Effects of Remifentanil on N -methyl-d-aspartate Receptor: An Electrophysiologic Study in Rat Spinal Cord

Type A porcine gelatin, glycine, HEPES, EGTA, glutamate and d-gluconic acid lactone were obtained from Sigma (Bornem, Belgium); NMDA, bicuculline, and d-2-amino-5-phosphonovalerate (D-AP5) were obtained from Tocris (Bristol, United Kingdom); strychnine was obtained from RBI (Bornem, Belgium); and tetrodotoxin was obtained from Alexis (Zandhoven, Belgium). Ultiva®, 1 mg, was obtained from GlaxoSmithKline (Genval, Belgium); naloxone was obtained from Bristol-Myers Squibb Pharma (Braine-l’Alleud, Belgium); and [³H]-d,l(E)-2-amino-4-propyl-5-phosphono-3-pentanoic acid (CGP39653), 44.0 Ci/mmol, was obtained from NEN (Zaventem, Belgium).

Whole cell patch clamp recordings were conducted on 300-μm-thick lumbar spinal cord slices obtained from 10-day-old Wistar rats (Iffa Credo, Reims, Paris, France).

After being anesthetized with halothane, rats were killed by cervical dislocation, and a lumbar laminectomy was performed. The lumbar enlargement was removed, cleared of any remaining dura membrane, embedded in a 10% gelatin block, and glued to the stage of a vibrating microslicer (Leica, Wetzlar, Germany) with cyanoacrylate glue (Roth, Karlsruhe, Germany). The spinal tissue was immersed in ice-cold Krebs solution and saturated with 95% oxygen and 5% carbon dioxide throughout all of these procedures. Slices were maintained in an incubation chamber at room temperature for at least 1 h before recordings. Krebs solution used for this procedure had the following composition: 206 mm sucrose, 25 mm NaHCO³, 25 mm D-(+) glucose, 2.5 mm KCl, 1.25 mm NaH²PO⁴, 1 mm CaCl², and 2 mm MgCl².33 

All efforts to avoid animal suffering were made according to the standards of the Institutional Ethical Committee of the School of Medicine of the Université Libre de Bruxelles (Brussels, Belgium).

Cells currents were recorded in the whole cell configuration of the patch clamp technique34at room temperature inside lamina II of the dorsal horn.

Cells from lamina II (substantia gelatinosa) were identified using a Zeiss (Oberkochen, Germany) Axioskop II FS MOT microscope. Because of its relatively high neuronal density, lamina II could be readily visualized at low magnification (×100). Lamina II could also be distinguished from lamina I characterized by a striated appearance.35At higher magnification (×400), morphologically distinct stalk and islet cells served as markers for the boundaries of lamina II, outer and inner, respectively.36Thus, the substantia gelatinosa contains three morphologic types of cells that can be visually identified: projection neurons on which recordings were performed and stalk and islet interneurons.

During the patch clamp recordings, slices were perfused with standard Krebs-Henseleit solution containing 125 mm NaCl, 25 mm NaHCO³, 25 mm D-(+) glucose, 2.5 mm KCl, 1.25 mm NaH²PO⁴, 1 mm CaCl², 10 μm bicuculline, 0.5 μm strychnine, and 0.1 μm tetrodotoxin, with a pH of 7.4 (95% oxygen and 5% carbon dioxide). Bicuculline is used as a blocker of γ-aminobutyric acid type A receptor, strychnine is used as a blocker of glycine receptors, and tetrodotoxin is used as a blocker of sodium channel.

The following components were diluted into the former solution: 10 μm NMDA; 10 or 3,000 μm glycine; Ultiva® containing 50 μm remifentanil hydrochloride and 3,000 μm glycine; 50 or 3 μm remifentanil hydrochloride, 50 μm D-AP5, the specific NMDA glutamate site antagonist; and 20 μm naloxone, the μ-opioid receptor antagonist.

Patch pipettes were pulled from borosilicate glass capillaries (Harvard Apparatus, Edenbridge, United Kingdom) and had a resistance of 6–10 MΩ. The intracellular solution contained 130 mm Cs⁺gluconate, 6.8 mm CsCl, 0.1 mm EGTA, 10 mm HEPES, and 2 mm Mg²ATP, with pH adjusted to 7.3 with NaOH and osmolarity adjusted to 305–310 mOsm with sucrose.

Whole cell currents were recorded with an EPC-8 amplifier (Heka Elektronik, Lambrecht/Pfalz, Germany) at a membrane potential of −60 mV.

Data were obtained using the acquisition software Pulse (Heka) and were analyzed with IGOR software (WaveMetrics, Lake Oswego, OR). Passive cellular parameters were extracted in voltage clamp by analyzing current relaxation induced by a 10-mV step change from a holding potential of −60 mV as described previously.37Visually identified neurons were selected based on a low and stable leak current. Access resistance was monitored to ensure that recording conditions remained stable during the whole experiment. Neurons showing more than 15% modification of this access resistance were excluded.

Remifentanil conditioned without glycine is not available; purification of Ultiva® is difficult and exposes the component to degradations and contaminations. Therefore, we proceeded to the synthesis of remifentanil. Synthesis has been conducted following two general published procedures. The first intermediate, i.e. , 4-[(1-oxopropyl) phenylamine]-1-benzylpiperidine-4-carboxylic acid methyl ester, has been obtained in five steps from N -benzylpiperidinone.38After N -debenzylation,39the piperidine derivative has been reacted with methyl acrylate.40Purification by column chromatography afforded remifentanil base, which was further transformed into hydrochloride salt by addition of hydrochloric acid solution in diethylether. The solid was filtered and crystallized into isopropanol–diethylether.

The resulting component was controlled with nuclear magnetic resonance spectra recorded on a Bruker Avance 300-MHz spectrometer (Illkirch, France) with tetramethylsilane as internal standard, at 293 K. Infrared analysis was performed with a Shimadzu IR-470 spectrophotometer (Antwerp, Belgium). In dimethyl sulfoxide (DMSO)-d  ⁶, a mixture of conformers has been observed (equilibrium between a minor isomer and a major isomer, ratio 15:85.

¹H-nuclear magnetic resonance (DMSO-d  ⁶): 11.36*– 11.02** (bs, 1H, NH⁺), 7.68*–7.52 (m, 3H, aromatics), 7.34 (m, 2H, aromatics), 3.72 (s, 3H, OCH³), 3.62 (s, 3H, OCH³), 3.40 (m, 2H, MeOOCCH²CH²), 3.25– 3.07 (m, 4H, H2–H6), 2.85 (t, 2H, J = 7.8 Hz, CH²COO), 2.26–1.97 (m, 4H, H3–H5), 1.75 (q, 2H, J = 7.4 Hz, CH²CH³), 0.8 (q, 3H, J = 7.3 Hz, CH²CH³) ppm.

¹³C-nuclear magnetic resonance (DMSO-d  ⁶): 174*–172.9** (piperidine COO), 173.5*– 172.5** (CH²COO), 170 (CON), 138.6*–137.8** (C1′), 130.5*–130** (C2′C6′), 129.4*–129.2** (C3′C5′), 128.7 (C4′), 59.9*–59.3** (C4), 52.3*–52** (piperidine COOCH³), 51.6 (CH²COOCH³), 50.5**–49.8* (C2C6), 48.5**–46.7* (NCH²CH²), 29.7 (CH²CON), 28 (NCH²CH²), 27.9**–27.7* (C3C5), 8.7 (CH³CH²) ppm.

IR (KBr disk): ν= 2920 (C-H), 2465 (N-H), 1724 (C = O ester), 1645 (C = O amide), 1438, 1367, 1247, 1206, 699 (aromatic) cm⁻¹. * Minor isomer; ** major isomer; no asterisk, both isomers.

Radioligand bindings were performed on 20-μm frozen spinal cord sections obtained from adult Wistar rats and mounted on gelatin-coated slides with increasing concentrations of remifentanil hydrochloride (0.1–100 μm). The sections were preincubated three times for 15 min each at room temperature in 50 mm Tris-HCl buffer, with a pH of 7.5, and incubated for 60 min at 4°C in buffer containing 10 nm [³H]CGP39653. Nonspecific binding of the ³H-labeled ligand was assessed by the addition of 100 μm l-glutamate. Slides were washed four times for 15 s in ice-cold buffer, dipped in ice-cold distilled water, and dried under a stream of cold air.

After completion of the procedures, slides were dried under a stream of cold air, exposed to tritium-sensitive screens, and revealed by a Cyclone storage phosphore system (PerkinElmer, Boston, MA). Analyses were performed using the public domain NIH image 1.61 program (National Institutes of Health, Bethesda, MD).

Data are reported as mean ± SEM, and statistical comparisons were performed using paired or unpaired Student t  tests for electrophysiologic results and analysis of variance for binding assay results (P > 0.05, not significant).

Location of lamina II is facilitated, at low magnification (×100), by the entrance of the dorsal root and the macroscopic aspect of the spinal slices. At higher magnification (×400), morphologically distinct stalk and islet cells serves as markers for the location of the cells that are recorded.36Membrane capacitances (23.52 ± 1.96 GΩ) and resistances (2.07 ± 0.27 GΩ) of the 22 selected cells exhibit little cell-to-cell variation. This result underlines the ability of the method to select a homogenous cells population.

Experiments were designed to determine whether remifentanil hydrochloride is able to activate NMDA receptors directly. Ultiva® contains remifentanil hydrochloride and glycine as vehicle, and NMDA receptors require both glutamate and its coagonist, glycine, to be activated. These specific characteristics mean that both components of Ultiva® can be involved in NMDA receptor activation. Therefore, we planned to perform experiments first with remifentanil hydrochloride, second with glycine, and third with the two components within Ultiva®.

Experiments were performed with 50 μm remifentanil hydrochloride, considering the saturating effect obtained with Ultiva® in a previous study.24Given that a 1-mg Ultiva® vial contains 15 mg glycine as adjunct, a solution of Ultiva® containing 50 μm remifentanil also contains 3 mm glycine. Therefore, we performed a first group of experiments with 50 μm remifentanil hydrochloride, a second with 3 mm glycine, and a third with Ultiva® containing 50 μm remifentanil and 3 mm glycine.

Whole cell current is first recorded in control conditions (standard Krebs-Henseleit solution). When the current is stable during 100 s, 50 μm remifentanil hydrochloride and 10 μm glycine are added to the solution perfusing bath, and the current is recorded during 400 s. Remifentanil hydrochloride does not induce any current variation (n = 5); the current recorded during the perfusion of remifentanil hydrochloride is similar to the one recorded during the control conditions (−14.69 ± 3.04 vs. −12.09 ± 2.61 pA; P > 0.05, n = 5; figs. 1A and B).

In accord with these results, the binding study conducted on spinal cord sections from adult rats shows that increasing concentrations (0.1–100 μm) of remifentanil hydrochloride do not modify the binding of [³H]CGP39653 from the glutamate site (fig. 2). These results indicate that remifentanil hydrochloride seems to be devoid of NMDA receptor activator properties.

Whole cell current is recorded using a similar protocol. When the current is stable, 3 mm glycine is added to the perfusing bath. Glycine induces an inward current (−94.53 ± 16.29 pA). When this current remains stable, 50 μm of the specific NMDA glutamate site antagonist D-AP5 is added to the bath. The glycine-induced current is abolished by application of D-AP5, revealing an NMDA current. There is a significant difference between the current induced by glycine and the one recorded during the control condition (−94.53 ± 16.29 vs. −20.90 ± 5.56 pA; P < 0.05, n = 5). No significant difference between the current induced by glycine plus D-AP5 and the current recorded during control condition is observed (−45.74 ± 10.54 vs. −20.90 ± 5.56 pA; P > 0.05, n = 5; figs. 3A and B). These results show that 3 mm glycine alone is able to activate NMDA receptors in rat spinal cord. Ultiva® contains remifentanil and glycine. When the dilution of Ultiva® provides a 50 μm remifentanil hydrochloride concentration, the concentration of glycine is 3 mm. This concentration of remifentanil does not induce an NMDA current, whereas 3 mm glycine does induce this current.

Using a similar protocol, application of Ultiva® containing 50 μm remifentanil hydrochloride and 3 mm glycine induces an inward current (−96.42 ± 10.99 pA) that is abolished by application of 50 μm D-AP5, revealing an NMDA current.

There is a significant difference between the current induced by Ultiva® and the one recorded during the control condition (−96.42 ± 10.99 vs. −23.92 ± 3.16 pA; P < 0.05, n = 7). No significant difference between the current induced by Ultiva® plus D-AP5 and the current recorded during the control condition is observed (−39.43 ± 5.47 vs. −23.92 ± 3.16 pA; P > 0.05, n = 7; figs. 4A and B).

There is also no significant difference between the NMDA current induced by Ultiva® containing 50 μm remifentanil hydrochloride and glycine and the NMDA current induced by 3 mm glycine (−96.42 ± 10.99 pA, n = 7 vs. −94.53 ± 16.29 pA, n = 5; P > 0.05).

These results indicate that the NMDA current induced by Ultiva® preparation is related to the presence of glycine and that remifentanil hydrochloride by itself seems to be devoid of NMDA receptor activation properties.

Modulation of NMDA receptors activity by remifentanil hydrochloride through the activation of μ-opioid receptors, as described for other opioids,27,28was our second hypothesis. Therefore, we applied remifentanil hydrochloride on cells with preactivated NMDA receptors. The opioid agonist DAMGO has been described to increase NMDA receptors activity when applied within the micromolar range.28In accord with this observation, we started our experiments with 3 μm remifentanil hydrochloride.

N -methyl-d-aspartate receptors are first activated by 10 μm NMDA and 10 μm glycine. When this current remains stable, 3 μm remifentanil hydrochloride, added to the bath, increases the NMDA-induced inward current by 61.3 ± 13% (−158.76 ± 22.66 vs. −247.31 ± 40.45 pA; P < 0.05, n = 5). This increase is abolished by 20 μm of the μ-opioid receptor antagonist naloxone (−158.76 ± 22.66 vs. −126.63 ± 18.81 pA; P > 0.05, n = 5; figs. 5A and B).

These results indicate that remifentanil hydrochloride potentiates the NMDA current through a μ-opioid receptor pathway.

The current study provides strong evidence that remifentanil hydrochloride does not directly activate NMDA receptors in rat dorsal horn neurons. The NMDA current recorded after application of Ultiva® is related to the drug vehicle glycine. In these neurons, the NMDA current is majored by application of remifentanil hydrochloride in the micromolar range, and this potentiation is abolished by the μ-opioid receptor antagonist naloxone. Therefore, remifentanil hydrochloride increases NMDA current, and this potentiation is mediated by μ-opioid receptor activation, probably through an intracellular pathway.

Application of remifentanil hydrochloride on rat dorsal horn neurons does not induce any current. The binding assay technique does not provide any evidence that remifentanil hydrochloride is able to bind to the glutamate sites of the NMDA receptor. Considering these latter results together with the patch clamp recordings, it is unlikely that remifentanil hydrochloride directly stimulates the NMDA receptor in rat spinal cord.

N -methyl-d-aspartate receptor is also activated in the presence of Ultiva® containing remifentanil hydrochloride and glycine. Considering the results of the experiments performed first with remifentanil hydrochloride and second with glycine, it seems that this current is related to the presence of glycine. NMDA receptors require both glutamate and its coagonist, glycine, to be activated, and the potentiation of NMDA receptor by glycine is specific for this receptor.41Previous studies in the dorsal horn of spinal cord showed that, in the presence of tetrodotoxin, bicuculline, and strychnine, nerves terminals spontaneously release quanta of glutamate. These small amounts of glutamate do not induce NMDA currents.42Nevertheless, when the concentration of glutamate remains unchanged, increasing concentrations of glycine induce increasing NMDA currents.43Therefore, a possible explanation would be that the small amount of endogenous glutamate present inside the spinal cord slices is sufficient to activate NMDA receptors with a 3 mm saturating concentration of glycine.

Moreover, the Ultiva®-induced current is abolished by the specific glutamate site antagonist D-AP5, suggesting that remifentanil hydrochloride contained in Ultiva® cannot activate NMDA receptors through another site or mechanism; otherwise, the induced current would not be totally reversed. This result confirmed the first experiment performed with remifentanil hydrochloride alone.

We also show that application of remifentanil hydrochloride on dorsal horn neurons in the micromolar range potentiates the NMDA-induced current. These results are in accord with recent reports showing similar results with the μ-opioid agonist DAMGO.27,28This potentiation is reversed by the μ-opioid antagonist naloxone, indicating the role of μ-opioid receptor as a trigger of an intracellular pathway, able to positively regulate the activity of NMDA receptors. Therefore, during application of Ultiva® on dorsal horn neurons, glycine appears as a primer of NMDA receptor activation, whereas remifentanil by itself enhances this activation.

Glycine was recently described as an important NMDA receptor facilitator.44During acute pain, spillover of glycine from spinal cord interneurons causes NMDA receptor facilitation and contributes to the development of chronic pain. Intravenous glycine injection dose-dependently increases the concentration of glycine contained inside the cerebrospinal fluid,45and extracellular concentrations are below the levels required for maximal NMDA receptor activation.46In contrast, behavioral studies with rats showed that the infusion of glycine alone did not modify the pain threshold while it induced a transitory motor impairment and that infusion of remifentanil hydrochloride with its vehicle glycine did not modify the pain threshold compared with infusion of remifentanil hydrochloride.47,48It should not also be forgotten that glycine receptors are the major inhibitory receptors in spinal cord.49 

Glycine contained in Ultiva® has not been incriminated as an NMDA receptor activator in a recent study conducted by Hahnenkamp et al.  24on NMDA receptors expressed in Xenopus  oocytes. Despite the absence of experiments using remifentanil hydrochloride out of the presence of glycine, these results suggest a direct allosteric activation of NMDA receptor by remifentanil hydrochloride. Differences with our results could be explained by the differences between the two models. Considering that the functional properties of NMDA receptors are modulated by the composition of the subunits,50relating data obtained with recombinant NMDA receptors to those obtained with native NMDA receptors is challenging. Moreover, the composition of the multiproteic complexes known to anchor and regulate NMDA receptors at the postsynaptic membrane plays also a critical role.51We must note here that the current study was conducted under physiologic conditions. Currents were recorded from neurons inside their environment. The opportunity to perform experiments with a pure and controlled preparation of remifentanil hydrochloride in the absence of glycine must also be noted.

In conclusion, remifentanil hydrochloride does not directly stimulate NMDA receptors in spinal cord. Ultiva® activates NMDA receptors, but this activation is related to the presence of glycine. Application of remifentanil hydrochloride increases the activity of NMDA receptors, and this potentiation is mediated through a μ-opioid receptor–initiated cascade.