The Analgesic Action of Nitrous Oxide Is Dependent on the Release of Norepinephrine in the Dorsal Horn of the Spinal Cord 

The experimental protocol was approved by the Animal Care and Use Committee at the Palo Alto Veterans Administration Medical Center. Male Sprague-Dawley rats (Bantin and Kingman, Fremont, CA) weighing 250–380 g were used. All tests were performed between 9 AM and 4 PM. A total of 104 rats were used. Each animal was used for only one set of studies to eliminate possible interaction between different doses and routes of drugs.

The modified methods of Skilling et al. , 9Liu et al. , 10and Peng et al.  11were followed. Briefly, male Sprague-Dawley rats (300–400 g) were anesthetized with isoflurane and the lateral surfaces of vertebra T13 were exposed. Bilateral holes were made through the bone to expose the spinal cord at the level of the dorsal horn.

Except for a 2-mm dialysis zone, dialysis fibers (diameter 200 mm; molecular weight cutoff = 9,000; Spectrum Laboratories, Laguna Hills, CA) were coated with a thin layer of silicon rubber. One end of the fiber was connected to a 90° angled stainless tubing made from a 22-gauge stainless steel needle. A stainless steel dissecting pin was affixed to the lumen of the other end of the fiber. By pushing the pin through the spinal cord and pulling it out the other side, the fiber was positioned so the uncoated portion of the fiber was located within the dorsal horn of the spinal cord. The pin was then cut and the free end of the fiber was attached to a length of PE 20 tubing (Clay Adams, Sparks, MD) with cyanoacrylic glue (Krazy Glue; Elmer's Products, Inc., Columbus, OH). The dialysis fiber, the initial part of the stainless tubing and PE 20 tubing were affixed to the exposed vertebra T13 with dental acrylic, and both tubing ends were externalized in the lumber region. The skin was then sutured around the tubing.

The next day, any animals displaying any signs of limb paralysis were rejected for further study. In the remaining, neurologically intact animals, the two tubing ends were attached to the fluid swivel in the CMA/120 system (Stockholm, Sweden) to allow for free movement, then attached to an infusion pump (Harvard Apparatus, South Natick, MA). The spinal cord was perfused with an artificial cerebrospinal fluid (aCSF) solution (NaCl: 125 mM; NaH²PO⁴: 0.5 mM; KCl: 2.5 mM; Na²HPO⁴: 2.0 mM; MgCl²: 0.25 mM; and CaCl²: 1.0 mM; pH adjusted to 7.2) at a flow rate of 1.3 μl/min for 180 min to establish a diffusion equilibrium. Samples were collected at 30-min intervals in a vial containing 0.9 μl perchloric acid to achieve a final concentration of 2% and analyzed immediately after collection. All tubing and collection vials were covered with aluminum foil to prevent degradation of norepinephrine by light.

Norepinephrine was separated by reverse-phase chromatography on an ESA column (catecholamine R-80; ESA Microdialysis, Medford, MA) maintained at 24°C with a column heater. The mobile phase (Cat-A-Phase; ESA) was delivered at a flow rate of 1.0 ml/min using a Beckman 118 Solvent Module (Fullerton, CA). Samples (20 μl) were injected with a BIO-RAD Model AS-100 HPLC Automatic Sampling System (Hercules, CA), and norepinephrine was detected coulometrically (model 5011; ESA). Potentials for the first and second electrodes were set at +100 mV and −300 mV, respectively. A conditioning cell was set at +350mV and was placed before the analytic cell. Retention time, peak area, and concentrations of norepinephrine in the dialysate were measured by comparison with known standards and were determined with the Dynamax MacIntegrator software system (Rainin Instrument Co., Inc., Emeryville, CA). The detection limit for norepinephrine in our assay varied between 0.5–1.6 pg/sample.

To verify that the fiber traversed the dorsal horn, the spinal cord was removed and fixed in 10% formalin for histologic confirmation of cannula placement. Only animals with the cannula located below lamina I and above the central canal were included in the study.

Rats were anesthetized with halothane, and laminectomy was performed at the T3–T4 level. Spinous processes and laminae were removed to expose a circular region of dura. The dura was opened and the spinal cord was severed at the T3–T4 level. The muscles were sutured over the laminectomy site and the skin was closed with wound clips. The animals were exposed to N²O and tested for tail flick response within 6 h after surgery.

Rats were anesthetized with isoflurane, an incision was made over the cervical spine, and a small puncture was made in the dura mater. PE-10 polyethylene tubing (0.28 mm ID) was threaded 8.5 cm into the intrathecal space so the tip of the catheter was positioned at the lumbar level. This tubing was then sutured in place, and the skin was sutured over the tubing. After allowing 7 days for recovery, DSP-4 (100 or 300 μg) was administered in 10 μl normal saline using a perfusion pump at a rate of 10 μl/min followed by a 10-μl flush of normal saline. Behavioral testing or killing for the determination of spinal norepinephrine levels was performed 10 days later.

The levels of norepinephrine in the lumbar enlargement of spinal cord were measured using the high-performance liquid chromatography. Rats were exposed to 100% carbon dioxide for 35 s and then killed. The spinal cord was rapidly extruded from the spinal canal using ice-cold saline; the lumbar enlargement was isolated and weighed. The tissue was put into 600 μl perchloric acid, 2%, and 2 × 10⁻⁸M dihydroxybenzylamine, and homogenized and centrifuged at 1,200 g  for 15 min at 4°C. The supernatant was removed and stored at −80°C for later analysis.

Nociception was assessed by the tail flick response to a noxious thermal stimulus, as previously described. 12In brief, a high-intensity light beam was focused on the tail, and the time for the rat to move its tail out of the light was recorded as tail flick latency. The latency from three sites on the tail were averaged and a cut-off time of 10 s was predetermined to prevent tissue damage. Baseline measurements consisted of a set of three tail flick determinations at 2-min intervals. Baseline tail flick latencies ranged between 3 and 4 s. In some cases, percent maximal possible effect (%MPE) was calculated as

All gas exposures were performed in a clear plastic chamber (92 × 48 × 38 cm) with a sliding door on one side (for insertion of the rats). This airtight chamber was large enough to contain the infusion pump and the analgesimeter device. Fresh test gases (10 l/min) were introduced into the chamber via  an inflow port, circulated throughout the chamber by a small fan, and purged by vacuum set to aspirate at the same rate as the fresh gas inflow. Oxygen concentration in the chamber was maintained between 22–30%, while N²O concentration was maintained at 0 or 70% by adjusting the flow rates of N²O, air, and nitrogen (Liquid Carbonic, Houston, TX). Gas concentrations were measured continuously and flow rates were adjusted appropriately to maintain the desired concentrations.

Release data were analyzed by analysis of variance for repeated measures and a posteriori  by Scheffé or Bonferroni tests. Nociceptive data were analyzed by unpaired Student t  test or analysis of variance for repeated measures and a posteriori  by the Bonferroni multiple comparisons test when appropriate.

Spinal transection at level T3–T4 did not affect the baseline tail flick latency but did block the analgesic action of 70% N²O (fig. 1).

Nitrous oxide caused approximately a fourfold increase in norepinephrine release within the first 30-min collection period (fig. 2). This level of release decreased in subsequent collection periods, reaching baseline values during the 60- to 90-min collection period.

Systemic opiate antagonists, such as naloxone, block the analgesic action of N²O, 13possibly by antagonizing the action of endogenously released opiates at the level of the PAG. 7To determine whether an opiate antagonist also suppressed norepinephrine release evoked by N²O, the levels of spinal norepinephrine release were measured before and after naltrexone, a long-acting opiate antagonist. Naltrexone (10 mg/kg intraperitoneal), a dose that antagonized N²O-induced analgesia for a least 3 h (fig. 3 A) but did not affect baseline tail flick latencies (data not shown), had no effect on the basal release of norepinephrine alone but blocked the N²O-evoked release (fig. 3B).

In animals pretreated with DSP-4 in which the norepinephrine levels were depressed to 22% of control values (fig. 4 A), the baseline tail flick latency was not changed (fig. 4B), but the analgesic action of 70% N²O as measured by percent maximal possible effect was greatly attenuated (fig. 4C).

This study shows that N²O causes the release of norepinephrine in the spinal cord in the awake freely moving rat. This release decreases with continued N²O exposure and is dependent on the presence of a functional opiate receptor. In keeping with the hypothesis that N²O analgesia is mediated by the release of norepinephrine, depletion of norepinephrine stores by DSP-4 or elimination of descending noradrenergic transmission by spinal cord transection attenuates N²O analgesia. Although spinal cord transection and the insertion of the microdialysis fiber may cause a degree of nerve injury that could modify norepinephrine release characteristics, all effects were temporally related to the onset of N²O exposure. Although DSP-4 treatment is known to also deplete serotonin, 14the lack of analgesic effect of N²O in DSP-4–treated animals coupled with the knowledge that spinal α²adrenoceptors are necessary for N²O analgesia 1indicates that the noradrenergic system is necessary for N²O analgesia.

By measuring neurotransmitter turnover. others have found that N²O stimulates norepinephrine turnover in various brain regions. 15N²O also has been found to increase brain dopamine turnover, 16although when turnover in discrete regions of the brain was evaluated, N²O caused a decreased turnover rate of dopamine in the hippocampus and striatum but an increase in the olfactory bulb. 15These studies indicate that N²O causes region-specific alterations in steady state levels and turnover rates of dopamine and norepinephrine within the central nervous system. In addition, N²O suppression of the activity of wide-dynamic-range neurons, in which activity has been linked to pain transmission, depends on an intact descending inhibitory pathway. 17 

The mechanism by which N²O affects norepinephrine release is not clear, although current evidence supports a pivotal role for release of endogenous opiate peptides. Previously, we found that an opiate antagonist administered into the PAG reduced the analgesic action of N²O. 7Coupled with the current finding that N²O-evoked norepinephrine release is blocked by naltrexone, these data are consistent with the hypothesis that N²O causes the release of endogenous opiate peptides in the PAG, as has been observed by others. Candidate peptides known to be present in the PAG are enkephalin, β-endorphins, or dynorphin. 18There is some evidence that β-endorphins may be involved because their release is stimulated by N²O both in vivo  3and in vitro.  4Antiserum against β-endorphin but not against metenkephalin also blocked the N²O-induced antinociception in rats in the hot plate test. 19However, other studies indicate that metenkephalin may be involved in the analgesic effect of N²O because cerebrospinal fluid levels of metenkephalin taken from the third ventricle of awake dogs increased significantly; no changes were noted in concentrations of dynorphin A, dynorphin B, or β-endorphin. 5N²O also caused an increase in metenkephalin-like immunoreactivity in the brain stem, spinal cord, hypothalamus, and corpus striatum in rats. 20Enkephalin- and dynorphin-containing cells are present in the PAG but are also found in many other areas of the brain 21; therefore, N²O may have an action in brain regions rostral to the PAG. It is known that analgesia can be evoked by electrical stimulation of various sites in the brain, such as the habenular complex, 22,23arcuate nucleus, 24and amygdala. 25Similar to N²O analgesia, this analgesia is sensitive to blockade by opiate 23–25and α-adrenergic antagonists. 23,25Further work is necessary to elucidate the identity of the opiate peptides and mechanism by which N²O causes their release.

Several lines of evidence indicate that spinally projecting noradrenergic neurons mediate the antinociception produced by the activation of the PAG. Intrathecal injection of α²-adrenergic antagonists can reduce the antinociception produced by either electrical 26or chemical 27stimulation of PAG neurons. Intrathecal injection of an α²antagonist can also attenuate the antinociception produced by microinjection of morphine in the PAG. 28Additionally, discrete injection of morphine into the PAG produces an increase in norepinephrine metabolites in the spinal cord, and its analgesic effects are attenuated by previous depletion of norepinephrine stores in the spinal cord. 29Electrophysiologic studies also showed that α²adrenoceptors in the spinal cord contribute to the mediation of the PAG-induced inhibition of dorsal horn cell activity. 11,30 

There are three nuclei from which noradrenergic neuronal projections to the spinal cord originate. Using the tract tracing methods and combinations of lesions with histochemical methods, numerous studies have shown that noradrenergic neuronal projection to the spinal cord originates from the A5, A6 (locus coeruleus, sub-coeruleus), and A7 cell groups. 31,32In the same substrain of animals used in this study, Basbaum's 33laboratory showed that the locus ceruleus is the major source of noradrenergic fibers to the dorsal horn region. All these nuclei receive projections from the PAG. 34–37Subsequent studies will seek to identify which noradrenergic nucleus is responsible for the analgesic action of N²O.

We have not conclusively established whether N²O-induced norepinephrine release is mediated by activation of the descending noradrenergic pathways at a supraspinal or spinal site. In the context of our previous studies that the analgesic response to N²O is caused by activation of opiate receptors in the PAG, 7and our current finding that naltrexone blocks N²O-induced norepinephrine release in the dorsal horn of the spinal cord, we believe that the weight of evidence supports a supraspinal site of activation.

There is clear evidence that the acute antinociceptive properties of N²O decrease over a relatively short time during continuous administration. 38–41The results of this study show that the release of norepinephrine also diminishes over roughly the same time span, pointing to a site for tolerance proximal to norepinephrine release. Interestingly, the ability of N²O to affect dopamine turnover also diminishes progressively. 16However, the precise site of tolerance is not known. Prolonged exposure (18 h) to N²O decreases opiate receptor density in rat brain stem, 42and this would be consistent with increased release of endogenous opiate peptides causing a down-regulation of the receptor system. This raises the possibility that, depending on the mechanism involved, “cross-tolerance” may develop to either exogenous opiate or α²-agonist administration, or both, after N²O exposure. Knowledge of the mechanism for N²O tolerance may help to identify people who might be less sensitive to the analgesic action of N²O, and also allow the design of strategies to mitigate the development of tolerance to prolong the analgesic effect of N²O.