| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
From the Division of Pharmacology and Toxicology, Indian Veterinary Research Institute, Izatnagar, Bareilly, Uttar Pradesh, India.
Address correspondence and reprint requests to S.K. Tandan, PhD, Principal Scientist, Division of Pharmacology & Toxicology, Indian Veterinary Research Institute, Izatnagar, Bareilly, UP 243 122, India. Address e-mail to sktandan{at}ivri.up.nic.in.
Abstract
BACKGROUND: An interaction between nitric oxide (NO) and cyclooxygenases (COX) in the production of prostaglandins in carrageenan-induced inflammation has been established. However, limited information is available about the interaction between inducible NO synthase (iNOS) and COX inhibitors in pain perception. Therefore, in the present study we assessed the nature of the interaction between S-methylisothiourea (a moderately selective iNOS inhibitor) with rofecoxib (selective COX-2 inhibitor) and mefenamic acid (a nonselective COX inhibitor) in formalin- induced pain in mice.
METHODS: The dose-response relation of S-methylisothiourea, rofecoxib, mefenamic acid, and their combination was studied in the late phase of formalin-induced pain in mice over the time spent in licking the hindpaw after formalin injection. The interaction was evaluated by simultaneous administration of fixed proportions of S-methylisothiourea with each COX inhibitor and the nature of the interaction was determined by isobolographic analysis.
RESULTS: Each drug alone produced a dose-dependent suppression of the late stage of formalin-induced behaviors with rank order of potency being rofecoxib > mefenamic acid > S-methylisothiourea. Isobolographic analysis of the combination of S-methylisothiourea with rofecoxib or mefenamic acid revealed a synergistic interaction. The experimental ED50 of the combination was significantly lower than the theoretical additive ED50 of the corresponding drug combination that substantiated the synergistic interaction between iNOS or NO and COX isoforms.
CONCLUSIONS: Our results explicitly indicate the synergistic nature of the interaction between NOS and COX inhibitors in formalin-induced nociceptive behavior in mice, and provide an alternative approach for controlling pain.
Cyclooxygenase (COX) and nitric oxide (NO) synthase (NOS) enzyme isoforms are often co-expressed in peripheral inflammatory cells and central neuronal and nonneuronal cells.1,2 The expressions of several of these isoforms are increasingly expressed during inflammation.3 The role of these enzyme systems in inflammatory pain states is supported by the antihyperalgesic effects of their respective inhibitors. Previous work has attributed the acute antihyperalgesic action of nonsteroidal anti-inflammatory drugs to the inhibition of constitutive spinal COX-2,4 whereas other reports have concluded that both COX-1 and COX-2 inhibition are required to achieve effective analgesia in inflammation.5,6 It has been demonstrated that carrageenan-induced hyperalgesia is inhibited by several COX-2 selective inhibitors, such as Dup-697, SC-58635, rofecoxib, and celecoxib.7–9 Mice preexposed to bacterial lipopolysaccharide for induction of COX-2 exhibited a higher writhing response when subjected to the acetic acid writhing test, thereby suggesting a role for COX-2 in nociception.10 The role of COX-2 in formalin-induced acute inflammatory pain has also been reported by other investigators.11 Importantly, the acuity of the effects of the COX-2 inhibitors in the formalin model emphasizes the relevance of constitutively expressed COX-2.
NO is a gaseous mediator molecule generated in high levels from the amino acid, l-arginine, by the catalytic action of the cytokine inducible enzyme, inducible NOS (iNOS), and constitutive enzyme isoforms endothelial NOS, and neuronal NOS (nNOS).12 NO is a potent modulator of pain perception and there is ample evidence to support its role in nociception. Intraplantar co-administration of l-arginine with formalin enhances the second phase nociceptive response to this noxious agent in mice and l-NAME (NG-nitro-l-arginine methyl ester) administration suppresses the behavioral pain response in this model.13 Similarly, l-NG-nitroarginine-p-nitroanilide has been found to inhibit the late phase of formalin-induced inflammatory pain and acetic acid-induced writhing in mice,14 whereas l-NAME and NG-monomethyl-l-arginine exhibited a dose-dependent suppression of intraplantar bradykinin-induced hyperalgesia in rats.15 Whereas all of these studies indicate involvement of NO in pain, it was later recognized that intraplantar formalin leads to an increase in the nNOS and iNOS in dorsal spinal neurons.16,17 There is evidence that iNOS, NOS-2 is constitutively expressed in the spinal cord which mediates inflammation-induced hyperalgesia18 Aminoguanidine, an inhibitor relatively selective for iNOS, has been found to exhibit antinociceptive properties in the second phase of formalin-induced pain.19
Given the convergent roles of these two proinflammatory systems, an important question concerns the possible nature of their interaction in the development of the hyperalgesic state observed after inflammation. Simultaneous treatment with NOS and COX-2 inhibitors produces an enhanced antinociceptive effect in formalin-induced pain in mice,20 but the nature of the interaction between inhibitors of NOS and COX-2 has yet to be determined. Therefore, we explored the nature of the interaction between iNOS and COX-2 using S-methylisothiourea (a moderately selective iNOS inhibitor), rofecoxib (selective COX-2 inhibitor), and mefenamic acid (nonselective COX inhibitor) as experimental tools in formalin-induced nociceptive behavior in mice.
METHODS
All experimental protocols were approved by the Institute Animal Ethics Committee.
Drugs
Rofecoxib is a diaryl substituted furanone compound with one heterocyclic ring and two benzene rings in its structure. It has more than 50-fold selectivity for the COX-2 enzyme isoform.7 S-methylisothiourea, an S-substituted isothiourea, shows selectivity towards iNOS over endothelial NOS.21,22 Mefenamic acid is an N-arylanthranilic acid that inhibits both COX isoforms in a nonspecific manner.
Rofecoxib, gifted by Ranbaxy Laboratories Ltd. India, S-methylisothiourea as its hemisulfate salt and mefenamic acid procured from Sigma Chemical Company (St. Louis, MO) were used in this study. S-methylisothiourea was dissolved in distilled water; mefenamic acid and rofecoxib in the form of aqueous suspension in 1% Tween-80 were given orally. All drugs were delivered in a volume of 10 mL/kg body wt.
Animals
Adult male Swiss albino mice weighing between 25 and 30 g were procured from the Laboratory Animal Resource Section of the institute. These mice were housed at controlled temperature (22 ± 2°C) and allowed free access to food and drinking water. The animals were acclimatized for a day before the experiment was conducted. The doses of different drugs used in the experiment were selected on the basis of results of a pilot study and are depicted in Table 1.
|
Formalin Model
One hour after the drug administration per os, 20 µL of 2.5% formalin solution prepared in normal saline was injected into the right hindpaw.23 After the intraplantar injection of formalin, mice were placed under a glass bell jar and cumulative time spent in paw licking was recorded during the late phase (15–30 min) after formalin injection.
Data Analysis
The effect of a particular dose of a drug or drug combination on pain perception was calculated using following formula:
|
|
where % efficacy measures the percent decline in paw licking response. Data were used to plot a least squares linear regression and the median effective doses (ED50) of the drugs administered were calculated.
The interaction of S-methylisothiourea with the COX inhibitors (rofecoxib and mefenamic acid) was evaluated by simultaneous administration of fixed proportions of S-methylisothiourea with each COX-inhibitor, and by performing an isobolographic analysis, as described by Tallarida et al.24 The isobologram was constructed by connecting the ED50 of the corresponding COX-inhibitor, plotted on the abscissa with the ED50 of S-methylisothiourea plotted on the ordinate to obtain the additive line. For each drug combination, ED50 and an associated 95% confidence interval were determined by linear regression analysis of the log-dose response data (six animals at each dose) and compared by t-test to the theoretical additive ED50. For interaction studies, fixed-ratio proportions were selected by first combining the ED50 of each compound and then constructing a dose-response curve in which ED50 fractions (1/2, 
, and 1/8) of iNOS-COX inhibitor combinations were administered. Theoretical additive ED50, ED25, and ED12.5 of the combinations were determined by combining 1/2 ED50 (e.g., 1/2 ED50 S-methylisothiourea + 1/2 ED50 COX-inhibitor), 1/2 ED25, and 1/2 ED12.5 of each drug and then following the linear regression analysis.
Variances of theoretical additive combinations were calculated using the following formula.25
|
|
where x is 50, 25, or 12.5.
From the individual variances, 95% confidence limits were calculated for theoretical additive combinations. When the drug combination gives an experimental ED50 statistically significantly lower than the theoretically calculated ED50, the interaction is said to be supra-additive or synergistic in nature. Statistical significance between theoretical and experimental median effective dose was assessed by using Student's t-test. The coincidence of a theoretical additive line and experimental combination line was assessed by F-test, as described elsewhere.24 Slopes of the regression lines were also calculated as per the standard statistical procedure.
RESULTS
Dose-effect data pertaining to the time spent in paw licking after intraplantar injection of formalin are summarized in Table 1. Slopes of the log dose-response regression lines and their 95% confidence intervals are presented in Table 2. For slope, overlapping confidence intervals indicated that all the dose-response curves do not differ from parallelism. Each drug or drug combination resulted in a significant decrease in the licking time during late phase of nociception. S-methylisothiourea-treated groups exhibited significantly reduced licking time during the late phase of nociception at the doses of 100 and 200 mg/kg, compared with that of vehicle-treated control mice. Groups of mice receiving rofecoxib (3, 10, and 30 mg/kg) also exhibited a significant and dose-dependent decrease in the late phase licking time compared with that of vehicle-treated control animals. Similarly, treatment with 30, 100, and 300 mg/kg of mefenamic acid also resulted in a significant and dose-dependent decrease in the late phase paw licking time, as compared with that of vehicle-treated control mice. There was a significant decline in the late phase licking time in the groups of mice receiving S-methylisothiourea + rofecoxib or mefenamic acid, compared with that of vehicle-treated control mice.
|
Based on the reduction in late phase paw licking, the effects produced by different drugs or drug combinations are expressed as "% Decline in licking response" in Figures 1a–e. The points of a theoretically derived additive line and experimentally derived combination line are represented in Table 3. ED50 values with 95% confidence intervals are presented in Table 4. The values of experimentally derived ED50 for S-methylisothiourea + rofecoxib or mefenamic acid drug combinations were significantly lower than their corresponding theoretical ED50, as evidenced by nonoverlapping confidence intervals. For either of the drug combinations, analysis of covariance revealed noncoincidence of a theoretical additive line and experimental combination line (P < 0.001). In isobolograms, Figs. 2a and b, the experimental ED50 points were well below the additive line, towards the left side indicating a synergistic type of interaction between the iNOS and COX inhibitors.
|
|
|
|
DISCUSSION
The present studies demonstrate a potent synergistic interaction between NOS and COX inhibition in a model of acute injury-induced hyperalgesia.
Formalin Model
The formalin test is a model of inflammatory pain and involves two distinct phases of pain-like behavior. The first phase, which has an immediate onset and short duration (
5 min), appears to reflect the nocifensive response to the direct activation of primary afferent fibers, whereas the second phase of prolonged duration that starts 10–15 min after injection of formalin is believed to reflect the nocifensive response to the continuing inflammation-evoked afferent activity, as well as the facilitatory processes in the spinal cord.26 There was a significant decrease in hindpaw licking response during the late phase that indicates involvement of prostaglandins (PGs) and/or NO in the processing and perception of formalin-induced pain. Similar observations have been made by some investigators working with models of formalin-induced pain.13,14,27 There is evidence that formalin injection causes increased expression of iNOS and nNOS in the spinal cord.17,18 Given the selectivity of rofecoxib, for COX-2 and the lack of difference in efficacy between a COX-2 selective and a COX nonselective agent, that inhibition of COX-2 is sufficient to produce the observed effects. With regard to NOS isoforms, S-methylisothiourea is modestly selective for iNOS versus nNOS, by a factor of two.22 Although this is more selective than the NG-substituted l-arginine analogues, it is not possible with this agent to exclusively argue for the role of either isoform. Nevertheless, these data are in accordance with results that have been reported for more selective iNOS inhibitors such as 1400 W.18,28
Mechanisms of NOS/COX Effects
A spinal site of the NO-mediated effects during the second phase of nociception has been suggested by electrophysiological studies. Firing from single, dorsal root horn neurons has been shown to be greatly reduced by NOS inhibitor treatment prior to intraplantar formalin. Thus, it is likely that NO-dependent sensitization may occur at the level of the spinal cord, where the release of excitatory amino acids leads to the generation of NO.29 In one study, formalin-induced pain was found to be associated with an increase in glutamate, NO and PGE2 in the dorsal horns of the spinal cord, suggesting that NO acts as a retrograde transmitter.30,31
It is believed that peripheral tissue injury, after formalin injection, provokes the release of excitatory neurotransmitters, such as glutamate and substance P, from primary afferents and dorsal horn neurons.32 The subsequent glutamate-induced activation of N-methyl- d-aspartate receptors leads to a Ca2+-influx and stimulates enzymatic activities that finally evoke the release of NO and PG which, in turn, increase the glutamate release,26 leading to increased sensitivity of dorsal horn neurons and, finally, central sensitization. During inflammatory pain states, PGE2 inhibits glycine receptor activity in the spinal cord by facilitating its phosphorylation via activation of protein kinase A.33 Because glycine receptors suppress neuronal firing, inhibiting them with PGE2 could facilitate transmission of pain signals to the brain. Thus, reduction of the paw-licking response associated with pretreatment with COX inhibitors should not be surprising.
NOS and COX Synergy
In the present study, the experimental ED50 of S-methylisothiourea plus a COX-inhibitor combination was significantly lower than the theoretical antinociceptive ED50 for the combination, indicating a synergistic nature of interaction between PGs and NO. The same was evident from the comparison of a theoretical additive line and the experimental mixture line or from isobolographic analysis. Although the present study did not explore the mechanism of interaction, the type of interplay between PGs/COX and NO/NOS can be deduced from the works of other investigators. It is presumed that co-production of NO and PGs associated with the activation of N-methyl-d-aspartate receptors during peripheral tissue injury is responsible for the central sensitization. Therefore, simultaneous inhibition of COX and NOS resulted in synergistic antinociceptive action. COX is a heme-containing enzyme, and heme has very high affinity for NO. It is also probable that NO produced by the activity of iNOS during the inflammation, stimulates COX-1 and/or COX-2 by combining with heme moiety or by nitrosylating the tyrosine residues of COX,33 thereby increasing the catalytic activity of COX manifold. Alternatively, the interaction of NO with COX might render the COX resistant to suicide inactivation, leading to enhanced PG production under the influence of NO.34
An earlier report from this laboratory involved study of iNOS and COX-2 inhibitors using variable doses of one in the presence of a fixed, but subthreshold dose of the other.20 Although this study revealed an enhanced effect of a subthreshold dose of one drug on the efficacy of another when used in combination, it necessitated the conducting of an isobolographic study with equieffective doses of iNOS and COX inhibitors. Induction of iNOS or COX-2 within a time period of 10–15 min after formalin injection is debatable. But, selective inhibitors of iNOS/COX-2 enzyme isoforms have prevented the perception of pain in formalin-induced nociception, indicating involvement of NOS (perhaps iNOS) and COX-2 in the facilitated state.11,19,20 Interestingly, although NOS inhibitors and nonsteroidal anti-inflammatory drugs have traditionally been viewed as providing pain relief primarily by decreasing signs of inflammation and sensitization of pain receptors peripherally, it appears that NOS and COX pathways are upregulated at the level of the spinal cord, which suggests that NOS and COX inhibitors may actually provide pain relief predominantly through this central effect.
Footnotes
Accepted for publication November 21, 2007.
REFERENCES
3: an essential target for spinal PGE2-mediated inflammatory pain sensitization. Science 2004;304:884–7
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
