General Anesthetics Do Not Affect Release of the Neuropeptide Cholecystokinin from Isolated Rat Cortical Nerve Terminals

These studies were approved by Weill Medical College of Cornell University Institutional Animal Care and Use Committee (New York, New York).

CCK8s was obtained from RBI (Natick, MA); Percoll was obtained from Pharmacia (Uppsala, Sweden); isoflurane was obtained from Abbott Laboratories (North Chicago, IL); and halothane (thymol-free) was obtained from Halocarbon Products (River Edge, NJ). Propofol was purchased from Aldrich Chemicals (Milwaukee, WI) or was a gift from AstraZeneca Pharmaceuticals (Wilmington, DE), and etomidate was a gift from Janssen Biotech n.v. (Olen, Belgium). Pentobarbital, thiopental, ketamine, polymeric L-glutamic acid (poly-Glu), 1-ethyl-3-dimethylaminopropylcarbodiimide hydrochloride, horseradish peroxidase (HRP)–labeled goat antirabbit IgG, rabbit anti-CCK8s IgG, bacitracin, Tween 20, veratridine, tetrodotoxin, ω-conotoxin MVIIC, dimethyl sulfoxide (DMSO), o -phenylenediamine, tablets for preparation of sodium carbonate-bicarbonate buffer pH 9.6, and HRP substrate solution (phosphate-citrate buffer with sodium perborate) were from Sigma Chemical Co. (St. Louis, MO). ω-Agatoxin IVA and ω-conotoxin GVIA were obtained from Alamone Labs (Jerusalem, Israel). All other chemicals were of analytical grade. Spectra/Por 12,000–14,000 dialysis membrane was obtained from Spectrum (Houston, TX); bovine serum albumin (fraction 5) was obtained from J.T. Baker (Phillipsburg, NJ); Nunc-immuno MaxiSorp flat-bottom modules (No. 469949), CoStar Spin-X centrifuge tube filters with 0.45-μm cellulose acetate membrane (Corning No. 8163), CoStar 96-well microtiter plates (No. 3797), and 0.4-ml polypropylene microcentrifuge tubes (No. 20170-326) were obtained from VWR Scientific Products (Bridgeport, NJ); and multiplates for filtration (No. MAFBNOB10) were obtained from Millipore (Bedford, MA).

CCK8s was coupled to poly-Glu to form CCK8s-polyglutamic acid (CCK8s-poly-Glu). 23A solution of poly-Glu (1.4 mg) in 1.3 ml water was added to a solution of 1-ethyl-3-dimethylaminopropylcarbodiimide hydrochloride (3.9 mg) in 0.39 ml water and stirred. A solution of CCK8s (0.2 mg) in 0.2 ml water was added over 45 s with mixing and left for 10 min at room temperature. The pH was adjusted to 6.0 with 0.27 ml sodium phosphate buffer, 1 m (pH 5.0), and the solution was stirred for 1 h at room temperature. The solution was diluted to 4.4 ml and dialyzed against 2 l water for 5 days at 4°C with a daily change of the dialysate. The dialyzed solution of CCK8s-poly-Glu conjugate was divided into aliquots, lyophilized, and stored at −70°C.

Purified synaptosomes were prepared from rat cerebral cortex by the Percoll gradient method 24as described. 25The synaptosome fraction was washed from Percoll with HEPES-buffered medium (HBM) composed of the following: 130 mm NaCl, 3 mm KCl, 5 mm NaHCO³, 1 mm MgCl², 1.2 mm Na²HPO⁴, 10 mm d-glucose, and 20 mm HEPES, titrated to pH 7.4 with Tris base. Protein content was measured using the Bio-Rad Protein Assay Kit (Hercules, CA) based on the method of Bradford 26using bovine serum albumin as a standard. Synaptosomes were stored as a pellet on ice and used within 5 h of preparation.

The effects of intravenous anesthetics on CCK8s release were assayed using a multifiltration system (Millipore Multiscreen Assay System; Millipore, Bedford, MA) by minor modifications of the procedure of Walaas. 27All buffers contained 0.5 mg/ml bacitracin to minimize degradation of CCK8s. 28Dilutions from stock solutions of propofol, thiopental, and etomidate in DMSO were made into HBM immediately prior to assays. DMSO as a vehicle at a final concentration of up to 0.5% (v/v) did not affect CCK8s release in control experiments (data not shown). Stock solutions of pentobarbital and ketamine were prepared in water and diluted into HBM.

Synaptosomes (120 μg protein in 60 μl HBM) plus any drugs for preincubation, and HBM (60 μl) containing drugs and/or the indicated secretogogues (15 or 30 mm final KCl; 10 or 20 μm veratridine) were preincubated in wells of separate plates (96-well microtiter plates, CoStar, No. 3797; VWR Scientific Products) with shaking for 5 min at 35°C. Reactions were initiated by transfer of synaptosomal aliquots to HBM aliquots (usually in triplicate or quadruplicate). After 5 min, reactions were stopped by transfer of synaptosomal suspensions to a third 96-well plate containing 15 μl EGTA, 100 mm, in HBM (final concentration approximately 10 mm). Synaptosomes were separated from incubation medium rapidly by vacuum filtration of samples through Millipore multiscreen filtration 96-well plates.

The effects of volatile anesthetics on CCK8s release were assayed using a centrifugation assay that allowed incubations in closed containers with minimal head space to reduce anesthetic losses. This assay was also used in control experiments to determine the effects of Ca²⁺channel blockers on CCK8s release. Volatile anesthetics were prepared as saturated solutions in HBM (10–12 mm) at room temperature. Required volumes were diluted into 0.4-ml polypropylene microcentrifuge tubes containing synaptosomes. Synaptosomes (100–150μg in 330 μl HBM) were preincubated for 5 min at 35°C with or without volatile anesthetics in capped tubes. Ca²⁺channel blockers prepared in HBM were added to synaptosomes in 1.5-ml polypropylene microcentrifuge tubes and preincubated at 35°C for 5 min. Concentrated KCl or veratridine (20 μl) was added through the cap with a Hamilton microsyringe, and tubes were sealed quickly with several layers of Parafilm (American National Can Co., Menasha, WI).

After incubation for 5 min at 35°C with mixing, reactions were stopped with EGTA (final concentration 10 mm). Synaptosomes separated from incubation solution rapidly from low-speed centrifugation for 30 s in CoStar Spin-X centrifuge tube filters (VWR Scientific Products) with an additional GF/B glass fiber filter layer.

CCK8s was quantified by enzyme-linked immunoassay 23,29with minor modifications.

The solutions used were as follows: (1) antigen-coating solution, 50 mm sodium carbonate–bicarbonate buffer, pH 9.6; (2) blocking solution, 20 mm Tris HCl, pH 7.4, 0.1% (w/v) BSA, 0.05% (v/v) Tween 20; (3) washing solution, blocking solution plus 0.15 mm NaCl; (4) substrate solution, 0.1 m phosphate–citrate buffer, pH 5.0, 0.03% (w/v) sodium perborate; and (5) stop solution, 4 N H²SO⁴.

CCK8s-poly-Glu was dissolved in antigen-coating solution to 2 μg/100 ml (calculated for CCK8s content), and 200 μl of this solution was placed in each well of a Nunc-immuno MaxiSorb module (VWR Scientific Products) mounted in a frame. The plate with lid was incubated for 2 h at room temperature. The plate was rinsed twice with 200 μl blocking solution, filled again with 375 μl of this solution and left for at least 1 h at room temperature, and then rinsed thrice with 200 μl washing solution, which was completely decanted on a paper towel. Samples (60–120 μl synaptosome incubation medium) containing CCK8s were added to wells. Standard dilutions of pure CCK8s in duplicate in the same buffer were added to separate wells as standards. Concentrated washing solution was added to each well to achieve final concentrations of 0.15 m NaCl, 0.1% BSA, and 0.05% (v/v) Tween 20 as in the washing solution. An appropriate dilution of anti-CCK8s antibody in washing solution was added, and the covered plate was incubated overnight (12–16 h) at 4°C. Wells were washed five times over 30 min with washing solution, following which 200 μl HRP-conjugated goat antirabbit IgG (1:1000 dilution in washing solution) was added to each well. Covered plates were incubated for 2 h at 35°C. Wells were rinsed five times with washing solution, two times with 20 mm Tris HCl, pH 7.4, and two times with substrate solution. The HRP reaction was initiated by addition of 200 μl o -phenylenediamine (final concentration 0.4 mg/ml) in substrate solution to each well. Plates were incubated at room temperature for 30–45 min in the dark, and reactions were stopped by addition of 50 μl H²SO⁴, 4 N. Absorbance at 492 nm was measured using a Bio-Rad Microplate Reader (model 3550). Release of CCK8s is expressed as fmol CCKs/mg synaptosomal protein for 10 min, which includes 5 min of preincubation and 5 min of stimulation. In figures 1–3, CCK8s release was normalized to the mean basal release of CCK8s in HBM in the absence of added CaCl²(400–600 fmol/mg synaptosomal protein).

Free propofol concentrations in assay mixtures were determined by high-performance liquid chromatography 30; the concentrations of propofol refer to the measured free concentrations. Initial and final volatile anesthetic concentrations in assay mixtures were determined by gas chromatography following extraction into n-heptane. 31The extract (5 μl) was injected into a gas chromatograph (GC-8A; Shimadzu Corp., Kyoto, Japan) equipped with a thermal conductivity detector. Separation was achieved on a 1.8 m/6 mm ID glass column packed with Porapack Q (Supelco, Bellefonte, PA). The column temperature was 210°C, the injector temperature was 230°C, and carrier gas (He) flow was 40 ml/min. Volatile anesthetic concentrations refer to the average measured concentrations; there were no significant differences between the initial and final concentrations.

Values are expressed as mean ± SD. Statistical significance was assessed by two-tailed Student t  test or analysis of variance (ANOVA) with the Newman-Keuls multiple range test or Dunnett multiple comparison test using GraphPad Prism, version 2.01 (GraphPad Software, Inc., San Diego, CA);P < 0.05 was considered statistically significant.

Content of CCK8s in synaptosomes prepared from rat cerebral cortex was 6.7 ± 1.9 pmol/mg synaptosomal protein (n = 12). Basal (spontaneous) release in the absence of CaCl²was 400–600 fmol/mg synaptosomal protein over 10 min. Basal release was not increased significantly by exposure to 1.3 mm CaCl²(fig. 1A). The efflux of CCK8s evoked by elevated KCl or veratridine was concentration dependent and Ca²⁺dependent. Release was increased threefold to fourfold by 30 mm KCl or by 20 μm veratridine. In control experiments, known P/Q-type Ca²⁺channel antagonists inhibited KCl-evoked release, while an N-type Ca²⁺channel antagonist was ineffective (fig. 1B). The specific Na⁺channel antagonist tetrodotoxin completely inhibited veratridine-evoked release (see below).

Propofol (12.5–50 μm), pentobarbital (50–100 μm), thiopental (20 μm), etomidate (20 μm), or ketamine (20 μm) did not affect basal (fig. 2), 30 mm KCl-evoked (fig. 3), or 20 μm veratridine-evoked (fig. 4) release of CCK8s. Similar results were obtained whether synaptosomes were preincubated with the anesthetic for 5 min before addition of secretogogue, or anesthetic and secretogogue were added simultaneously (data not shown). Potentiation or inhibition of release was not observed when submaximal stimuli (15 mm KCl or 10 μm veratridine) were used (data not shown).

Isoflurane or halothane at 0.6–0.8 mm (approximately 2 times minimum alveolar concentration [MAC]) did not affect basal or stimulus-evoked release of CCK-8s (figs. 5 and 6).

Neither basal nor evoked release of the excitatory neuropeptide CCK was affected by clinically relevant concentrations of a number of intravenous and volatile general anesthetics. We employed a specific and sensitive enzyme-linked immunoassay to quantify CCK release from isolated rat cortical nerve terminals (synaptosomes). Synaptosomes are pinched-off nerve terminals prepared by gentle homogenization and subcellular fractionation of nervous tissue. 32They represent the most accessible experimental system for the study of the biochemical and functional properties of presynaptic terminals. Our observations suggest that neither the ion channels that couple depolarization of CCK-containing nerve terminals to Ca²⁺influx nor the molecular machinery that couples this increase in Ca²⁺to vesicle exocytosis is sensitive to concentrations of general anesthetics, which significantly affect the release of other transmitters under comparable conditions.

In agreement with previous studies, elevated KCl or veratridine evoked CCK release in a Ca²⁺-dependent manner. Previous studies have analyzed CCK release from nerve terminals isolated from guinea pig 19and rat hippocampus 20and rat cerebral cortex. 33,34Verhage et al.  20reported basal release of approximately 150 fmol CCK/mg protein/3-min assay (50 fmol/mg protein/min), compared to our value of approximately 500 fmol CCK/mg protein/10-min assay (50 fmol/mg protein/min), which was increased approximately 2.3-fold by 30 mm KCl, compared to a stimulation of threefold to fourfold by 30 mm KCl in our experiments. Synaptosomes are too small for electrical field stimulation; transmitter release must be stimulated pharmacologically. We used two secretogogues that activate secretion by distinct mechanisms. Step elevations of extracellular KCl induce a “clamped” depolarization of the synaptosomal plasma membrane above the threshold for activation of voltage-gated Ca²⁺channels present in the presynaptic terminal. KCl-evoked release is sensitive to P/Q-type Ca²⁺channel blockade (Leenders et al.  34and fig. 1A) but resistant to Na⁺channel blockade (e.g. , with tetrodotoxin). Na⁺channel activation by veratridine, a neurotoxin that inhibits channel inactivation, leads to prolonged synaptosome depolarization and influx of Ca²⁺. Since Na⁺channel activation is required, release is tetrodotoxin-sensitive (figs. 4 and 6). 32CCK release activated either by elevated KCl or by veratridine was not affected by general anesthetics, which suggests that the Na⁺and Ca²⁺channel subtypes coupled to CCK release are insensitive. The interpretation of these studies must consider that artificial secretogogues are required to evoke transmitter release from isolated nerve terminals. Chemical stimulation may not accurately reproduce action potential-evoked release, which could result in underestimation of the sensitivity of neuropeptide release to anesthetics. Previous studies 34and our control studies indicate that CCK release from rat cortical nerve terminals evoked with 30 mm KCl is inhibited by the P/Q-type Ca²⁺channel antagonists ω-agatoxin IVA and ω-conotoxin MVIIC, and we showed that veratridine-evoked release is sensitive to the Na⁺channel antagonist tetrodotoxin. Together, these findings indicate that this assay is sensitive to blockade of presynaptic Ca²⁺and/or Na⁺channels. Nevertheless, it is possible that the stimuli used to evoke CCK from isolated nerve terminals do not mimic action potential–evoked release and thus do not accurately reflect their sensitivity to anesthetics. Our findings do not rule out anesthetic actions on synaptic transmission by CCK mediated by mechanisms other than direct effects on release or on transmission by peptides other than CCK, such as those demonstrated in mollusks. 35 

The fundamental mechanisms underlying transmitter release are largely conserved among the various classes of neurotransmitters. 7–9Vesicles are released by Ca²⁺-dependent fusion with the plasma membrane upon Ca²⁺influx utilizing the highly conserved SNARE (soluble N-ethylmaleimide-sensitive fusion protein [NSF]-attachment protein [SNAP] receptor) core complex proteins syntaxin, synaptobrevin, and SNAP-25, as well as the Ca²⁺-binding protein synaptotagmin. Genetic evidence for anesthetic effects on SNARE proteins was provided by a recent screening of existing mutants of the nematode C. elegans  for alterations in sensitivity to volatile anesthetics, which identified mutations in all three components of the core fusion complex. 36These findings imply that inhibition of transmitter release by general anesthetics, which is observed for many agents at clinical concentrations, 3should be observed for all transmitter classes, if these mechanisms are shared by all transmitter classes. The results reported here indicate that this generalization does not hold for the excitatory neuropeptide cholecystokinin. Thus, concentrations of both volatile and intravenous anesthetics which inhibit depolarization-evoked release of glutamate (e.g. , isoflurane IC⁵⁰= 0.50 mm for inhibition of veratridine-evoked release in rat cortex), 25,37–39GABA, 4and acetylcholine 40from small synaptic vesicles had no significant effect on CCK release from large dense core vesicles. This difference suggests that the mechanism involved in the presynaptic actions of general anesthetics is not a common property of all transmitter classes. Since classic transmitters are often colocalized with neuropeptides in the same nerve terminals (e.g. , GABA and CCK in cerebral cortex), 7it is possible that general anesthetics selectively affect the release of the transmitters contained within small synaptic vesicles without affecting release of neuropeptides contained within dense core vesicles from the same terminals. Whether release of other neuropeptides is similarly resistant to general anesthetics will require further study.

Neurotransmitter release involves many steps, from action potential invasion of the presynaptic terminal; depolarization; ion channel gating; Ca²⁺influx; vesicle translocation, docking and priming; Ca²⁺-release coupling; vesicle fusion/exocytosis; to vesicle endocytosis. Each of these processes is modulated by repetitive activity, trophic factors, and presynaptic receptors. Certain differences in the molecular and cellular mechanisms that underlie the exocytotic release of fast transmitters (amino acids, monoamines) and slow transmitters (neuropeptides) have been described which could result in their differential sensitivities to general anesthetics. Neurons possess two types of secretory vesicles that undergo Ca²⁺-dependent exocytosis with distinct kinetic differences despite their overall similar biochemical compositions. 7Classic transmitters are released by fast exocytosis from small synaptic vesicles which cluster at active zones, whereas neuropeptides are released by fusion of large dense core vesicles with slower kinetics at sites away from the active zone. 9,41Fast transmitters are released at active zones by relatively high local concentrations of Ca²⁺(Ca²⁺microdomains) produced by close physical coupling to specific Ca²⁺channel subtypes, while neuropeptide release occurs by a more delocalized bulk phase increase in nerve terminal Ca²⁺as a result of repetitive, high-frequency stimulation. 19,42The apparent Ca²⁺sensitivity of amino acid transmitter release from small synaptic vesicles is lower than that of neuropeptide release from large dense core vesicles 19(but see Bollmann et al.  43), which suggests the involvement of distinct Ca²⁺sensors. Release of different classes of transmitters may also be coupled to distinct Ca²⁺channel subtypes; release from small synaptic vesicles was coupled to N- and P-type Ca²⁺channels, whereas CCK release from large dense core vesicles was coupled primarily to P/Q-type Ca²⁺channels. 12,34Neuropeptides are packaged into large dense core vesicles in the trans -Golgi network, whereas small synaptic vesicles are loaded by specific transporters in the terminal. Although exocytosis of both small and dense core vesicles involves the highly conserved SNARE protein machinery for fusion/exocytosis, 9,44,45these different properties are consistent with subtle differences in their regulatory components, such as differential expression of synaptotagmin isoforms 46and CAPS, 47which is expressed on large dense core, but not small synaptic, vesicles. One or more of these differences may be crucial in conferring sensitivity to general anesthetics of release of classic transmitters such as glutamate from small synaptic vesicles and insensitivity to anesthetics of the release of CCK, and possibly other transmitters, from large dense core vesicles.