Nitrous Oxide Activates GABAergic Neurons in the Spinal Cord in Fischer Rats

Adult male Fischer and Lewis rats were used throughout this study (B&K Universal, Grimston Aldbrough Hull, UK). The study protocol of the animal experiments was approved by the Home Office of the United Kingdom (London, UK), and all efforts were made to minimize animal suffering and reduce the number of animals used.

Gas exposure was performed in a Plexiglas chamber (45 cm long, 22.5 cm wide, and 20 cm high). A mixture of nitrous oxide and oxygen gas and air was continuously delivered from an anesthetic machine into the chamber through an inflow port, which was exhausted through an outflow port following through the Anesthetic Gas Scavenging System (Ohmeda, Louisville, CO). Gas flow rate was 10 l/min. Gas concentrations, including those for nitrous oxide, oxygen, and carbon dioxide, in the chamber were measured continuously by infrared gas spectrometry (Ohmeda 5250 RGM).

After exposure either to nitrous oxide or to air (control), animals were injected intraperitoneally with 100 mg/kg sodium pentobarbital. After being deeply anesthetized, the animals had thoracotomy by transverse incision at the level of diaphragm and midline sternotomy and were perfused with 0.1 m phosphate-buffered saline (PBS) followed by 4% paraformaldehyde in 0.1 m phosphate buffer through a 16-gauge cannula inserted through the left ventricle into the ascending aorta. (An incision was made in the right ventricle for drainage.) Animals were decapitated, and the entire spinal cord was expelled from the spinal canal by rapid injection of PBS at the sacral vertebral level. The spinal cord was stored in 30% sucrose in 0.1 m phosphate buffer for at least 24 h at 4°C. A portion of the spinal cord at the lumbar level (i.e.,  approximately 5-mm length) was cut by a razor blade and freeze-mounted in embedding matrix, and 30-μm transverse sections at approximately L5 level were cut under −15°C, which were collected in PBS as free-floating sections.

Free-floating spinal cord sections were first incubated for 1 h in blocking solution consisted of 3% rabbit serum and 0.3% Triton X in PBS (PBT). They were incubated overnight with goat anti–c-Fos antibody (1:10,000, catalog No. sc-52-G, Santa Cruz Biotechnology, Santa Cruz, CA) in blocking solution consisting 1% normal rabbit serum in PBS on a shaker at room temperature. The sections were rinsed with 1% rabbit serum in PBT and were incubated for 1 h with biotinylated rabbit anti-goat immunoglobulin (1:200; Vector Laboratories, Burlingame, CA) in the same solution. The sections were rinsed with PBT and were incubated for 1 h with avidin-biotin-peroxidase complex (Vector Laboratories) in PBT. Visualization of the immunohistochemical reaction was achieved by incubation with 3,3′-diaminobenzidine (DAB) with nickel-ammonium sulphate to which hydrogen peroxide was added (DAB kit, Vector Laboratories). After the staining procedure was completed, the sections were rinsed in PBS followed by distilled water and placed on slide glasses, which were dehydrated in 100% ethanol, and cleared in 100% xylene, and cover slips were placed.

Free-floating spinal cord sections were first incubated for 1 h in blocking solution consisting of 3% donkey serum (Chemicon International, Temecula, CA) in PBS. They were incubated overnight with goat anti–c-Fos antibody (1:1,000, catalog No. sc-52-G, Santa Cruz Biotechnology) and rabbit anti–glutamic acid decarboxylase (GAD) antibody (1:1,000, catalog No. GC3008, Affiniti Research Products, Mamhead, UK) in 1% donkey serum in PBS on a shaker at −4°C. Sections were rinsed with PBS and incubated for 1 h in darkness with a mixture of Cy3-conjugated donkey anti-goat secondary antibody (1:200, Jackson ImmunoResearch Laboratories, West Grove, PA) and fluorescein isothiocyanate (FITC)–conjugated donkey anti-rabbit secondary antibody (1:200, Jackson ImmunoResearch Laboratories) in 1% donkey serum in PBS. Sections were rinsed with PBS, floated in water, and mounted on slide glasses. After being dried in darkness, one drop of VectaShield (Vector Laboratories), a special mounting medium for fluorescence, was put on each slide glass, and a cover slip was placed.

Using a DAB staining, c-Fos–positive cells were identified by dense black nuclear staining under bright-field microscope (Olympus Model BX50 Research Photomicroscope; Olympus Optical, Southall, Middlesex, UK). Randomly selected sections were photographed using a digital camera (Olympus Digital Camera, model C2020Z, Olympus Optical). The number of c-Fos–positive cells was counted for each area of the spinal cord (i.e.,  laminae I–II [superficial area], laminae III–IV [nucleus proprius area], laminae V–VI [neck area], and laminae VII–X [ventral area]), according to the method by Presley et al. , 16based on the laminar scheme in rats originally developed by Molander et al.  17Five photos were taken from each rat, and the mean number of c-Fos–positive cells per section was calculated. At least three to four animals were examined for each group, and the number of c-Fos–positive cells in each group was calculated as mean ± SD. The investigator was blinded to the treatment that the experimental groups had received.

Colocalization of c-Fos and GAD staining was examined by confocal microscopy (Bio-Rad Microradiance MR/AG II System equipped with Argon/Green HeNe lasers; Bio-Rad Laboratories, Hercules, CA). Cy3 and FITC staining were detected separately with E570 LP (red) and HQ500 LP (green) emission filters, respectively.

The numbers of c-Fos–positive cells were analyzed by one-way analysis of variance, and Bonferroni correction was used as a post hoc  test. A P  value < 0.05 was considered to be statistically significant.

The number of c-Fos–positive cells in the control group was approximately 40 cells/section, which was consistent among different experiments. The number of c-Fos–positive cells increased in a time-dependent manner after 75% N²O exposure reaching to the peak level of approximately 130 cells/section at approximately 90 and 120 min (fig. 2). Increase in the number of c-Fos–positive cells after 90 min of nitrous oxide exposure was dose-dependent (fig. 3) and was observed in the entire area excluding laminae I and II (figs. 4 and 5), but appeared to be specifically concentrated in laminae III and IV (fig. 5).

The number of c-Fos–positive cells in the control group was 35.0 ± 15.7 cells/section (fig. 5). After 75% N²O exposure for 90 min, the number of c-Fos–positive cells increased slightly in all laminae but statistical difference between the control groups was achieved only in laminae I and II (fig. 5).

Representative pictures of double staining for c-Fos and GAD are shown in figure 6. Compared with DAB staining for c-Fos alone, the number of c-Fos–positive cells in both control and nitrous oxide groups was lower in double-stained sections. The most well-stained undamaged section was selected from each animal, and all c-Fos–positive cells in half of each section were examined for colocalization with GAD. The prevalence of c-Fos and GAD colocalization was calculated for all the animals in each group (table 1). In the control group, a total of 53 c-Fos–positive cells were examined, and in 26 (49.1%) of them, GAD was colocalized. In the nitrous oxide group, a total of 216 cells were examined, and in 182 (84.3%) of them, GAD was colocalized. Because the number of c-Fos–positive cells that were not colocalized with GAD was almost the same between the control (27) and nitrous oxide (34) groups, these results indicate that almost all c-Fos–positive cells induced by nitrous oxide also contained GAD.

c-Fos, a protein product of the immediate early gene, c-fos , has been commonly used as a histologic marker of transsynaptic neuronal activation since Hunt et al.  18first reported it in 1987. In the present study, we used c-Fos as a marker for neuronal pathways activated by nitrous oxide to examine whether GABAergic inhibitory interneurons are involved in the antinociceptive effect of nitrous oxide in the rat spinal cord. First, we characterized the expression pattern of c-Fos after nitrous oxide administration. Second, using double-staining techniques and confocal microscopic analyses, we demonstrated that the c-Fos–positive cells induced by nitrous oxide were GABAergic neurons. In Fischer strain, exposure to 75% N²O increased the number of c-Fos–positive cells in the spinal cord approximately threefold beyond baseline expression. This effect was time-dependent (fig. 2) and dose-dependent (fig. 3). To the contrary, this effect was observed less prominently in the Lewis strain, which is consistent with a lack of antinociceptive effect by tail-flick test in this strain in the previous study of Fender et al.  15In another recent study, our group has shown that nitrous oxide administration increased c-Fos expression in noradrenergic neurons of the brainstem (i.e.,  A5, A6 [locus ceruleus], and A7) in Sprague-Dawley rats, 13a strain in which nitrous oxide shows reasonably strong antinociceptive effect in the tail-flick test. Furthermore, when the rats were injected intracerebroventricularly with a mitochondrial toxin, saporin, which was coupled to the dopamine β hydroxylase antibody, those noradrenergic cells were almost entirely eliminated and the animals no longer showed the antinociceptive effect of nitrous oxide. Together with the results from the present study, our findings provide functional evidence for the involvement of descending noradrenegic neurons in the antinociceptive effect of nitrous oxide. In addition, we have found a striking colocalization of nitrous oxide–induced c-Fos expression with GABAergic neurons (table 1and fig. 6), suggesting that GABAergic inhibitory interneurons and their downstream effectors are at least partly involved in the antinociceptive effect of nitrous oxide.

The distribution patterns of c-Fos–positive cells in the control and nitrous oxide groups of Fischer strain within the spinal cord are summarized in table 2and juxtaposed with information regarding the localization of different neurons and receptors from previously published reports. In control animals that were exposed to air, the few c-Fos–positive cells were randomly distributed throughout all laminae. In the nitrous oxide group, the number of c-Fos–positive cells increased significantly, especially in laminae III–VI, and were almost entirely colocalized with GABAergic neurons (table 1and fig. 6). Although GABAergic neurons are ubiquitous in the entire spinal cord (except for lamina IX), 19they are most prevalent in laminae I–III, which are proposed to be inhibitory interneurons. 20The noradrenergic descending neurons from the brainstem have been shown to terminate in the spinal cord mainly in laminae I–IV, 21although the distribution of the termini depends on the genetic background and its origin in the brainstem (i.e.,  A5, A6 [locus ceruleus] or A7). 22,23Based on our data, we suggest that nitrous oxide–induced c-Fos–positive cells, particularly in laminae III and IV, are activated by nitrous oxide through descending noradrenergic neurons, although this has not yet been tested definitively. Nitrous oxide– induced c-Fos–positive cells in other laminae could be the result of (1) the activation of a minority group of descending noradrenergic neurons that reach laminae other than I–IV, (2) activation of other descending neurons (e.g.,  serotonergic or opioidergic neurons), or (3) direct activation by nitrous oxide in the spinal cord. Further investigations are necessary to determine their origins and whether the activation of these neurons is functionally involved in the antinociceptive effect of nitrous oxide. Interestingly, the distribution pattern of nitrous oxide–induced c-Fos–positive cells does not correspond with that of the primary afferent neurons, which are mainly in laminae I, II, and V, 24or that of α²adrenoceptors, which mostly corresponds to that of the primary afferent neurons. 25,26The α¹adrenoceptors are localized throughout the spinal cord. 27Because nitrous oxide administration did not increase the number of c-Fos–positive cells in laminae I and II in Fischer strain, descending noradrenergic neurons that terminate at these laminae may be acting in an inhibitory manner, possibly through α²adrenoceptors (any inhibitory action caused by nitrous oxide would not result in increasing c-Fos expression).

GABA usually produces postsynaptic inhibition by hyperpolarizing the postsynaptic cell (fig. 1). In support of this mechanism in the spinal cord, a recent electrophysiologic study has shown that norepinephrine applied in the perfusion solution of the sliced rat spinal cord preparation activates GABAergic inhibitory activity in the substantia gelatinosa of the spinal cord (i.e.,  lamina II in the dorsal horn). 28It also has been demonstrated that this effect of norepinephrine is mediated by α¹but not by α²adrenoceptors. 28GABA may also act as a depolarizing transmitter on the presynaptic terminals of the primary afferent neurons to produce presynaptic inhibition (fig. 1). 29In an immunohistochemical study using anti-GABA antiserum, it has been shown that some unmyelinated primary afferent neurons are subject to presynaptic inhibition at GABAergic neurons. 30Furthermore, supporting the existence of the latter function, gene expression of the GABA^Areceptor has been demonstrated in the dorsal root ganglia in addition to the dorsal horn. 31 

The hypothetical neuronal pathways shown in fig. 1are not an exhaustive representation because both serotonergic and opioidergic inhibitory neurons also exist, although their functional importance in the antinociceptive effect of nitrous oxide is not clear. Yet, nitrous oxide increased c-Fos–positive cells not only in laminae III and IV, where noradrenergic descending inhibitory neurons are mainly distributed, but also in other laminae, suggesting the possibility that other pathways are activated, although these may not necessarily be involved in the antinociceptive effect of nitrous oxide. It is entirely possible that multiple pathways work in concert to produce the antinociceptive effect of nitrous oxide; although all may be needed, none is sufficient.

In summary, we have demonstrated in the present study that nitrous oxide activates GABAergic neurons in the dorsal horn of the rat spinal cord. Our finding suggests that GABAergic interneurons and their downstream effectors play at least a part in the antinociceptive action of nitrous oxide in the spinal cord.

The authors thank Miss Laura Nelson, B.A. (Department of Anaesthetics and Intensive Care, Imperial College of Science, Technology and Medicine, and Magill Department of Anaesthesia, Intensive Care and Pain Management, Chelsea and Westminster Hospital, London, UK), a doctoral student of their department, for general assistance provided.