Introduction
Hyperalgesia, one of the main features of chronic pain, develops closely associated with increased glutamatergic neurotransmission in the dorsal horn of the spinal cord, especially to
N-methyl-D-aspartate (NMDA) receptor activation. Accordingly, a variety of NMDA receptor antagonists, acting on different sites of the receptor, have demonstrated antinociceptive efficacy on chronic experimental inflammatory and neuropathic pain syndromes [
1‐
5]. NMDA receptor activation is followed by downstream modifications of intracellular signaling, including activation of nitric oxide synthase (NOS), which catalyzes the formation of nitric oxide from arginine. Nitric oxide is a gaseous mediator that seems to have a pivotal role in multisynaptic local circuit nociceptive processing in the spinal cord. It is generated by three major NOS isoforms: nNOS (neuronal NOS) and eNOS (endothelial NOS), which are calcium-dependent constitutive enzymes, and iNOS (inducible NOS), which a calcium-independent inducible isoform [
6‐
8].
Intrathecally administered NMDA induces short-term hyperalgesia, whereas systemic and intrathecal administration of the non-selective NOS inhibitor
Nω-nitro-L-arginine methyl ester (L-NAME) blocks NMDA-induced hyperalgesia, suggesting that the generation of nitric oxide contributes to this response [
9]. In addition, intrathecal L-NAME prevents thermal pain hypersensitivity in rats after carrageenan injection [
10] and sciatic nerve constriction-induced injury [
11], as well as thermal and mechanical hypersensitivity induced in mice by the intraplantar administration of complete Freund's adjuvant (CFA) [
12]. Besides, increased expression of one or more of the three NOS isoforms has been shown in the spinal cord of rodents after carrageenan injection into a hindpaw [
13], intraplantar injection of CFA [
12] and formalin [
14], and intradermal injection of capsaicin [
15]. However, in these models of tonic experimental pain, only fast and short-term hyperalgesia and allodynia are tested. With regard to changes in NOS expression in long-term experimental models of chronic pain, the available data refer only to the spinal nerve ligation model in rats [
16,
17], whereas expression of NOS in the spinal cord in rat models of arthritic pain was only partly studied [
18]. It has been shown that monoarthritic pain is highly sensitive to NMDA antagonists [
19] and to L-NAME [
20], suggesting an involvement of the nitric oxide/cyclic GMP cascade in downstream NOS activation in the spinal cord. However, there have been no studies exploring the effect of NMDA receptor blockade on NOS expression in the dorsal horn. The aim of this work was therefore to study the expression of nNOS, iNOS and eNOS in the dorsal horns of monoarthritic rats, and to explore how the expression of NOS isoforms in this model of chronic pain is modified by pharmacological blockade of spinal cord NMDA receptors with competitive and uncompetitive antagonists.
Discussion
Results showed that the nNOS isoform was expressed in the lumbar dorsal horn of intact control rats, whereas the iNOS and eNOS isoforms could not be detected in the same spinal cord regions of these animals. This is in accordance with earlier studies showing moderate expression of nNOS in the spinal cord of intact rats [
25]. In addition, the present results showed that the three major NOS isoforms were expressed bilaterally in the dorsal horns of monoarthritic rats, which is in agreement with previous observations that injection of incomplete Freund's adjuvant into the knee joint cavity increased the expression of the nNOS and iNOS isoforms in the lumbar enlargement of rats [
18]. Increased expression of nNOS and/or iNOS has also been shown in the lumbar spinal cord of rats after carrageenan injection into a hindpaw [
12], intraplantar administration of formalin [
14,
26] or CFA [
12], and intradermal injection of capsaicin [
15]; large increases in the expression of nNOS and iNOS have been observed in dorsal horns taken from L4-L6 spinal cord segments of rats subjected to ligation of the sciatic nerve [
16,
17].
The differences in NOS expression levels found in these studies could be the result of the different procedures used to induce persistent pain (namely long-lasting chronic pain after sciatic nerve ligation and intra-articular administration of Freund's adjuvant, versus tonic pain after administration of carrageenan, formalin or capsaicin), the different time course of the pain induced, and/or the different times within the hyperalgesic process at which NOS determinations were made, but they could also be the result of differential glial activation in the spinal cord. In fact, a role of spinal glia has recently been described for the initiation and early maintenance of inflammatory pain facilitation in monoarthritic rats [
27]. The fact that levels of eNOS in the spine have been shown to increase in astrocytes but not in neurons after the injection of carrageenan [
28] suggests the participation of spinal cord glia in the increases in eNOS reported here. In this regard, the regulation of only the genes encoding iNOS in glia has been well described [
29], but more recently reverse transcriptase-mediated polymerase chain reaction and Western blot analyses have revealed that mRNAs encoding the constitutive nNOS and eNOS as well as the corresponding proteins were expressed in human astrocytes [
30], thus opening the possibility that glial activation could account, at least in part, for the observed increases in expression of the three NOS isoforms in the dorsal horn of monoarthritic rats. Further investigation will be required to determine to what extent increases in NOS isoforms in dorsal lumbar spinal cord are dependent on neuronal and/or glial function.
Although in the present experiments the inflammation was induced on the right hind limb only, the changes in expression of NOS isoforms were found in the dorsal lumbar horn of both sides, suggesting a functional role for the contralateral innervation, at least during conditions of arthritic pain. In this respect it has been recognized that neurons in the spinal cord receive inputs from the contralateral side that, under normal conditions, are ineffective in generating an active response. However, there exist studies showing that on iontophoretic administration of NMDA or strychnine on one dorsal horn, the neurons of the opposite dorsal horn increased their excitability, thereby indicating that the contralateral input participates in the circuit dynamics of spinal nociceptive transmission [
31]. This makes possible a functional role for such crossed connections in neuronal sensitization after unilateral peripheral injury. Support for this hypothesis can be found in the study by Ondarza and colleagues [
32] showing a bilateral massive increase in calcitonin gene-related peptide staining (a marker for nociceptive endings) colocalized with GAP-43 (a marker for neurite sprouting) in the dorsal horn of rats subjected to unilateral spinal cord injury, indicating that mechanisms participating in the reorganization of nociceptive neuronal connections in dorsal horn circuits may be bilaterally activated in spite of the unilateral condition of the injury. An alternative explanation for the bilateral expression of NOS isoforms in monoarthritic rats could be that the process of disease spreads to the contralateral side, thereby stimulating the upregulation of certain molecules (namely NOS) in the contralateral dorsal horn. However, this mechanism no longer seems sustainable because in the monoarthritic model used here the contralateral hindpaw did not show gross inflammatory alterations [
22].
The administration of either NMDA antagonist, CPP or ketamine, produced similar changes in NOS expression in the dorsal horns of monoarthritic rats; that is, a decrease in nNOS but increases in iNOS and eNOS. In this regard, it is tempting to speculate that the decrease in nNOS expression induced by the intrathecal administration of CPP or ketamine could be related to the depressant action of these drugs on the activity of pain-transmitting dorsal horn neurons, but there are no conclusive data on this matter. For example, ketamine depresses the expression of c-Fos protein (an index of neuronal activation) in various brain areas [
33], but its effect on c-Fos expression in dorsal horn cells is still unexplored. In contrast, the intrathecal injection of NMDA blockers, such as CPP and ketamine, in rats exerts a rapid (a few minutes) but brief (about 1 hour) antinociceptive effect [
34‐
38], whereas in the present study the action of these drugs on NOS expression persisted for more than 24 hours, thus indicating that changes in NOS expression resulted from genomically mediated mechanisms occurring downstream of NMDA receptor blockade, an effect that clearly outlasts the antinociceptive action of the drugs. The fact that blockade of spinal NMDA receptors produced a substantial decrease in nNOS only 6 hours after intrathecal injection of the NMDA antagonists reveals an active and fast turnover of this constitutive enzyme in dorsal horn cells. Importantly, the present results indicate that the decrease in dorsal horn nNOS after NMDA receptor blockade was accompanied by simultaneous increases in iNOS and eNOS, suggesting compensatory interactions in the expression of the different NOS isoforms.
With regard to iNOS, it has been pointed out that nNOS inhibition could activate the nuclear factor NF-κB, which may lead to the induction of iNOS [
39,
40] through transcriptional activation of the genes encoding iNOS [
41]. Conversely, iNOS knockout mice showed increased expression of nNOS in the lumbar enlargement in comparison with wild-type mice 24 hours after challenge with carrageenan [
13]. With regard to eNOS, it has been shown that endothelial NOS expression in the spinal cord of nNOS knockout mice was upregulated compared with that in wild-type mice [
28]. The apparent compensation for the decrease in nNOS by rapid increases in iNOS and eNOS that we observed after administering CPP or ketamine is expected to have a functional meaning because all NOS isoforms are involved in producing the pronociceptive mediator nitric oxide, which may be relevant to prolonged treatment of chronic pain conditions, such as arthritic pain, with drugs that block the NMDA receptor.
Taken together, the present results suggest that blockade of spinal cord NMDA receptors by competitive and uncompetitive antagonists produces complex regulatory, genomically mediated, rapid (less than 6 hours) but long-lasting (more than 24 hours) changes in the expression of NOS isoforms in monoarthritic rats that may have some relevance for nitridergic neuronal/glial mechanisms involved in the pathophysiology of monoarthritis and in the pharmacological response to drugs interacting with NMDA receptor-dependent transduction pathways. In fact, it has been reported that L-NAME dose-dependently inhibits wind-up activity in the spinal cord of monoarthritic rats but not in normal controls [
20], suggesting that NOS-dependent nitridergic mechanisms have a non-significant role in acute pain, whereas it may be essential in chronic pain processing.
Several studies have sought to clarify the contribution of several neural components to joint injury, rather than to determine the effect of pharmacological blockade of specific NMDA receptors in rats with adjuvant-induced arthritis. For instance, the pioneer work of Levine and colleagues [
42] pointed out that no one class of nerve fiber is wholly responsible for the neurogenic component of inflammation in experimental arthritis but that large-diameter and small-diameter afferents, sympathetic efferents, and central nervous system circuits that modulate those fiber systems all influence the severity of joint injury in arthritic rats. However, no further studies establishing a relationship between lesion of specific neural components (namely dorsal rhizotomy) and changes in expression of NOS in the dorsal horn have been done. In addition, no previous pharmacological studies exploring the modifications of NOS expression in the lumbar spinal cord of arthritic animals had been performed with NMDA antagonists, even though pharmacological modulation of NOS proteins may be relevant for the successful treatment of long-lasting painful conditions, such as arthritic pain. Although the present study provides new evidence on this subject, several questions remain unanswered, such as the functional significance of decreased nNOS together with increased iNOS/eNOS after NMDA receptor blockade in spinal cord of monoarthritic rats, because no studies have examined whether nNOS loss and its possibly associated antinociceptive effect could be functionally compensated for by increases in iNOS and/or eNOS.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
CI and MD performed most of the experiments. TP performed experiments in inducing monoarthritis. CI, AH, LC and TP conceived the study and participated in the design, in the interpretation of results, and in drafting the manuscript. All authors read and approved the final manuscript.