Dialysis Delivery of an Adenosine A¹Receptor Agonist to the Pontine Reticular Formation Decreases Acetylcholine Release and Increases Anesthesia Recovery Time

All procedures involving animals were reviewed and approved by the University of Michigan Committee on Use and Care of Animals (Ann Arbor, Michigan) and strictly adhered to the recommendations in the Guide for the Care and Use of Laboratory Animals . 19Four adult male cats were anesthetized with isoflurane and implanted with electrodes to enable objective recording of arousal states. Multiple lines of previous evidence made the cat the logical choice for these preclinical studies. First, most of the cellular-level data regarding brain stem regulation of arousal states have been derived from the cat. 12,20,21Second, use of the cat made it possible to interpret the current results relative to evidence that pontine acetylcholine release is significantly altered by opioids, 15halothane, 14and ketamine. 16Third, the loss of wakefulness associated with sleep or anesthesia causes respiratory depression. An extensive body of evidence from cats shows that alterations in pontine cholinergic neurotransmission cause upper airway hypotonia, decreased minute ventilation, and diminished responsiveness to hypercapnia (reviewed in Lydic and Baghdoyan 13). Finally, previous neurobiological studies of cats support the working hypothesis that neuronal networks that have evolved to generate naturally occurring states of sleep and wakefulness contain neurons that, when altered by anesthetic molecules, generate the physiologic and behavioral traits comprising anesthetic states. 22,23Thus, the current use of the cat provided another opportunity to evaluate the hypothesis that sleep neurobiology contributes novel insights into the mechanisms of anesthetic action.

As described in detail elsewhere, 24,25implanted electrodes permit long-term recording of electroencephalogram, electromyogram, and eye movements (electrooculogram). These electrodes were encased in acrylic and permanently affixed to the calvarium. Cylinders were imbedded in the acrylic for attachment to a Kopf 880 (David Kopf Instruments, Tujunga, CA) head holder. The Kopf 880 cylinders prevent nociceptive stimulation by permitting subsequent stereotaxic positioning without touching tissue. A craniotomy with a removable cover provided brain access for a microdialysis probe. 15Cats were allowed 3 weeks to recover before beginning experiments.

The probes used for microdialysis (CMA/Microdialysis, Stockholm, Sweden) had a 70-mm shaft and a polycarbonate membrane (2 mm length, 0.5 mm diameter, 20 kd pore size). Probes were perfused continuously with Ringer's solution at a flow rate of 3 μl/min using a CMA/100 pump. Before each experiment, a standard curve was created with six concentrations of acetylcholine ranging from 0.1 to 1 pmol. Prior to the in vivo  portion of each experiment, preexperimental probe recoveries were determined by in vitro  dialysis of a known concentration of acetylcholine. These probes typically have an acetylcholine recovery rate of approximately 8%. The true amount of pontine acetylcholine is unknown, and the acetylcholine values are therefore expressed as pmol/10 min of dialysis. Likewise, the precise amount of pontine tissue affected by dialysis delivery of adenosinergic drugs is unknown. One of the advantages of microdialysis over microinjection is that dialysis drug delivery influences a smaller area of tissue than microinjection. 26Diffusion rates and tissue area affected by drug vary as a function of drug properties and tissue. In the medial pontine reticular formation, microinjection effects have been shown to vary within 1 mm. 27One may speculate that microdialysis affects a more limited region of tissue surrounding the probe.

Each 30-μl sample was injected into an LC4C amperometric detector (Bioanalytical Systems, West Lafayette, IN). The electroactive product (H²O²) produced by an immobilized enzyme reactor column passed over the electrochemical detector and generated chromatograms proportional to the amount of acetylcholine in the dialysis sample. The area under each chromatographic peak was integrated with a computer using the ChromGraph Report and Control programs (Bioanalytical Systems). The percentages of acetylcholine recovery from the probe were compared by t  test to values obtained after removing the probe from the brain. This comparison of preexperimental and postexperimental probe recoveries ensured that variation in acetylcholine reflected neurochemical changes rather than changes in performance of the dialysis probe.

The adenosine receptor agonist N⁶-p-sulfophenyladenosine (SPA; Sigma-Aldrich Corp., St. Louis, MO) was used because of its high selectivity for A¹versus  A²receptors (affinity ratio of A¹/A²= 120). 28SPA was prepared in Ringer's solution in concentrations of 0.088 or 8.8 mm. The adenosine receptor antagonist 8-cyclopentyl-1, 3-dipropylxanthine (DPCPX; Sigma-Aldrich) shows a 700-fold selectivity for the A¹receptor versus  the A²receptor. 29DPCPX was dissolved in dimethyl sulfoxide (Sigma-Aldrich) and then diluted to 100 μm in Ringer's solution, with a final dimethyl sulfoxide concentration of 0.008%. Antagonist + agonist coadministration consisted of 8.8 mm SPA and 100 μm DPCPX prepared in Ringer's solution. Drug solutions were made fresh prior to every experiment. Drug solutions were filtered, pH adjusted to 6.0 ± 0.4, maintained at room temperature, and protected from light. These drug concentrations were chosen as equimolar to the concentration of cholinergic agonist shown by previous studies to alter sleep 30or acetylcholine release 31when administered into the pontine reticular formation. A limitation of microdialysis drug delivery is the inability to specify the quantity of administered drug. Since the current dialysis probes exhibit an 8% acetylcholine recovery, one can assume dialysis drug delivery of approximately 8% of the dialyzed concentration. If this assumption is correct, the delivered drug concentrations would have been approximately 7 or 700 μm SPA and 8 μm DPCPX.

Each experiment began by anesthetizing a cat with 3% halothane. The trachea then was intubated with a cuffed pediatric endotracheal tube (size 3.5–4.5 French), and the cat was placed in a Kopf stereotaxic frame and connected to a mechanical ventilator. End-tidal halothane and carbon dioxide fractions were monitored continuously with an Ohmeda Rascal II spectrometer (Datex-Ohmeda, Madison, WI). Ventilation and inspired halothane concentration were adjusted to maintain end-tidal halothane at 1.5% and end-tidal carbon dioxide at 30 mmHg. Oxygen saturation and heart rate were monitored with an Ohmeda Biox 3700 Pulse Oximeter (Datex-Ohmeda). Blood pressure was monitored using a Critikon Dinamap (Critikon, Tampa, FL), and body temperature was measured with a rectal thermometer (Yellow Springs Instrument Co, Yellow Springs, OH) throughout the experiment. A T500 T/Pump Heat Therapy System (Gaymar, Orchard Park, NY) was used to maintain body temperature at 37°C. A microdialysis probe was aimed for the mPRF or cerebellar cortex using a stereotaxic atlas. 32Stereotaxic coordinates for the mPRF dialysis sites were posterior 1–3, lateral 1–3, and horizontal −5. Microdialysis sites in the cerebellar cortex were posterior 10, lateral 2–3, and horizontal 5.

This study used an intensive within-subjects design. Each experiment was divided into two phases (fig. 1). Care was taken to ensure that evaluation of the effects of adenosinergic compounds was not confounded by significant variations in duration of halothane anesthesia (fig. 1, lines A–D). The first phase involved microdialysis for simultaneous collection of acetylcholine and drug delivery (fig. 1, lines B–C). The second phase involved quantification of postanesthetic recovery time (fig. 1, lines D–E). Prior to the microdialysis portion of the study, the probe was positioned stereotaxically in the mPRF (fig. 1, lines A–B). The probe was perfused with one of four different solutions: Ringer's solution alone, 0.088 mm SPA, 8.8 mm SPA, or SPA (8.8 mm) plus DPCPX (100 μm). During drug delivery, sequential 30-μl mPRF dialysis samples were collected on ice for quantifying acetylcholine release (fig. 1, lines B–C). The duration of sample collection (fig. 1, lines B–C) always was 180 min. After the microdialysis phase of the experiment was complete (fig. 1, line C), the dialysis probe was removed from the brain (fig. 1, lines C–D). Time required for removal of dialysis probe to cessation of halothane (fig. 1lines C–D) averaged approximately 15 min. After the halothane anesthesia was discontinued (fig. 1, line D), recovery time from anesthesia was quantified (fig. 1, lines D–E). The results quantify variations in time for recovery (interval between lines D and E in fig. 1) as a function of dialysis drug delivery (conditions shown between lines B and C in fig. 1).

To evaluate postanesthetic emergence, a recovery score developed for humans 33was modified for use in the current study. The recovery score was determined at each of the following time points after turning off the halothane: every min for the first 10 min, every 5 min up to 1 h, every 15 min up to 1.5 h, every hour up to 4 h, and at 24 h. This resulted in a total of 26 time points per experiment at which anesthetic recovery was evaluated. The five criteria for recovery score were provided by measures of (1) motor activity, (2) respiration, (3) blood pressure (BP) and heart rate (HR), (4) responsiveness, and (5) oxygen saturation and skin color. These variables were evaluated and scored as follows:

1. Motor activity: Spontaneous motor activity was scored as 0 if no movement was observed; 1 when the head or one or two limbs moved; and 2 when all four limbs moved and the cat started walking.

2. Respiration: No spontaneous respiratory activity was scored as 0; apparent dyspnea, superficial breathing, or effort to breathe were scored as 1; and deep, regular breathing was scored as 2.

3. Blood pressure and heart rate. Systolic arterial BP and/or HR variation greater than 50% of the initial values was scored as 0; a 20–50% variation from initial values was scored as 1; and less than 20% variation from initial values was scored as 2.

4. Responsiveness: Failure to respond to auditory stimulation (head or eye movement toward the sound) was scored as 0; any behavioral sign of responsiveness to auditory stimulation was scored as 1; and full alertness and responsiveness to visual and auditory stimuli was scored as 2.

5. Oxygen saturation and skin color: Cyanosis was scored as 0; alterations from the normal color, such as pale, dusky, or blotchy discoloration, were scored as 1; and normal pink skin coloration and/or oxygen saturation values above 90% were scored as 2.

For each experiment, at each time point, the total recovery score was determined by summing the scores for dependent measures 1–5. Thus, the maximal score possible was 10 and indicated complete recovery from anesthesia.

On completion of in vivo  experiments, brains from all animals were examined histologically to confirm dialysis probe placement in the mPRF or cerebellar cortex. Cats were deeply anesthetized with sodium pentobarbital (35–40 mg/kg intraperitoneal) and perfused transcardially with isotonic saline followed by 10% formalin. Brains were removed, fixed, serially sectioned (40 μm thick), and stained with cresyl violet for histologic localization of microdialysis probe sites. All tissue sections containing dialysis probe–induced lesions were digitized and compared with a brain stem atlas. 32 

Postanesthetic recovery scores are reported as mean ± SD. Recovery scores from experiments in which the same drug was administered were averaged at each of the 26 time points. Linear regression analysis was used to quantify recovery score as a function of time for Ringer's solution (control), SPA, and SPA + DPCPX dialysis conditions. Analyses were performed using multiple linear comparisons to evaluate significant differences in slopes and intercepts (GraphPad Prism 3.0 Software Inc., San Diego, CA). One-way analysis of variance (ANOVA) and Tukey-Kramer procedures compared recovery scores following Ringer's solution, 0.088 mm SPA, and 8.8 mm SPA at 15 min, 45 min, and 24 h after drug dialysis completion. Acetylcholine release as a function of dialysis drug delivery was evaluated using ANOVA and the Tukey-Kramer procedure. Probability values of P < 0.05 were regarded as statistically significant.

The pontine region where drugs were delivered by microdialysis is illustrated in figure 2. Histologic analyses confirmed that all pontine dialysis sites were localized to the mPRF, which lies within the gigantocellular tegmental field as described by Berman. 32 

Figure 3shows that dialysis delivery of 8.8 mm SPA into the mPRF significantly increased time required for recovery from halothane anesthesia compared to recovery time following dialysis with Ringer's solution alone. The three functions shown in figure 3were determined from recovery scores obtained after dialysis with Ringer's solution, 0.088 mm SPA, or 8.8 mm SPA (fig. 1, lines B–C). The function labeled Ringer's had a slope of 0.18 and an intercept of 5.4, representing the control profile of emergence from halothane anesthesia (fig. 3). Following control dialysis with Ringer's solution, all cats completely recovered in 30 min after delivery of halothane ended. Dialysis delivery of 0.088 mm SPA produced a recovery score profile similar to the control, with a slope of 0.14 and an intercept of 4.95 (fig. 3). All cats that received the low concentration of SPA (0.088 mm) recovered within 30–45 min after cessation of halothane. Dialysis delivery of 8.8 mm SPA increased postanesthetic recovery time. The recovery score determined following treatment with 8.8 mm SPA had a slope of 0.05 ± 0.007 and a y-axis intercept of 3.24 ± 0.19, reflecting the delayed emergence from anesthesia (fig. 3). Multiple regression comparisons indicated a significant difference (F = 22.7;df = 2, 83;P < 0.0001) between the recovery score slopes for SPA (8.8 mm) and Ringer's solution (control). The histograms of figure 4show that dialysis delivery of 8.8 mm SPA significantly delayed recovery for up to 4 h following discontinuation of halothane. The figure 4recovery scores were obtained during the interval indicated in figure 1between lines D and E.

The SPA-induced delay in recovery from halothane anesthesia was significantly reduced by coadministration of the adenosine A¹receptor antagonist DPCPX. Figure 5summarizes the functions generated by linear regression analysis of recovery scores obtained after dialysis with Ringer's solution, DPCPX + SPA, and SPA alone. By min 60, the mean recovery score following administration of DPCPX + SPA was 9.67, indicating complete postanesthetic recovery. The linear function representing postanesthetic recovery scores following DPCPX + SPA had a slope of 0.09 ± 0.01 and an intercept of 4.89 ± 0.27. These values were intermediate between slope and intercept values for Ringer's solution and 8.8 mm SPA. Multiple linear regression analysis demonstrated that DPCPX and SPA coadministration caused a significantly shorter (F = 21;df = 2, 78;P < 0.0001) wake-up time than 8.8 mm SPA dialysis alone.

Figure 6compares recovery scores obtained following cerebellar SPA administration with recovery scores obtained following mPRF Ringer's solution (control) and mPRF SPA administration. Following dialysis delivery of SPA into cerebellum, two cats achieved complete recovery by min 30, and the third cat recovered completely by min 45. Multiple linear regression analyses demonstrate that cerebellar administration of SPA failed to delay recovery from anesthesia, and that the recovery score was significantly different from the delayed recovery caused by 8.8 mm SPA delivered to mPRF (F = 20.53;df = 2, 75;P < 0.0001). Thus, the delay in recovery was significantly dependent on SPA administration into the medial pontine reticular formation. There was no significant difference in the duration of halothane administration between experiments that dialyzed the mPRF with Ringer's solution (control), experiments that dialyzed the mPRF with SPA or SPA + DPCPX, and experiments that dialyzed the cerebellar cortex with SPA.

Figure 7Aplots acetylcholine release measured during a typical experiment in which dialysis delivery of 8.8 mm SPA decreased mPRF acetylcholine. Figure 7Bsummarizes mPRF acetylcholine release during dialysis with Ringer's solution (control) and during dialysis delivery of 0.088 or 8.8 mm SPA. ANOVA revealed a significant SPA-induced decrease in acetylcholine release (F = 22.4;df = 2, 117;P < 0.0001). Dialysis with 8.8 mm SPA (fig. 7B, solid bar) decreased acetylcholine release to 78.6% of control (fig. 7B, hatched bar). Multiple comparison statistic showed that the lower concentration of SPA (0.088 mm) did not significantly alter acetylcholine release (fig. 7B, shaded bar). Figure 8shows that coadministration of DPCPX with SPA blocked the SPA-induced decrease in mPRF acetylcholine release (F = 39.9;df = 2, 102;P < 0.0001).

Adenosine is relevant to anesthesiology in part because of compelling evidence that adenosine alters hyperalgesia and neuropathic pain. 34–38Adenosine infusion reduces the requirement for isoflurane, 39and the current results show that time required for recovery from halothane anesthesia was significantly increased by pontine delivery of an adenosine A¹receptor agonist. Normal brain acetylcholine concentrations are required for electroencephalographic and behavioral arousal. 40Thus, the finding that pontine acetylcholine was decreased by the adenosine A¹receptor agonist is considered below as one potential mechanism modulating the delayed recovery from halothane anesthesia.

N -methylated xanthine molecules such as caffeine, theobromine, and theophylline all promote arousal. Caffeine ingestion by humans increases latency to sleep onset and reduces the electroencephalographic δ power of non-REM sleep. 41The ability of caffeine to promote wakefulness is mediated in part by A¹adenosine receptors. 5The data reported here show a significant delay in posthalothane recovery caused by dialysis delivery of the adenosine A¹receptor agonist SPA into the medial pontine reticular formation (figs. 3 and 4). The results provide additional support for the view that medial regions of the pontine reticular formation, known to modulate sleep, 12also contribute to the loss of wakefulness caused by a chemically diverse set of anesthetic agents. As noted previously, acetylcholine release in the pontine reticular formation is significantly decreased by morphine, 42halothane, 14fentanyl, 15and ketamine. 16 

The results are consistent with the interpretation that the SPA-induced increase in postanesthetic recovery time can be modulated by adenosine A¹receptors. The delay in postanesthetic recovery caused by dialysis delivery of the adenosine A¹receptor agonist varied significantly as a function of agonist concentration (figs. 3 and 4) and was blocked by coadministration of the adenosine A¹receptor antagonist DPCPX (fig. 5). Dialysis delivery of the agonist into the cerebellar cortex failed to prolong emergence (fig. 6). Thus, the delayed recovery from halothane anesthesia was site-specific to the mPRF, varied significantly with SPA concentration, and was blocked by an adenosine agonist.

Studies of sensory afferent neurons, central nervous system neurons, and efferent motor systems show that adenosine A¹receptors alter cholinergic neurotransmission. In sensory neurons, purinergic neuromodulators such as ATP and acetylcholine can activate similar nonselective cation channels via  common intracellular pathways. 43Adenosine A¹receptors inhibit synaptic transmission by decreasing N-type Ca²⁺current. 44Adenosine A¹receptors exert an inhibitory action on Ca²⁺entry coupled to acetylcholine release in retina cells. 45In motor systems, ATP suppresses acetylcholine release via  protein kinase C–coupled purinergic receptors, and adenosine alters quantal acetylcholine release. 46In central nervous system neurons, acetylcholine release is inhibited by adenosine A¹receptors in optic tectum, 47striatum, 48cortex, 49,50and hippocampus. 51 

By what cellular mechanisms might adenosine alter acetylcholine release and arousal? Adenosine inhibits electroencephalographic and behavioral arousal, which are modulated by forebrain and brain stem cholinergic neurons. Basal forebrain cholinergic neurons provide acetylcholine to the cortex, and cortically activated states of wakefulness and REM sleep are associated with greater concentrations of cortical acetylcholine than are states of NREM sleep. 52Measures of basal forebrain acetylcholine also show that acetylcholine release is greater during wakefulness and REM sleep than during NREM sleep. 53These microdialysis data are consistent with studies showing that administering adenosine to the basal forebrain decreases wakefulness. 7,54In contrast, wakefulness is increased by administering an adenosine A¹receptor antagonist to the basal forebrain. 55The foregoing neurochemical studies are supported by electrophysiological evidence. Putatively cholinergic neurons of the magnocellular basal forebrain exhibit the greatest discharge frequency during wakefulness and REM sleep. 56The discharge of these wake-active basal forebrain neurons is inhibited by an adenosine A¹receptor agonist. 57 

Cholinergic neurons in the pontine brain stem also modulate electroencephalographic and behavioral arousal. 13,58These cholinergic neurons are located in the laterodorsal tegmental and pedunculopontine tegmental nuclei (LDT/PPT) and provide acetylcholine to the mPRF. Neurons in the mPRF do not produce acetylcholine and, in the current study, the acetylcholine measured in the mPRF is known to originate from LDT/PPT neurons. 59,In vitro  electrophysiological recordings found that adenosine A¹receptors cause postsynaptic inhibition in approximately two thirds of LDT/PPT neurons. 60LDT/PPT neuronal excitability also is presynaptically disfacilitated by adenosine A¹receptors on presynaptic glutamatergic terminals that innervate the LDT/PPT. 61Microdialysis delivery of adenosine to the LDT/PPT decreases wakefulness. 7The foregoing evidence fits well with the current finding that dialysis delivery to the mPRF of the adenosine A¹agonist SPA decreased acetylcholine release in the mPRF (figs. 7 and 8). These results also are consistent with the sleep-enhancing effects of microinjecting an adenosine agonist into the pontine reticular formation. 8 

The current results provide a novel extension of the sleep-promoting effects of adenosine by showing that an adenosine A¹agonist delays resumption of wakefulness following halothane anesthesia. This extension is consistent with a potentially larger relevance of the current findings for efforts to elucidate the specific neurons and molecules causing traits that characterize the state of anesthesia. The current data support the working hypothesis that adenosine A¹receptors in medial regions of the pontine reticular formation, known to modulate sleep states, also contribute to the generation of anesthetic states. 22,23 

Four adenosine receptors (A¹, A^2a, A^2b, A³) have been cloned, 44and use in this study of the agonist SPA restricts conclusions to inferences regarding the A¹subtype. Adenosine A¹receptors inhibit the release of “virtually every classic neurotransmitter (including glutamate, γ-aminobutyric acid, acetylcholine, norepinephrine, 5-hydroxytryptamine, dopamine, and other transmitters as well)”. 44Dialysis delivery of SPA was limited to concentrations that were ineffective (0.088 mm) or effective (8.8 mm) in decreasing acetylcholine release. A complete concentration–response curve is required to determine potency and maximal efficacy of SPA. Limitations associated with applying Aldrete and Kroulik's 33clinical assessment tool to the evaluation of feline posthalothane emergence also are acknowledged. The current data, however, demonstrate the face validity of modifying the human recovery score 33for operationally quantifying posthalothane recovery in felines (fig. 4). The results encourage future studies aiming to determine if quantitative electroencephalographic analyses can provide a more objective assessment of SPA-induced alterations in levels of arousal. 62 

The current data cannot specify the cellular mechanisms by which the adenosine agonist SPA prolonged the recovery from halothane anesthesia for 4 h (figs. 3–6). Available data, however, permit speculation regarding the long duration of SPA effects on anesthetic recovery time and acetylcholine release. The adenosine A¹receptor is coupled to an inhibitory guanine nucleotide–binding (Gi) protein which inhibits adenylyl cyclase and activates calcium and potassium channels. 63One function of G proteins is signal amplification in the time domain, which may account for the long-duration SPA actions. SPA causes a concentration-dependent activation of G proteins in the medial pontine reticular formation. 64Although acetylcholine is a neurotransmitter that alters cell excitability with a time course of milliseconds, the cholinergic regulation of sleep and wakefulness has a time course of minutes and hours. 25,30The modulation of arousal by muscarinic cholinergic receptors in the mPRF involves a signal transduction cascade comprised of adenylate cyclase, a pertussis toxin–sensitive G protein, cyclic AMP, and protein kinase A. 65–67This cholinergic transduction pathway would be anticipated to cause slow and long-lasting changes in neuronal excitability. 12The clinical relevance of these cholinergic pathways is illustrated by the ability of acetylcholinesterase inhibitors to reverse propofol-induced unconsciousness. 68Thus, the current results suggest the working hypothesis that SPA caused a long-lasting depression in acetylcholine and anesthetic recovery time by interacting with cholinergically activated G proteins. This hypothesis is testable and consistent with data showing that in the pons, SPA and the μ-opioid agonist DAMGO cause a partially additive activation of G proteins. 64 

The authors thank Grant Bowman (Research Assistant, Department of Anesthesiology, University of Michigan, Ann Arbor, Michigan), Daniela Hangan, M.D. (Research Assistant, Department of Anesthesiology, University of Michigan), and Carrie A. Lapham, B.A. (Research Secretary IV, Department of Anesthesiology, University of Michigan), for assistance.