Spinal N-Methyl-d-Aspartate Receptors May Contribute to the Immobilizing Action of Isoflurane

Abstract

We examined whether N-methyl-d-aspartate (NMDA) receptors influence the immobilizing effect of isoflurane by a spinal or supraspinal action. We antagonized NMDA receptors by intrathecal (IT), intracerebroventricular (ICV), and IV administration of MK 801 (a noncompetitive NMDA antagonist) and measured the decrease in isoflurane minimum alveolar anesthetic concentration (MAC). We also measured MK 801 tissue concentrations in homogenates of upper and lower spinal cord, a slice of cerebral cortex, and the whole brain. IT infusion of MK 801 decreased isoflurane MAC more potently than ICV or IV infusions. The change in MAC correlated with the MK 801 concentration in the lower part of the spinal cord (P < 0.01) but not with concentrations in supraspinal tissue. The maximal effect of IT MK 801 reached a plateau without achieving anesthesia. IV doses 270-fold larger than the largest IT dose also did not produce anesthesia in the absence of isoflurane. These results suggest that the capacity of MK 801 to decrease the MAC of isoflurane results from an effect on the spinal cord but that spinal NMDA receptors provide only partial mediation of the immobility produced by isoflurane. Because neither IT nor IV MK 801 provide complete anesthesia, these findings also call into question the notion that NMDA blockade alone suffices to produce anesthesia as defined by immobility in the face of noxious stimulation.

Ligand-gated ion channels are likely targets of volatile anesthetic action. Enhancement of inhibitory neurotransmitter function or inhibition of excitatory neurotransmitter function or both are plausible mechanisms of anesthesia. Ionotropic glutamate receptors mediate excitotoxic neurotransmission ⁽¹⁾. The present investigation focuses on the role of glutamate N-methyl-d-aspartate (NMDA) receptors as mediators of immobility, as defined by the minimum alveolar anesthetic concentration (MAC; in this case of the prototypic volatile anesthetic isoflurane), preventing movement in 50% of subjects in response to a noxious stimulus.

Inhaled anesthetics, including isoflurane, ethanol, nitrous oxide, and xenon, inhibit NMDA receptors in receptor expression systems ⁽²⁾. Clinically relevant (approximately 1 MAC) concentrations of isoflurane depress field excitatory postsynaptic potentials mediated by NMDA receptors ⁽³⁾. Both enflurane ⁽⁴⁾ and ethanol ⁽⁵⁾ can decrease NMDA glutamate currents in motor neurons in the spinal cord, an effect independent of effects on γ-aminobutyric acid type A and glycine receptors. These studies suggest that NMDA receptors may mediate the actions of at least some inhaled anesthetics.

If NMDA receptors mediate the immobilizing effect of isoflurane, then the administration of NMDA receptor antagonists should decrease the MAC of isoflurane. Consistent with this prediction, Ishizaki et al. ^(6,7) found a maximal decrease of 30% in isoflurane MAC in rats given intrathecal (IT) bolus doses of the NMDA antagonists AP5, MK 801, CPP (a competitive antagonist at the NMDA receptor), and 7CKA (a selective antagonist at the glycine site of the NMDA receptor). McFarlane et al. ⁽⁸⁾ reported an 80% decrease in MAC of halothane from IV application of the competitive NMDA receptor antagonist CGS 19755. If isoflurane and halothane similarly depress NMDA receptors, these results suggest that either the IT infusion of NMDA antagonists produced a submaximal MAC-sparing effect or that the IV infusion of an NMDA blocker added an immobilizing action at a supraspinal target. A supraspinal effect is suggested by the results reported recently by Masaki et al. ⁽⁹⁾. They applied the NMDA antagonist D(-)-2-amino-5-phosphonopentanoic acid (D-AP5) intracerebroventricularly (ICV) and noted a sevoflurane MAC-sparing effect.

The spinal cord mediates most of the capacity of volatile anesthetics to produce immobility in the face of noxious stimulation. Cervical transection of the spinal cord of rats does not change MAC ⁽¹⁰⁾, and separate perfusion of the brain and spinal cord of goats reveals that the cord is more sensitive to anesthetic than the brain (nearly three times the cerebral concentration of isoflurane is required to prevent pain evoked movement when the brain alone is perfused with anesthetic compared with perfusion of both the brain and spinal cord) ⁽¹¹⁾.

In this study, we assessed the relative contributions of spinal versus supraspinal NMDA receptor blockade on the MAC for isoflurane. To accomplish this, we examined the maximal effects of IT, ICV, and IV infusion of ascending doses of the NMDA antagonist MK 801 on the MAC of isoflurane. Our analysis used a regression model to correlate the MAC-sparing effect with the concentration of MK 801 at four sites in the central nervous system.

Methods

With approval of the Committee on Animal Research of the University of California, San Francisco, we studied 80 male (Crl:CD(SD)BR) rats weighing 206–440 g obtained from Charles River Laboratories (Wilmington, MA).

To place IT catheters, rats were anesthetized with isoflurane, and a 32-gauge polyurethane catheter (Micor Inc, Allison Park, PA) was placed through the atlanto-occipital membrane, as described by Zhang et al. ⁽¹²⁾. The catheter was caudally threaded 6 to 8 cm towards the lumbar sac, with the length depending on the size of the rat. Sutures fixed the catheter to adjacent muscle and skin at the neck. Rats were allowed to recover from anesthesia and surgery for at least 24 h before the study.

To place cannulae into the third ventricle, rats were anesthetized with isoflurane, their skulls were exposed, and a hole was drilled 1.8 mm posterior to the bregma in the midline. Through this hole, we inserted a 24-gauge stainless steel guide cannula to a depth of 4.1 mm from the skull surface. The cannula was secured by wiring to 2 screws that were placed into the skull approximately 5 mm to either side of the cannula. During isoflurane anesthesia, IV catheters made of polyurethane 10 were placed in the right internal jugular vein.

IT, ICV, and IV studies

MAC was determined concurrently in four rats placed in individual clear plastic cylinders. A rectal temperature probe was inserted, and the temperature probe and the tail of the rat were separately drawn through holes in the rubber stopper used to seal one end of the cylinder. Ports through the rubber stoppers in each end of the cylinder allowed gas delivery at the head end of the cylinder and exit of gas at the tail. A total flow rate of 4 L/min of oxygen and isoflurane was delivered (average 1 L/min/cylinder), and the exiting gases were scavenged.

To determine the control MAC (MAC₀) for the upper and lower limbs (i.e., MAC was separately defined by movement of the upper versus lower limbs), isoflurane was introduced into the system using a conventional vaporizer and starting with a partial pressure of approximately 1.0% atm. Anesthetic partial pressures were monitored using an infrared analyzer (Datascope, Helsinki, Finland). After equilibration with isoflurane for 30 min, a tail clamp was applied for up to 1 min or until the rat moved the upper or lower limbs. The isoflurane partial pressure was then measured by gas chromatography. If the rat moved both the upper and lower limbs, the isoflurane partial pressure was increased by 0.1% to 0.2% atm. After equilibration for 30 min, a tail clamp was applied and isoflurane partial pressure again measured. This procedure was repeated until a partial pressure was reached at which each rat did not move the upper and lower limbs. MAC was calculated separately using movement of the upper limbs and the lower limbs.

Each study consisted of three determinations of isoflurane MAC. In the first (MAC₀), we infused artificial cerebrospinal fluid (aCSF) alone at 1 μL/min through the IT catheter or saline alone at 5 μL/min through the IV catheter. The aCSF was made up daily from stock solutions, with the pH value adjusted to approximately 7.2 (measured with a pH meter) by bubbling CO₂ through it. The final composition of the aCSF was 154.7 mM of Na⁺, 0.82 mM of Mg⁺⁺, 2.9 mM of K⁺, 132.49 mM of Cl⁻, 1.1 mM of Ca²⁺, and 5.9 mM of glucose.

In the second part of the IT and ICV studies (MAC₁), the aCSF was replaced with aCSF containing MK 801 and 0.01% methylene blue. For the IV studies, the saline was replaced with saline containing MK 801. The interval between the first and second parts of the study was a half hour, during which time the isoflurane concentration was decreased to the point that each rat again responded to the tail clamp. The procedure used to determine MAC₀ was used to determine MAC₁, defined again as the average of the partial pressures that just prevented and permitted movement in the upper and lower limbs in response to clamping the tail. The change in MAC was calculated as the percent decrease in the MAC₁ for the second part of each study from the first part (MAC₀). That is, 100 × (MAC₀ − MAC₁)/MAC₀. This process was repeated one more time to give MAC₂. The MAC₂ determination was a check to confirm that MAC was being measured at a point in the infusion at which the drug effect was stable, that is, at which MAC₁ and MAC₂ did not differ. We calculated the mean and sd for the change at each dose of MK 801.

After determining MAC₂ for the upper and lower limbs, isoflurane was given to assure immobility, and a 1-mL blood sample was taken from the aorta and transferred to a 10 mL Vacutainer^® filled with sodium heparin (Becton Dickenson, Franklin Lakes, NJ). Subsequently, the rats were decapitated, and a necropsy was performed to determine the location of the methylene blue and therefore the correctness of the position of the ICV cannula or the IT catheter.

The concentrations of MK 801 in plasma, upper spinal cord, lower spinal cord, and whole brain were determined in selected groups, as shown in Table 2. After determining MAC, the brain and spinal cord were removed. The spinal cord was cut in half, and the upper and lower spinal cords were immersed separately two times into saline 0.9% (pH value of 4.56). Excess water was removed with a dry sponge, weighed, and placed into a 10-mL glass tube (Fisher Scientific, Pittsburgh, PA) filled with 5 μg of carbamazepine (Aldrich, St. Louis, MO), as an internal standard, and 3 mL of sodium carbonate 0.5 M (pH value of 10.8). Similarly, a cortical slice distant from the third ventricle was taken, and the slice and remaining brain were immersed two times into saline 0.9% (pH value of 4.56), excess water was removed with a dry sponge, and the tissue was weighed and placed into a 10-mL glass tube filled as described below.

Table 2: Concentrations of MK 801 in the Plasma, Brain, Cortex, and Upper and Lower Spinal Cord After Infusion of MK 801 IV, Intrathecally (IT), and Intracerebroventricularly (ICV)^a

We conducted separate studies to determine whether MK 801 infusion permanently changed MAC (e.g., via neural injury). Ascending concentrations of MK 801 were infused IT. Recovery of normal motor coordination (as determined by visual inspection with movement and stimulation of the rat) was observed for 48 h after the study. Then, recovery MAC (MAC_r) was determined in four rats at a time. MAC_r was measured 2 days after the study for both the upper and lower limbs.

The blood samples were centrifuged (TJ-6, Beckman Dickenson) at 2000 g rpm for 20 min, and 300 μL of the plasma phase was transferred to 15-mL glass tubes (Fisher Scientific) filled with 5 μg of the internal standard and 500 μL of sodium carbonate 0.5 M (pH value of 10.8). The brain and spinal cord samples together with the carbamazepine solution and 3 mL of sodium carbonate 0.5 M (pH value of 10.8) were homogenized (Kinematica Co, Kriens, Lucerne, Switzerland). The homogenate was transferred to 15-mL glass tubes. Eight milliliters of pentane (Fisher Scientific) was added to the tissue and the plasma samples in the glass tubes. All samples were vortexed for 1 min and centrifuged at 2000 g rpm for 20 min. We transferred 5.5 mL of the organic phase to 10-mL glass tubes and dried this phase with a flow of nitrogen. The dried samples were frozen at −80°C until analyzed.

Calibration samples were prepared using various amounts of stock MK 801 (2.5 μg, 5 μg, and 20 μg), a constant amount of carbamazepine (5 μg), acetonitrile (100 μL) (Fisher Scientific), and 300 μL of rat plasma placed into 15-mL glass tubes. The calibration samples were then treated as the samples described above.

A model 1100 high-performance liquid chromatograph (Hewlett-Packard 1100 series, Agilent Technologies Inc, Mountain View, CA), equipped with an autosampler injector and a detector, was used. Analyses were performed on a 5-μm C18-A Polaris column (15 cm × 4.6 mm internal diameter) operating at ambient room temperature (20°C–25°C) (Ansys Technologies, Inc, Lake Forest, CA).

Acetonitrile and 0.05 M of monobasic potassium phosphate (1:1) with a pH value of 4.5 ± 0.05 were used as mobile phase. The flow rate was 0.5 mL/min. The frozen plasma and tissue samples were dissolved in 100 μL of mobile phase, and a 20-μL aliquot was injected. Quantitation was performed using the peak area ratio method with carbamazepine as the internal standard. Calibration graphs were obtained from unweighted least-squares linear regression analysis of the data after injection of two separately prepared samples for each of the four calibration levels. The detector wavelength for the internal standard carbamazepine and for MK 801 detection was 265 nm.

We used a Gow-Mac 580 flame ionization detector gas chromatograph (Gow-Mac Instrument Corp, Bridgewater, NJ) to analyze isoflurane concentrations. The 4.6-m-long, 0.22-cm (infective dose) column was packed with SF-96. The column temperature was 152°C with the detector maintained at a temperature approximately 30°C greater. The carrier gas flow was nitrogen at a flow of 16 mL/min. The detector received 35 mL/min of hydrogen and 320 mL/min of air. Primary standards were prepared for isoflurane and the linearity of the response of the chromatograph determined. We often used secondary (cylinder) standards referenced to the primary standards.

Mean and sd values were calculated, and Student’s t test was performed as appropriate. A ceiling effect was deemed to exist if two successively larger doses of MK 801 (producing the greatest change in MAC) did not differ significantly from each other.

A regression model was used to determine the neuroanatomic site(s) of NMDA receptor inhibition that mediate isoflurane-induced immobility. The absolute value of the change in MAC (MACΔ) was fit to the lower cord (lcord), upper cord (ucord), brain (brain), and cortex (cortex) MK 801 concentrations via the equation:MATH

This equation describes a sigmoid curve, where MACΔ asymptotically approaches a maximum value of A and a minimum value of B. a₀ through a₄ are the coefficients to be determined by the regression. To implement this model, the data were transformed, and multiple linear regression to the following equation was performed:MATH

The variables that had an effect significantly different from zero were included in the final regression model. Regression analysis was performed using SPSS 10.0 (SPSS Inc, Chicago, IL). P < 0.05 was taken as statistically significant.

Results

Infusion of ascending concentrations of MK 801 into the subarachnoidal space decreased isoflurane MAC₂ (Table 1). An infusion of 10 μg · kg⁻¹ · min⁻¹ of MK 801 produced a decrease of approximately 65% (Fig. 1) without nerve injury. MAC defined by movement of upper limbs versus lower limbs did not differ significantly. Infusion of 20 μg · kg⁻¹ · min⁻¹ of MK 801 decreased isoflurane MAC 73% ± 17% in the upper limbs and 87% ± 19% in the lower limbs (not significantly different); but, at this dose, after a recovery period of 2 days, neurologic deficits in the lower limbs of the rat were observed. Redetermination of MAC revealed a 10% decrease in MAC defined by movement of the upper limbs and approximately a 30% decrease in MAC defined by movement of the lower limbs. Because of this, larger IT doses of MK 801 were not studied. There was no injury in rats receiving smaller doses of MK 801 IT.

Table 1: The Effect of MK 801 Infused IV, Intrathecally (IT), and Intracerebroventricularly (ICV) on the Minimum Alveolar Anesthetic Concentration^a (MAC) of Isoflurane

Figure 1: Intrathecal (IT), intracerebroventricular (ICV), and IV infusions of MK 801 produce a dose-related decrease in isoflurane MAC (minimum alveolar anesthetic concentration producing immobility in 50% of rats). ICV infusion of MK 801 decreased MAC for isoflurane <25%. The dose-response curve of IT versus IV infusion is shifted to the left by a factor of 3 at doses larger than 1.25 μg · kg⁻¹ · min⁻¹. IV infusion of MK 801 in the absence of isoflurane did not produce immobility, even with doses 270-fold larger than the largest applied IT dose (not shown). Values are presented as mean, sd.

Ascending concentrations of IV MK 801 decreased isoflurane MAC significantly without evidence of neural injury. The dose-response curve for IT infusion was to the left of the curve for IV infusion by a factor of 3 at doses more than 1.25 μg · kg⁻¹ · min⁻¹. IV infusion of MK 801 in the absence of isoflurane did not produce immobility, even at doses 270-fold larger than the largest IT dose. The largest concentrations of MK 801 given ICV that did not produce apparent neurological deficit in the rats decreased the MAC of isoflurane <25%.

MK 801 was infused for 310 ± 88 min IT, 228 ± 47 min IV, and 260 ± 23 min ICV. When MK 801 was given at 10 μg · kg⁻¹ · min⁻¹ IT, the MK 801 concentration in the lower spinal cord (16.8 ± 10.8 μg/g) exceeded the concentration found in the spinal cord (6.76 ± 1.57 μg/g) after IV infusion of 50 μg · kg⁻¹ · min⁻¹ by a factor of >2 (Table 2). Note that neither the IV nor the IT infusion produced immobility in the absence of isoflurane.

In our multiple-regression model, neither the upper cord, brain, or cortex MK 801 concentrations correlated significantly with isoflurane MAC; only the lower spinal cord concentrations of MK 801 correlated significantly (Fig. 2;P < 0.001). The correlation coefficient for the regression was 0.8.

Figure 2: The percentage decrease in minimum alveolar anesthetic concentration (MAC) from the administration of MK 801 correlated significantly (P < 0.001) with MK 801 concentrations in lower spinal cord samples. No other MAC change significantly correlated with MK 801 at other central nervous system concentrations (either whole brain, cerebral cortex, or upper spinal cord). The final regression model chosen was a logistic model in which the absolute value of the change in MAC was expressed only as a function of lower cord concentrations. The correlation coefficient, r, was 0.8.

Necropsy examination after IT and ICV infusion of MK 801 with 0.01% methylene blue at a rate of 1–2 μL/min confirmed that the infusions were confined to the lumbar and lower thoracic portions of the IT space after IT infusion and to the third and fourth ventricle after ICV infusion. Studies of MK 801 infusion with and without application of methylene blue 0.01% did not differ in the decrease in isoflurane MAC that the infusion produced.

Discussion

Our results demonstrate that MK 801 has a MAC-sparing effect and that this effect results primarily from an action on the spinal cord. The MAC-sparing effect of MK 801 correlated with the lower cord MK 801 concentrations in a dose-dependent manner in the absence of cord injury, producing a maximum change in MAC of approximately 65%. Furthermore, IV administration of MK 801 at doses 270-fold larger than the largest IT doses applied did not produce anesthesia in the absence of isoflurane.

Thus, spinal NMDA receptors may partially mediate the capacity of isoflurane to produce immobility. This accords with the observation of McFarlane et al. ⁽⁸⁾ who found that IV infusion of MK 801 produced a maximal decrease of 80% in halothane MAC. The observation that MK 801 is not a complete anesthetic suggests that blockade of NMDA receptors alone is not sufficient to produce anesthesia. Similarly, Ishizaki et al. ^(6,7) found that IT infusion of MK 801, CPP, or 7CKA produced a maximal decrease of 30% in MAC. We obtained a larger decrease, presumably because we applied larger doses. Ishizaki et al. ^(6,7) gave boluses of 0.1 to 30 μg of MK 801 (in 10 μL of saline), whereas we gave infusions of 1.25 to 20 μg/min of MK 801 for three to four hours for a much larger cumulative dose (∼200 to 4000 μg).

The ratio of the MK 801 concentrations found in the spinal cord versus brain in our study after IV infusion (1:2) differed from the ratio (2:1) found in the study of Mather et al. ⁽¹³⁾. Studies in cats by Young ⁽¹⁴⁾ indicate that blood flow to the spinal cord is substantially less than in the brain. Because we used a 2.5 times shorter equilibration time than Mather et al. ⁽¹³⁾, MK 801 might have had insufficient time to accumulate in the more lipid-rich white matter of the spinal cord compared with the cortical regions of the brain.

In summary, our results suggest that the capacity of MK 801 to decrease the MAC of isoflurane primarily results from an effect on the spinal cord, with a far smaller contribution of supraspinal centers. NMDA receptor blockade by MK 801 does not produce complete anesthesia; however, NMDA receptors may mediate, in part, the immobilizing action of isoflurane.