenuates post-hypoxia frequency decline

4. Discussion This study indicates that a 2-adrenergic receptor activation is necessary for full expression of post-hypoxia fre-

Systemic pre-treatment with RX821002 attenuated PHFD, but not the short-term depression of phrenic burst frequency during hypoxia. Similarly, the frequency decline

K.B. Bach et al.r Brain Research 817 (1999) 25–33

during carotid sinus nerve stimulation can occur without an ensuing post-stimulus frequency decline or short-term depression in rats w26x. Thus, it appears that the frequency decline during hypoxia Žor carotid sinus nerve stimulation. and that observed post-hypoxia Žor carotid sinus nerve stimulation. represent unique mechanisms with a similar manifestation. In a study similar to ours, PHFD was not prevented by intra cerebro-ventricular infusion of the a 2 adrenoreceptor antagonists RX821002 or SK and F-86466 w9x. There are several differences in experimental protocol that could account for these discrepant results. For instance, Coles et al. w9x used Equithesin, a mixture of sodium pentobarbitone Ž0.3 mgrkg. and chloral hydrate Ž1.33 mgrkg., to anesthetize their rats, whereas we used urethane Ž1.6 grkg.. Moreover, their protocol involved a shorter, more severe hypoxic stimulus Ž FI O s 0.08 for 45–60 s; resulting paO 2 2 values were not reported.. Finally, they used a different substrain of Sprague–Dawley rats ŽZivic-Miller vs. Sasco. which, given the substantial differences in noradrenergic innervation of the spinal cord in Sprague–Dawley rats from different colonies w6x, could play an important role in noradrenergic modulation of respiratory motor output. In our laboratory, we have observed differences in the magnitude of PHFD between Sprague–Dawley rats from Harlan and Sasco Žboth from Madison, WI; unpublished observation.. Each of these factors may contribute to the different results between studies, suggesting that studies of plasticity in respiratory motor control must adequately consider genetic and protocol differences. 4.3. Effect of episodic hypoxia on post-hypoxia frequency decline: metaplasticity Three successive episodes of isocapnic hypoxia, separated by 5 min, result in long-term facilitation of phrenic nerve amplitude and frequency in anesthetized rats w2x. If the hypoxic episodes are separated by 30 min or more, significant long-term facilitation does not occur Žpresent study.. Although each hypoxic episode was separated by 30 min of recovery in the present study, a diminished PHFD in successive hypoxic episodes during sham experiments revealed an unexpected form of plasticity that may influence the interpretation of our results. However, the results from the sham protocol do not completely account for the attenuation of post-hypoxia frequency decline following RX821002 administration. The PHFD observed after the second and third hypoxic episodes was still significantly greater than in rats pre-treated with RX821002 ŽFig. 6., thereby suggesting that a 2 adrenoreceptor activation contributes to PHFD. The observation that a single hypoxic episode can have long term effects on hypoxic responses conducted 30 min later is intriguing. It suggests a degree of plasticity beyond that which is usually reported in the literature. Abraham and Bear w1x refer to this type of plasticity as ‘metaplastic-

31

Fig. 6. Changes in phrenic burst frequency Žexpressed as change from pre-hypoxic baseline values. illustrating short-term depression during hypoxia Žbetween 0 and 5 min. and post-hypoxic frequency decline Žbetween 5 and 10 min. in untreated Ž`; ns6., RX821002 treated Žv; 0.25 mgrkg; ns6. rats, and rats pretreated with a single 5 min episode of hypoxia 30 min prior to a second hypoxic episode Ž^; sham trial 2; ns8.. This figure illustrates that pre-treatment with RX821002 attenuates post-hypoxia frequency decline more than a previous hypoxic trial Žsham trial a2., suggesting that previous hypoxic trials influence the return toward baseline during the second hypoxic trial, but cannot completely account for the effects of RX821002 pretreatment. )Post-hypoxic value significantly different from baseline Ž p- 0.05.. a response significantly different from sham trial a2 Ž p- 0.05.. †Indicates a value at 5 min significantly different from the corresponding peak Ž50 s. value.

ity,’ or the modulation of Žsynaptic. plasticity by prior ‘experience’ Žsynaptic activity.. For example, they report that activation of hippocampal NMDA receptors with a weak tetanus inhibits the induction of long-term potentiation and predisposes the system to long-term depression. It is reasonable to suspect that parallel mechanisms of metaplasticity are functional throughout the brain, including the respiratory control system. The ‘metaplasticity’ we observed may reflect interactions between a 2-dependent PHFD and opposing mechanisms such as Žsub-threshold. serotonin-dependent long-term facilitation w2,3,13,27x. 4.4. Possible mechanisms of post-hypoxia frequency decline We hypothesize that PHFD is caused, at least in part, by noradrenergic activation of central a 2 receptors which inhibit phrenic nerve burst frequency following hypoxia in rats. There are also a 2 receptors located peripherally in the carotid bodies w25x and intracarotid infusions of norepinephrine can depress ventilation in goats w29x. However, it is unlikely that increased circulating norepinephrine is responsible for PHFD since a previous study in our laboratory indicates that systemic norepinephrine injections in anesthetized rats cause only a short-lasting Žseconds. depression, followed by a long-lasting Žminutes. increase in respiratory nerve activity w3x.

32

K.B. Bach et al.r Brain Research 817 (1999) 25–33

Norepinephrine is produced in several discrete regions within the brainstem. Of these, the A5 region of the caudal ventrolateral pons and the locus coeruleus Žarea A6., are activated by hypoxia w12,17,31x. Electrical stimulation in the ventrolateral pons near the A5 noradrenergic area slows phrenic nerve burst frequency w23x, and bilateral chemical Žmuscimol. or electrolytic lesions of the A5 noradrenergic area block PHFD w8x. In contrast, locus coeruleus activation increases phrenic motor output w7x. Since norepinephrine tonically inhibits the medullary respiratory rhythm generator via a 2 receptors in neonatal rat preparations w21,22x, our working hypothesis is that hypoxia elicits PHFD at least partially via activation of ŽA5. noradrenergic neurons with projections to medullary regions involved in respiratory rhythmogenesis w11x and the subsequent activation of inhibitory a 2-adrenergic receptors at these sites. Chemoafferent activation with hypoxia can result in post-hypoxia augmentation Žlong-term facilitation w2,27x. andror inhibition ŽPHFD. of respiratory motor output. The absence of observable effects on phrenic burst amplitude during PHFD may be due to concurrent activation of short-term potentiation andror long-term facilitation, which preferentially influence amplitude vs. frequency w2x. Long term facilitation may offset, and mask potential decreases in nerve burst amplitude associated with the mechanisms that underlie PHFD.

w9x

w10x

w11x

w12x

w13x

w14x

w15x

w16x

w17x

w18x

Acknowledgements This study was supported by grants from the National Institutes of Health ŽHL 36780 and HL 53319., the Neuroscience Training Program ŽGM07507., and a post-doctoral fellowship from the Medical Research Council of Canada ŽRK..

w19x

w20x

w21x

References w1x W.C. Abraham, M.F. Bear, Metaplasticity: the plasticity of synaptic plasticity, TINS 19 Ž1996. 126–130. w2x K.B. Bach, G.S. Mitchell, Hypoxia-induced long-term facilitation of respiratory activity is serotonin dependent, Respir. Physiol. 104 Ž1996. 251–260. w3x K.B. Bach, G.S. Mitchell, Hypercapnia-induced long-term depression of respiratory activity requires a 2 adrenergic receptors, J. Appl. Physiol. 84 Ž6. Ž1998. 2099–2105. w4x G.E. Bisgard, J.A. Neubauer, Peripheral and central effects of hypoxia, Regulation of Breathing, Vol. 79, Dekker, New York, 1995, pp. 617–668. w5x M.D. Champagnat, J.L. Henry, V. Leviel, Catecholaminergic depressant effects on bulbar respiratory mechanisms, Brain Res. 160 Ž1979. 57–68. w6x F.M. Clark, H.K. Proudfit, Anatomical evidence for genetic differences in the innervation of the rats spinal cord by noradrenergic locus coeruleus neurons, Brain Res. 591 Ž1992. 44–53. w7x E.L. Coates, A. Li, E.E. Nattie, Widespread sites of brainstem ventilatory chemoreceptors, J. Appl. Physiol. 75 Ž1993. 5–14. w8x S.K. Coles, T.E. Dick, Neurones in the ventrolateral pons are

w22x

w23x w24x w25x

w26x

w27x

w28x

required for post-hypoxic frequency decline in rats, J. Physiol. 497 Ž1. Ž1996. 79–94. S.K. Coles, P. Ernsberger, T.E. Dick, Post-hypoxic frequency decline does not depend on a 2 -adrenergic receptors in the adult rat, Brain Res. 794 Ž2. Ž1998. 267–273. T.E. Dick, S.K. Coles, J.S. Jodkowski, A ‘pneumotaxic centre’ in the ventrolateral pons of rats, Ventral Brainstem Mechanisms and Control Functions, Marcel Dekker, New York, 1995, pp. 723–737. E.G. Dobbins, J.L. Feldman, Brainstem network controlling descending drive to phrenic motoneurons in rat, J. Comp. Neurol. 347 Ž1994. 64–86. M. Elam, T. Yao, P. Thoren, T.H. Svensson, Hypercapnia and hypoxia: chemoreceptor-mediated control of locus coeruleus neurons and splanchnic, sympathetic nerves, Brain Res. 222 Ž1981. 373–381. F.L. Eldridge, D.E. Millhorn, Oscillation, gating, and memory in the respiratory control system, Handbook of Physiology, The Respiratory System, American Physiological Society, Washington, DC, 1986, pp. 93–114. J.T. Erickson, D.E. Millhorn, Hypoxia and electrical stimulation of the carotid sinus nerve induce Fos-like immunoreactivity within catecholaminergic and serotoninergic neurons of the rat brainstem, J. Comp. Neurol. 348 Ž1994. 161–182. R. Fregosi, G.S. Mitchell, Long term facilitation of inspiratory intercostal nerve activity following repeated carotid sinus nerve stimulation in cats, J. Physiol. 477 Ž3. Ž1994. 469–479. E.A. Gallman, D.E. Millhorn, Two long-lasting central respiratory responses following acute hypoxia in glomectomized cats, J. Physiol. 395 Ž1988. 333–347. P.G. Guyenet, N. Koshiya, D. Huangfu, A.J.M. Verberne, T.A. Riley, Central respiratory control of A5 and A6 pontine noradrenergic neurons, Am. J. Physiol. ŽRegulatory Integrative and Comparative Physiology. 33 Ž1993. R1035–R1044. M. Halme, B. Sjoholm, J.M. Sovola, M. Scheinin, Recombinant human alpha 2-adrenoceptor subtypes: comparison of wrauwolscine, 3 Hx3 Hxatipamezole and w3 HxRX821002 as radioligands, Biochim. Biophys. Acta 1266 Ž2. Ž1995. 207–214. M.A. Haxhiu, C.H. Chang, I.A. Dreshaj, B. Erokwu, N.R. Prabhakar, N.S. Cherniack, Nitric oxide and ventilatory response to hypoxia, Respir. Physiol. 101 Ž1995. 257–266. F. Hayashi, S.K. Coles, K.B. Bach, G.S. Mitchell, D.R. McCrimmon, Time dependent phrenic nerve responses to carotid afferent activation: intact vs. decerebellate rats, Amer. J. Physiol. ŽRegulatory Integrative and Comparative Physiology. 265 Ž1993. R811– R819. G. Hilaire, R. Monteau, S. Errchidi, D. Morin, J.M. Cottet-Emard, Noradrenergic modulation of the medullary respiratory rhythm generator by the pontine A5 area, Respiratory Control: Central and Peripheral Mechanisms, University of Kentucky Press, 1993, pp. 43–47. S. Homma, E. Kusano, S. Takeda, N. Murayama, Y. Asano, T.P. Dousa, Both beta-1 and beta-2 adrenoreceptors are coupled to the cyclic AMP system in connecting tubules of rabbit nephron, J. Pharmacol. Exp. Ther. 264 Ž1. Ž1993. 105–109. J.S. Jodkowski, S.K. Coles, T.E. Dick, A ‘pneumotaxic centre’ in rats, Neurosci. Lett. 172 Ž1–2. Ž1994. 67–72. R. Kinkead, G.S. Mitchell, Serotonin modulates post-hypoxia frequency decline ŽPHFD. in rats, FASEB J. 12 Ž5. Ž1998. A780. Y.R. Kou, P. Ernsberger, P.A. Cragg, N.S. Cherniack, N.R. Prabhakar, Role of a 2-adrenergic receptors in the carotid body response to isocapnic hypoxia, Respir. Physiol. 83 Ž1991. 353–364. L. Ling, E.B. Olson, E.H. Vidruk, G.S. Mitchell, Integrated phrenic responses to carotid afferent stimulation in adult rats following perinatal hyperoxia, J. Physiol. 500 Ž3. Ž1997. 787–796. D.E. Millhorn, F.L. Eldridge, T.G. Waldrop, Prolonged stimulation of respiration by endogenous central serotonin, Respir. Physiol. 42 Ž1980. 171–198. R.Y. Moore, F.E. Bloom, Central catecholamine neuron systems:

K.B. Bach et al.r Brain Research 817 (1999) 25–33 anatomy and physiology of the norepinephrine and epinephrine systems, Annu. Rev. Neurosci. 2 Ž1979. 113–168. w29x J. Pizzaro, F.E. Warner, M.L. Ryan, G.S. Mitchell, G.E. Bisgard, Intracarotid norepinephrine infusions inhibit ventilation in goats, Respir. Physiol. 90 Ž1992. 299–310. w30x F.L. Powell, G.S. Mitchell, W.K. Milsom, Time domains of the hypoxic ventilatory response, Resp. Physiol. 112 Ž1998. 123–134.

33

w31x L.J. Teppema, J.G. Veening, A. Kranenburg, A. Dahan, A. Berkenbosh, C. Olievier, Expression of c-fos in the rat brainstem after exposure to hypoxia and to normoxic and hyperoxic hypercapnia, J. Comp. Neurol. 388 Ž1997. 169–190. w32x P.G. Wagner, F.L. Eldridge, Development of short-term potentiation of respiration, Respir. Physiol. 83 Ž1991. 129–140.