INTRODUCTION

Top Abstract Introduction Methods Results Discussion Appendix References

THE INVOLVEMENT OF CENTRAL catecholamines in cardiorespiratory (16) or respiratory-cardiovascular (13) coordination has been postulated. A schema suggests that the rostral ventrolateral medulla (RVLM) is the final integrative structure for both the sympathetic baro- and the sympathetic chemoreflex (see Fig. 11 of Ref. 13). Cell bodies located immediately above the chemosensitive area of the ventral surface of the medulla project monosynaptically to the intermediolateral cell column (1). RVLM barosensitive bulbospinal (RVLMbb) neurons have been classified, with some overlap, as 1) slow conducting, presumably adrenergic neurons, sensitive to low concentrations of clonidine, with a slow baseline firing rate; and 2) fast conducting, presumably glutamatergic neurons, sensitive to higher concentrations of clonidine, with a high baseline firing rate (10, 31). If these adrenergic RVLMbb neurons are part of the sympathetic chemoreflex, they should be activated by local or systemic acidosis. By contrast, RVLMbb neurons characterized by a high baseline firing rate are not activated by iontophoresis of H⁺ (35). However, these neurons were not characterized by their conduction velocity or their clonidine sensitivity (35). Thus the sensitivity of presumed adrenergic RVLMbb neurons to acidosis has not been thoroughly studied.

In vivo electrochemistry (8) allows one to dynamically record the activity of catechol metabolism in the RVLM next to adrenergic cell bodies (7, 17, 27). This catechol signal is increased when 1) hypercapnia alone is imposed on the system (24) and 2) hypercapnia and/or systemic acidosis superimpose themselves on hypotension (3). However, previous design (3) has precluded an unequivocal conclusion with respect to the response of RVLM catechol metabolism after pure metabolic isocapnic acidosis, in the presence of unaltered mean arterial pressure (MAP).

The present study aims at documenting an RVLM catechol activation on systemic H⁺ load. Given the relationship between the RVLM catechol signal and minor changes in MAP (27) and the limitations of previous data (3), changes in MAP were accepted, in the present study, during HCl challenge if <15 mmHg. A functional difference is believed to exist between the RVLM itself, primarily cardiovascular, and the obex-ventrolateral medulla (obex-VLM), putatively respiratory (see Fig. 11 of Ref. 13 for details). This report restricts itself to recording sites located at least 1,000 µm rostral to the obex. An RVLM catechol activation was observed after mild systemic prolonged isocapnic acidosis in the presence of stable MAP.

	MATERIAL AND METHODS

Top Abstract Introduction Methods Results Discussion Appendix References

General procedures. With the approval of the Animal Care Committee, male Sprague-Dawley rats (Iffa-Credo, 350-400 g) were housed (3-6 per cage, temperature: 21 ± 1°C, light: 7 AM-7 PM) with rat chow and water up to the experimentation. Anesthesia was induced with halothane (5%) in O₂. After ensuring a MAP of 105 ± 5 mmHg with volume loading (22) and adjusting the halothane concentration (0.8-1%), a hemostat was clamped to full ratchet lock onto the tail to assess the behavioral response to a nociceptive stimulus. This did not induce movement in any of the animals. Then, muscle relaxation was induced using metocurine (400 µg/kg iv; Lilly). This dose was administered again before the beginning of the H⁺ challenge. After muscle relaxation, tail clamping was performed at regular intervals. When an increase in MAP of >10-15 mmHg occurred on tail clamping, the halothane concentration was increased by 0.1-0.2%. Then, the end-tidal concentration of halothane, monitored with a Siemens gas monitor GM120, was kept constant (0.87 ± 0.24%) within each experiment up to euthanasia. Rectal temperature was maintained at 37°C using a warming blanket. Mechanical ventilation was established (frequency 52 breaths/min, oxygen 40-50%) through a tracheotomy tube (ID 1.67 mm). The tidal volume was adjusted to maintain an end-tidal CO₂ of 30-35 mmHg (Engstrom Eliza Duo modified with a neonatal kit) throughout the experiments. Arterial blood gases were measured every 30 min on an ABL30 Radiometer blood gas analyzer and every 10 min during acid challenge. MAP was recorded through a femoral catheter with a Bentley Trantec 800 transducer-Gould pressure processor-Gould2600S recorder. The processor counts heart rate (HR) from pulse pressure. Ringer lactate (10.8 ± 1.8 ml · kg¹ · h¹) was continuously infused through a femoral venous catheter to minimize a progressive decline in systemic pressure (22).

Sinodeafferentation. Chemodenervation was assessed in a separate group of animals anesthetized with pentobarbital sodium (induction: 37.5 mg/kg ip; maintenance: 7.5 mg/kg iv every 30-45 min as required, at least 15 min away from injection of lobeline or cyanide) and maintained under spontaneous ventilation to observe changes in respiratory rate (RR) after lobeline or cyanide injections. The carotid sinus nerves were sectioned bilaterally next to their merging point with the glossopharyngeal nerve (IX) in the sinodeafferented animals and left untouched in intact animals. The vagi, aortic depressor, superior laryngeal nerves, and the sympathetic trunk were untouched. A three-way tracheotomy tube was devised 1) to allow spontaneous nonobstructive breathing with added oxygen (fraction of inspired O₂ 0.40-0.45) to lower baseline RR to 50-60 cycles/min and 2) to measure end-tidal CO₂ and thus respiratory rate. The external carotid was catheterized (Biotrol; ID 0.3 mm) retrogradely to place the distal tip of the catheter as closely as possible to the carotid sinus. Lobeline (Sigma; 100 µg/kg iv, dissolved in 1 drop of 1 N HCl then in saline; pH = 7) and sodium cyanide (Sigma; 100 µg/kg iv in saline; pH = 7) were injected as a bolus (100 µl) in 15 s. Then the catheter was rinsed with saline (100 µl, pH = 7) over a period of 15 s. This was performed before and after section of the carotid sinus nerve. The injections of stimulants were separated by at least 15 min to allow a return of the respiratory rate to baseline. From a stable respiratory rate of at least 60 s, the maximal change in respiratory rate (RR) was calculated and averaged for each stimulant and condition.

Protocol. Rats were allocated to three groups. 1) In the saline group (data from Ref. 27 with permission), baseline recording was carried out for at least 30 min; the MAP fluctuated spontaneously within the 100- to 110-mmHg range. The height of the catechol signal varied by <10%. After injection of saline (500 µl iv bolus), the recording was continued for 150 min under constant halothane concentration. 2) In the HCl intact group, a slow infusion of 1 N HCl in saline (3.54 ± 0.98 ml/kg) was administered through the internal jugular vein by means of an electric syringe over a period of 30-50 min (mean: 38 min) to lower the arterial pH to 7.20-7.25, with a modification of MAP <15 mmHg during acidosis. Acidosis, defined as a pH <7.35, was left uncorrected for at least 20 min, and then 0.5 M bicarbonate was infused over a period of 6-14 min (mean: 11 min) to raise the pH at or above 7.35. 3) In the HCl sinodeafferented group, the carotid sinus nerves were sectioned bilaterally as stated above.

In vivo voltammetry. A carbon fiber electrode was lowered vertically through a burr hole toward the VLM (incisor bar 3.3 mm; zero, interaural line; anteroposterior 3.2 mm; lateral 1.9 mm; depth from brain surface 9.4 mm) with the rat prone in a stereotactic frame. In vivo, the auxiliary electrode was made of tungsten (125 µm; Goodfellow, Cambridge, UK). The reference electrode was changed at least once a week. Differential normal pulse voltammetry (8) (scan rate: 2 mV/0.4 s, scan potential: 240 mV/+160 mV, pulse amplitude: 30-50 mV, pulse duration: 40 ms, prepulse duration: 80-120 ms; Biopulse, Tacussel, Villeurbanne, France) allowed the recording every 2 min of a catechol oxidation current (CAOC), as a peak appearing at +40-50 mV in a restricted area covering a depth of 300 µm. This potential did not change when the beginning and end of recording were considered (see APPENDIX for details). The baseline value for the catechol signal was calculated as the mean height of the five catechol peaks recorded during the 10-min period preceding the challenge. Then changes in catechol signal were measured and reported as a percentage of the baseline value (100%). The amplitude of the peak, which is related to the oxidation of ascorbic acid (100 mV), did not change throughout the experiments in any of the animals. The catechol nature of the signal was ascertained in 15 of 19 traces obtained in the VLM by its complete disappearance after the administration of -methyl-p-tyrosine (250 mg/kg ip, n = 5; Sigma) or of pargyline (75 mg/kg ip, n = 10; Sigma) at the end of the experiments (7, 27). The amplitude of the catechol signal measured at the end of the baseline period was 0.62 ± 0.42 nA (n = 15 rats) in the RVLM (estimated in vivo catechol concentration 1.41 ± 0.56 µM). A hypertensive response was observed in all animals when direct current (+5 V, 30 s) was passed through the carbon fiber electrode at the end of the experiment in which the recording site was confirmed histologically to be in the RVLM (systolic pressure +20 ± 6 mmHg after pargyline or -methyl-p-tyrosine administration; n = 14).

Histology. The brain was frozen in isopentane (40°C) in a vertical position. Sections 20 µm thick were cut serially from the obex, defined as the closing of the fourth ventricle into the central canal (9). One of five sections was stained with cresyl violet. In each animal, the sections in which a lesion was recovered were orientated with respect to the obex, referenced at 100-µm intervals on the rostrocaudal axis from the obex (Fig. 1) and to the Paxinos and Watson atlas (20). Finally, the recording sites for which the center was recovered 1,000 µm, or beyond, rostral to obex were considered to belong to the RVLM. By contrast, when the center of the recording site was recovered <1,000 µm rostral to the obex, the recording sites were classified as belonging to the obex-VLM (9).

View larger version (23K): [in this window] [in a new window]   Fig. 1.   For histological analyses, sections were serially recovered from obex (junction of central canal with 4th ventricle) to reconstruct the position of the recording site and to refer it to the rostral ventrolateral medulla (RVLM) when its center was located 1,000 µm and beyond rostral to obex. Left to right: interaural 3.3 mm plate of the Paxinos and Watson atlas (20) corresponds to +1,000 µm rostral to obex. , Centers of recording sites referred to the nearest plate. Calibration 1 mm. Amb, nucleus ambiguus; PGi, paragigantocellular reticular nucleus; Su7, suprafacial nucleus.

Data analysis. Data are expressed as mean ± SD unless otherwise stated. The following variables were analyzed: CAOC, MAP, HR, H⁺ concentrations, and arterial PCO₂ and PO₂ (Pa_CO2 and Pa_O2, respectively) using a two-way analysis of variance for repeated measures, testing for time and study condition (saline, intact, and sinodeafferented groups). To assess the effectiveness of the carotid sinus deafferentation, the changes in respiratory rate observed after injection of chemical stimulants were compared with zero values, both before and after deafferentation using a t-test. P < 0.05 was considered significant. When analysis of variance showed significance, a post hoc analysis was performed with a Tukey's test.

	RESULTS

Top Abstract Introduction Methods Results Discussion Appendix References

Miscellaneous. Forty-six rats were included in this study. Twenty-two rats were excluded for either voltammetric (n = 5), ventilatory (n = 1), or circulatory (n = 5) instabilities; inability to achieve a reproducible H⁺ challenge (n = 4); miscellaneous problems (n = 2); or preliminary experiments (n = 5). In no instance was a decrease in catechol signal observed on HCl infusion. Twenty-four rats gave traces suitable for inclusion. One recording site could not be recovered. Recording sites were recovered in the RVLM (n = 14; center of recording site, mean: +1,500 µm rostral to the obex; range: 1,000-2,100 µm; Fig. 1) and the obex-VLM (n = 4, data not shown). The reversible changes observed in the potential of the current related to catechol oxidation during acid load are detailed in the APPENDIX.

Sinodeafferentation. In pentobarbital sodium-anesthetized rats maintained under spontaneous ventilation, changes in respiratory rate after lobeline and sodium cyanide injections were entirely suppressed after section of the carotid nerve (n = 5; Fig. 2; see Tables 1 and 2 and Figs. 2 and 3 for all quantitative or statistical details..

View larger version (11K): [in this window] [in a new window]   Fig. 2.   Effect of carotid sinus nerve (CSN) section on changes in respiratory rate (RR, cycles/min) after lobeline (A) or cyanide (B) injection in pentobarbital sodium-anesthetized spontaneously breathing rats (n = 5). Before deafferentation, changes in RR were different from zero (* P < 0.05). After deafferentation, changes in RR were not different from zero.

View larger version (26K): [in this window] [in a new window]   Fig. 3.   Effect of mild systemic acidosis on RVLM catechol signal and circulatory variables in intact (B) and sinodeafferented (C) animals. A: saline-injected animals. Data (means ± SE) are expressed as follows: 1) catechol oxidation current (CAOC) displayed every 2 min [P < 104 for study condition between the 3 groups and time; P < 0.05 from baseline values from +120 to +150 min within HCl intact group; P < 0.05 from baseline values from +50 to +150 min within HCl deafferented group; P < 0.05 from +120 to +150 min between HCl intact and saline; P < 0.05 from +50 to +60 and from +80 to +150 min between HCl deafferented and saline; not significant between intact and deafferented groups] and 2) mean arterial pressure (MAP; mmHg; P < 104 between the 3 groups for study condition and time) and heart rate (HR, beats/min; not significant) displayed every 5 min. HCl (1 N) was infused over 30-50 min to lower pH to 7.25-7.20 for at least 20 min with changes in MAP <15 mmHg. Then, HCO3-Na over 5-10 min restored pH >7.35. A catechol activation occurred, with or without carotid sinus nerve section.

Metabolic acidosis. Saline injection led to no difference in any circulatory or acid-base variables (Fig. 3A and Table 1, respectively). By contrast, infusion of HCl lowered the pH to 7.25-7.20 for at least 20 min (Table 1; lowest pH 7.24 ± 0.02 and 7.26 ± 0.05 at +40 min in intact or deafferented animals, respectively; lowest individual pH 7.22 and 7.21 in intact and deafferented animals, respectively). In no instance, did the pH revert back to 7.35 or above spontaneously. Minor but significant changes in Pa_CO2 were observed between the three groups [not significant (NS) between intact and deafferented rats]. Overall differences in pressure were significant between the three groups. Changes in heart rate were not significant. A significant increase in the RVLM catechol signal followed within HCl intact and deafferented groups (n = 5 in each group; Fig. 3, B and C). No significant difference in the kinetics and amplitude of the RVLM catechol signal was observed between the intact and deafferented groups. This increase in catechol signal was irreversible despite restoration of systemic pH >7.35 with HCO₃-Na (intact: 7.7 ± 1.4 ml/kg; deafferented: 7.6 ± 2.1 ml/kg; NS).

	DISCUSSION

Top Abstract Introduction Methods Results Discussion Appendix References

After mild systemic prolonged isocapnic acidosis, increased catechol activity was observed in the RVLM, with arterial pressure maintained constant during the acid challenge. These data demonstrate an impact of H⁺ stimuli on RVLM catechol metabolism.

Methodology. First, the RVLM catechol signal appears to be closely linked to the activity of the rate-limiting enzyme for the synthesis of catecholamines, tyrosine hydroxylase (7), located in adrenergic cell bodies (6). Two lines of evidence are to be considered here: 1) RVLMbb neurons present very few collaterals (31) and 2) A1 and A5 noradrenergic groups present few or no projections onto the RVLM (4, 33). Thus the catechol signal originates most likely from a catechol metabolite (7) diffusing from adrenergic cell bodies or their dendrites. Second, the comparison among in vivo electrophysiological, electrochemical, and microdialysis data (2, 14, 15, 23, 33) indicates that the RVLM catechol signal gives a reasonable estimate of the direction of changes in firing rate of adrenergic RVLM neurons, although 1) the kinetics of the two signals are different (brief stimuli and fast response vs. long stimuli and slow onset) and 2) no inference can be made on the pattern of discharge itself, but only on the mean firing rate (25). Third, the changes in systemic pH observed here are followed by changes in medullary extracellular pH of similar amplitude (19). Fourth, the changes in H⁺ observed during acid infusion are compatible with a pure metabolic acidosis without changes in CO₂ between the intact and deafferented rats. Fifth, rats with unstable circulatory profiles were excluded from the analysis, thus in keeping with our aim to minimize changes in pressure during the acid load. Indeed, hypotension by as little as 15 mmHg previously led to a baroreflex-linked increase in the catechol signal (27). In keeping with the minimal changes in pressure observed here, only minor changes in pressure were observed in cats subjected to isocapnic metabolic acidosis (19). Thus the increase in catechol signal observed during the acid load itself is not contaminated by baroreflex influences (Fig. 3, B and C).

Metabolic acidosis. The major result is an increase in catechol metabolism during acid challenge. The present RVLM catechol-specific activation is fully in line with a sensitivity 1) in vivo, of the intermediate area of the ventral surface of the medulla (30) and 2) in vitro, of VLM neurons after acidosis of the perfusate (5). The comparison of previously published figures (5, 34) allows one to infer that H⁺-sensitive cells are immediately caudal to RVLM bulbospinal neurons that exhibit pacemaker properties in vitro (34). Furthermore, the catechol activation observed after prolonged systemic acidosis concurs with the minor sympathetic nerve activation observed after isocapnic systemic metabolic acidosis (19) or when the RVLM itself is briefly perfused with acidified solutions (32). An indication with respect to the relative specificity of the present activation is provided by the fact that the catechol signal recorded in the dopaminergic midbrain A10 area (ventral tegmental area) was not modified by respiratory acidosis (24).

The increase in RVLM catechol signal after systemic acidosis is at variance with the lack of change of activity of RVLMbb neurons iontophoresed locally with H⁺ (35). This apparent discrepancy may be explained as follows. First, iontophoretic application of H⁺ may be insufficient to activate the cells due to the localized nature of the applied ion. Second, iontophoretic stimulus is brief as opposed to the 20-min acidosis used here. Third, the histochemical specificity of the neurons warrants consideration. The catechol signal is an integrated, catechol-specific index that reflects the global activity of RVLM adrenergic neurons. The RVLM catechol signal cannot differentiate between several contributions, namely 1) adrenergic RVLMbb neurons; 2) adrenergic barosensitive neurons with rostral projections, e.g., RVLM-barosensitive neurons projecting to the locus ceruleus (11, 12); and 3) in addition, an adrenergic specificity of some RVLM preinspiratory bulbospinal neurons (36) cannot be excluded. In contrast, RVLMbb neurons iontophoresed with H⁺ were not identified according to their conduction velocity and their clonidine sensitivity (35). Furthermore, their baseline firing rate was high (~14-18 Hz) (35). If slow-conducting RVLMbb neurons are only adrenergic and fast-conducting RVLMbb neurons are only glutamatergic (10, 31), then an interpretation of the discrepancy between previous (35) and present results is as follows: 1) adrenergic cells involved in the respiratory generator or adrenergic RVLMbb neurons do respond to acidosis and 2) in contrast, presumed glutamatergic RVLMbb cells do not respond to systemic acidosis. An in-depth electrophysiological analysis of slow- and fast-conducting RVLMbb neurons after systemic and local H⁺ warrants consideration.

The second result pertains to the similarity of catechol activation irrespective of carotid sinus nerve section. No changes in respiratory rate were observed in spontaneously ventilated rats after sinodeafferentation (Fig. 2), in line with previous reports (29). Therefore, the changes in RVLM catechol signal are primarily centrally mediated. However, this statement has to be qualified: in the HCl intact rats, the catechol signal rose steadily from +10 min onward. In the HCl deafferented rats, the CAOC rose after +25 min. Although this did not achieve statistical difference between groups, the involvement of carotid bodies and/or a slower development of acidosis within the cerebrospinal fluid after systemic acid challenge cannot be excluded.

The third result pertains to the irreversibility of RVLM catechol activation after mild and reversible systemic acidosis. First, irreversible increases in catechol signals have been reported several times previously. 1) An irreversible increase occurred in the RVLM when severe hypotension was produced without monitoring the acid-base equilibrium (7). 2) An irreversible increase occurred in the RVLM when systemic acidosis superimposed itself on recovery from hypotension (3). In this experiment, the catechol activation was reversible only if systemic acidosis was absent (3). 3) The RVLM catechol activation observed on emergence from anesthesia was reversible on readministration of anesthesia only if systemic acidosis was absent (26). 4) Recordings performed in the obex-VLM (n = 4, data not shown) on identical HCl challenge led to an irreversible increase identical to the one observed in Fig. 3 (B and C). 5) An irreversible increase in catechol signal was observed previously when recordings were performed in the caudal VLM or the locus ceruleus (Ref. 21 and unpublished data) after systemic metabolic acidosis. Thus the irreversibility of the RVLM catechol activation on metabolic acidosis is beyond doubt. As the RVLM catechol signal was increased in a fully reversible manner by respiratory acidosis (24), the present phenomenon is probably specific to noradrenergic/adrenergic hindbrain areas on metabolic acidosis. Second, perfusion of the RVLM itself with acidified solution led to changes in sympathetic activity that were much longer than observed after perfusion with CO₂-bubbled solutions (32). Third, during the acid challenge itself, pressure was fully maintained; thus the catechol activation noticed during acid challenge is not likely to be contaminated by baroreceptor inputs. In contrast, after the acid challenge, the arterial pressure declined by ~20 mmHg in HCl intact and deafferented rats. This delayed decline did not become significant between groups when stringent post hoc analysis was used. However the catechol signal is curvilinearly related to pressure (27). Thus the small and nonsignificant drop in arterial pressure observed after acid challenge (+60/+150 min) may contribute, among other factors, to the irreversible significant activation of catechol metabolism observed after acid challenge (+60/+150 min). No attempt was made, in the present study, to maintain arterial pressure strictly at baseline level after acid challenge. Indeed, phenylephrine, used to maintain pressure on metabolic acidosis in a previous design, led to a further increase in RVLM catechol signal (3).

As maintenance of pressure with phenylephrine (3) or a nonsignificant decline in pressure (Fig. 3) led to irreversible catechol activation, a final question is raised as to the mechanism of such irreversible activation. As the unique goal of the present experiment was restricted to a documentation of catechol activation under pure H⁺ stimulus, no definitive answer can be offered. First, the stimulus provoking the rise in catechol signal may not be the H⁺ ion itself, but some element correlated with the ion in an undetermined manner. Second, lipid peroxidation and neuronal damage may have occurred during systemic acidosis. However, this is unlikely as the amplitude of the ascorbate signal was unchanged; this ascorbate signal was recorded here simultaneously with the catechol signal (3, 8) and is a very sensitive index of neuronal death. A third explanation would be a reduced catechol catabolism secondary to acidosis; however, the disappearance half-life of the RVLM catechol signal is unaffected by severe hypotension or hypovolemia (7), during which metabolic acidosis is likely. A fourth explanation would be that HCO₃-Na buffered only the extracellular space, leaving intact a prolonged intracellular acidosis. In keeping with this speculation, a slower recovery of intracellular pH was observed in VLM-chemosensitive neurons as opposed to nonchemosensitive neurons after hypercapnia (28).

Implications. An involvement of central catecholamines within the respiratory-cardiovascular coordination has been postulated (3, 16, 24). When the respiratory-cardiovascular coordination, i.e., the sympathetic chemoresponse, is considered, the present data allow one to go two steps further: 1) adrenergic RVLM neurons do respond to systemic H⁺ stimulus, and 2) catecholaminergic neurons located at obex-VLM (n = 4, data not shown) respond also to a systemic H⁺ stimulus. Adrenergic neurons, lying around small capillaries (18), may convey information with respect to oxygen (10, 11), H⁺ (Ref. 3 and present data), and CO₂ (24) to the midbrain via their rostral collaterals (11, 12) and to the sympathetic preganglionic neurons through their spinal axons. The prolonged catechol activation observed after mild systemic acidosis is in line with the suggestion that delayed recovery of intracellular pH in chemosensitive VLM neurons tends to maximize the response of the respiratory generator (28). The prolonged catechol activation observed after long mild systemic acidosis may be congruent with the prolonged circulatory and ventilatory activation observed after exercise or low flow in trauma or septic patients, as opposed to short physiological stimuli (5, 32).

RVLM catechol activation occurs after mild systemic acidosis. Thus adrenergic RVLM neurons are sensitive to changes in arterial H⁺. As such, adrenergic bulbospinal cells may be one of the central chemosensors for H⁺, themselves, or simply one relay within the sympathetic chemoresponse.