Introduction

Top Abstract Introduction Methods Results Discussion References

It is generally accepted that NSAIDs exert their anti-inflammatory effect by inhibiting cyclooxygenase, thereby blocking the synthesis of prostaglandins.

However, there is considerable evidence that suppression of prostaglandin synthesis at the mucosal level is the cause for gastric toxicity, the main side effect of NSAIDs (Insel, 1990).

TNF is a cytokine that plays a key role in inflammation, as demonstrated by the clinical efficacy of anti-TNF antibodies in rheumatoid arthritis and Crohn's disease (Elliot et al., 1994; Stack et al., 1997). Paradoxically, NSAIDs apparently increase, rather than inhibit, TNF production. In fact, it has been reported that NSAIDs can directly induce TNF release in vitro and in vivo, this effect being apparently critical in the pathogenesis of NSAID-induced gastric injury (Appleyard et al., 1996; Santucci et al., 1994; Tsuboi et al., 1995).

Up-regulation of TNF production by NSAIDs is due to inhibition of the synthesis of PGE₂, a potent feedback inhibitor of TNF synthesis (Renz et al., 1988; Jorres et al., 1997; Kunkel et al., 1988; Tannenbaum and Hamilton, 1989; Sironi et al., 1992). In animal models of endotoxic shock, the enhancement of TNF production by NSAIDs was associated with increase in animal mortality (Pettipher and Wimberly, 1994).

Cytokines can evoke hyperalgesia, even though this effect is only partially related to PGE₂ induction. In fact, whereas IL-1-evoked hyperalgesia is significantly decreased by NSAIDs, the hyperalgesic effect of TNF, which represents the main component of carrageenan nociception, is only partially reduced by cyclooxygenase inhibitors (Cunha et al., 1991, 1992). On the other hand, pain induction by the chemokine IL-8 is apparently a prostaglandin-independent process (Cunha et al., 1991). It is therefore suggested that some inflammatory mediators could cause sensitization of nociceptors by PGE₂-independent mechanisms. Ketoprofen, a well known 2-arylpropionic acid NSAID, is a racemic mixture of two enantiomeric forms, R and S isomers, due to the presence of an asymmetric carbon atom in the  position to the carbonyl function. The R-enantiomer of ketoprofen is known to transform to the S-enantiomer in vivo in several animal species, except in humans and guinea pig (Brune et al., 1992; Abas and Meffin, 1987; Hutt and Caldwell, 1983). Although the anti-inflammatory role of the two enantiomers is not fully characterized, it is known that R-ketoprofen is a weak cyclooxygenase inhibitor, being ~100 to 1000 times less potent than the S-enantiomer in vitro (Brune et al., 1992; Williams, 1990; Adams et al., 1976), and it is therefore supposed to contribute only marginally to anti-inflammatory protection. In the case of another 2-arylpropionic acid NSAID, flurbiprofen, the R isomer, was shown to exert little effect on prostaglandin synthesis and no influence on inflammation. R-Flurbiprofen did, however, block nociception almost as potently as the S isomer (Brune et al., 1991, 1992). On the other hand, S-ketoprofen efficiently inhibits PGE₂ production and is considered the main effector of the anti-inflammatory activity (Mauleon et al., 1996). In clinical studies, gastric damage caused by ketoprofen could be mainly attributed to its S-enantiomer (Jerussi et al., 1998). It is therefore important for NSAIDs to define the relative contribution of the two enantiomers to the various activities (analgesia, anti-inflammatory capacity, induction of TNF). To this end, to allow a better clinical targeting, this study was aimed at comparing the action of R- and S-ketoprofen on inflammatory events in vivo and in vitro. The effects of R- and S-ketoprofen have been investigated in (1) a model of inflammation in the guinea pig (carrageenan-induced paw edema), (2) a model of inflammatory pain in the guinea pig (carrageenan-evoked hyperalgesia), (3) a model of TNF production in the guinea pig (LPS-induced serum TNF), and (4) a model in vitro of TNF production (LPS-induced TNF from mouse and guinea pig peritoneal macrophages). Studies were carried out in guinea pigs because similar to humans, no interconversion of R- to S-enantiomer occurs (Brune et al., 1992). TNF production in vitro was assessed not only in guinea pig macrophages but also in mouse peritoneal macrophages, the experimental system in which the regulation of TNF production has been previously studied (Lehmmann et al., 1988; Tannenbaum and Hamilton, 1989; Strassmann et al., 1994). Interconversion in vitro in mouse peritoneal macrophages is negligible because most of the interconversion of R-2-arylpropionic acids (associated to a stereoselective activation of the R-enantiomer to its CoA thioester by an acyl-CoA synthetase; Menzel-Soglowek et al., 1992) occurs in mitochondria and microsomes of liver tissue (Menzel-Soglowek et al., 1992; Cox et al., 1985; Knihinicki et al., 1989; Sanins et al., 1990; Muller et al., 1990; Knadler and Hall, 1990). The results reported hereafter show that the analgesic effect of ketoprofen is mainly associated to its R-enantiomer, whereas S-ketoprofen is the major inhibitor of edema formation. On the other hand, the anti-inflammatory action of S-ketoprofen was associated to the NSAID side effects, such as up-regulation of inflammatory cytokines.

	Methods

Top Abstract Introduction Methods Results Discussion References

Animals and drugs. Male Dunkin-Hartley guinea pigs were obtained from Harlan-Nossan, S. Pietro al Natisone, Italy. Male CD1 mice (for cytokine production in vitro) were obtained from Charles River Italia (Calco, Como, Italy). Animals were housed and acclimatized for 1 week under conditions of controlled temperature (20° ± 1°C), humidity (55 ± 10%) and lighting (7 AM to 7 PM); standard sterilized food and water were supplied ad libitum during acclimatization and experiments. All the procedures were performed in the animal facilities according to ethical guidelines for the conduct of animal research (Authorisation Italian Ministry of Health No. 271/95-B, D.Lvo 116/92; Italian Legislative Decree 116/92, Gazzetta Ufficiale della Repubblica Italiana No. 40, February 18, 1992; EEC Council Directive 86/609, OJ L 358, 1 December 12, 1987; NIH Guide for the Care and Use of Laboratory Animals, NIH Publication No. 85-23, 1985). Indomethacin (Sigma, St. Louis, MO) was dissolved in 1 M Tris base solution and diluted to the appropriate concentration in saline solution (0.9% NaCl).

Carrageenan-induced edema in the guinea pig paw. Guinea pigs, weighing 600 to 650 g, were randomized and assigned to the experimental groups. Test compounds (1 ml/kg) were administered subcutaneously 30 min before the subplantar injection into the right paw of 0.15 ml of 1% (w/v) carrageenan type IV (Sigma) in saline at 37°C (Winter et al., 1963). Control groups received DL-lysine/saline or Tris/saline solutions at the appropriate dilution (vehicle). Evaluation of the paw volume was performed 1 hr before (basal) and 3 hr after carrageenan injection (final), using a hydropletismometer (Basile, Comerio, Italy). Paw swelling was calculated as the difference between basal and final measures, and the anti-inflammatory activity of test compounds was evaluated as percent inhibition of carrageenan-induced edema formation.

Carrageenan-induced hyperalgesia in the guinea pig paw. Guinea pigs were treated as described for edema formation. The pain threshold (Randall and Selitto, 1957) was measured in the same paw, immediately after the hydropletismometric measures (1 hr before and 3 hr after carrageenan), using a Randall-Selitto analgesy-meter (Basile, Comerio, Italy).

LPS-induced cytokine production in the guinea pig. Animals received an i.p. inoculum of bacterial endotoxin (LPS from E. coli 055:B5, Sigma; 40 µg/animal) 30 min after treatment with ketoprofen molecules (390 µmol/kg s.c.) and were bled 90 min later. Control groups received DL-lysine/saline solution s.c. and saline i.p. TNF production was determined in the serum as described below.

Macrophage preparation and LPS-induced PGE₂ and cytokine production. Peritoneal exudate cells were collected from peritoneal washings of mice or guinea pigs 5 days after i.p. inoculum of 3% thioglycolate (Difco, Detroit, MI) in saline (1.5 ml in mice and 15 ml in guinea pigs, respectively). Cells were plated at 3 × 10⁶/well in 24-well plates (Costar, Cambridge, MA), and nonadherent cells were removed by repeated washing 2 hr later. Ketoprofen enantiomers were then added to adherent macrophages 30 min before the addition of LPS (1 µg/ml). Control wells received DL-lysine/saline solution at the appropriate dilution (vehicle). Culture supernatants were harvested 4 hr after LPS stimulation for TNF determination and after 24 hr for IL-1 measurement. Total PGE₂ production (supernatants plus cell lysates) was determined 4 hr after LPS stimulation. Cell lysates were obtained by three cycles of freeze-thawing. Cell viability was >95% in all experiments, as measured by trypan blue dye exclusion.

	Results

Top Abstract Introduction Methods Results Discussion References

Inhibition of carrageenan-induced edema and hyperalgesia in guinea pig. The effect of ketoprofen enantiomers on carrageenan-induced edema in the guinea pig paw is shown in figure 1. Carrageenan induced paw swelling of 1.2 ± 0.1 ml (mean ± S.E. of 10 animals), whereas no detectable variation of paw volume was detected in the control group (pretreated with vehicle). S-Ketoprofen, administered subcutaneously 30 min before subplantar injection of carrageenan, significantly inhibited edema formation already at 75 µmol/kg and reached maximal inhibition at 250 µmol/kg (57% and 67% inhibition, respectively). The R-enantiomer showed a much reduced effect, which became significant only at 250 µmol/kg. The racemate appeared to have an intermediate effect between R and S isomers.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Inhibitory effect of R- and S-ketoprofen on carrageenan-induced edema formation. Guinea pigs were received a subplantar injection into the right paw of 0.15 ml of 1% (w/v) carrageenan type IV. The test compounds (R-ketoprofen, ; S-ketoprofen, ; ketoprofen racemate, ; indomethacin, ) were administered s.c. 30 min before carrageenan. Evaluation of the paw volume was performed 1 hr before and 3 hr after carrageenan injection. Percentage of edema inhibition was calculated as described in Methods. Data are expressed as percent of edema inhibition and represent the mean ± S.E. of 10 animals per group within one experiment representative of three performed. Statistical analysis was performed by ANOVA for random design (SAS/STAT 6.12). Multiple comparisons were performed using Dunnett's t test. Significance threshold was set at P < .05. * P < .05 vs. carrageenan alone. # P < .05 vs. R-ketoprofen.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Analgesic effect of R- and S-ketoprofen in carrageenan-evoked hyperalgesia. Guinea pigs received a subplantar injection into the right paw of 0.15 ml of 1% (w/v) carrageenan type IV. The test compounds (R-ketoprofen, ; S-ketoprofen, ; ketoprofen racemate, ; indomethacin, ) were administered s.c. 30 min before carrageenan. The pain threshold was measured in the same paw 1 hr before and 3 hr after carrageenan treatment. Pain index was determined as described in Methods. Data are expressed as percent variation of pain index and represent the mean ± S.E. of 10 animals per group within one experiment representative of three performed. Statistical analysis was performed by ANOVA for random design (SAS/STAT 6.12). Multiple comparisons were performed using Dunnett's t test. Significance threshold was set at P < .05. * P < .05 vs. carrageenan alone. # P < .05 vs. R-ketoprofen.

Effects on LPS-induced cytokine production. The effect of ketoprofen on LPS-induced TNF production was assessed in a therapeutical (0.1-10 µM; Insel, 1990) as well as lower concentration range (1 and 10 nM). As reported in figure 3A, S-ketoprofen induced a marked amplification of TNF production in LPS-stimulated mouse macrophages starting at 10 nM. On the contrary, no significant upregulation of TNF production was induced by R-ketoprofen. Ketoprofen racemate at a concentration of 10 µM induced a significant enhancement of LPS-induced TNF production, comparable to that induced by S-ketoprofen. Similarly, S-ketoprofen could also enhance LPS-induced IL-1 production, whereas R-ketoprofen did not have any significant effect on IL-1 release (fig. 3B). In the same assay, IL-1 production was enhanced by ketoprofen racemate at a concentration of 10 µM. In the absence of LPS stimulation, ketoprofen isomers alone were unable to induce TNF or IL-1 production over a wide range of concentrations (data not shown). The inhibitory effect of R- and S-ketoprofen on PGE₂ production by macrophages was also investigated. As reported in figure 3C, pretreatment of mouse macrophage with as little as 1 nM S-ketoprofen could significantly reduce LPS-induced cyclooxygenase activity (as measured by inhibition of PGE₂ production), this reduction being almost complete at 10 nM. On the other hand, R-ketoprofen was >100-fold less effective, as it could achieve complete inhibition of PGE₂ production only at 10 µM. Inhibition of PGE₂ production by ketoprofen racemate was somehow intermediate to that of R and S isomers.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Effect of R- and S-ketoprofen on TNF, IL-1 and PGE2 production in LPS-stimulated mouse peritoneal macrophages. Adherence-purified mouse peritoneal macrophages were exposed to ketoprofen molecules (R, ; S, ; or racemate, ) for 30 min before addition of 1 µg/ml LPS (white column). TNF (A) was measured in cell supernatants 4 hr after LPS stimulation, whereas IL-1 (B) was measured 24 hr after LPS stimulation, as described in Methods. TNF was not detectable in the control group, whereas spontaneous production of IL-1 in unstimulated macrophages was 4.6 ± 0.3 pg/ml. Total PGE2 production (C) was measured 4 hr after LPS stimulation, as described in Methods. Spontaneous PGE2 production was 20 ± 4 pg/well. Data are mean ± S.E. from five replicate wells within one experiment representative of three performed. Statistical analysis was performed by Student's t test and Mann-Whitney U test. Significance threshold was set at P < .05. * P < .05, ** P < .01 vs. LPS alone by Student's t test and Mann-Whitney U test.

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Effect of R- and S-ketoprofen on TNF and PGE2 production in LPS-stimulated guinea pig peritoneal macrophages. Adherence-purified guinea pig peritoneal macrophages were exposed to ketoprofen enantiomers (R, ; and S, ) for 30 min before addition of 1 µg/ml LPS (white column). TNF (A) and total PGE2 production (B) were measured 4 hr after LPS stimulation. No TNF was detectable in the control group, whereas spontaneous PGE2 production was 0.92 ± 0.3 pg/well. Data are mean ± S.E. from five replicate wells within one experiment representative of three performed. ** P < .01 vs. LPS alone by Student's t test and Mann-Whitney U test.

	Discussion

Top Abstract Introduction Methods Results Discussion References

Inhibition of cyclooxygenase activity by NSAIDs is at the basis of their anti-inflammatory action. However, suppression of prostaglandin synthesis is also the mechanism underlying NSAID unwanted effects, such as gastric damage. Among NSAIDs, 2-arylpropionic acid derivatives (e.g., ibuprofen and ketoprofen) are a racemic mixture of the two enantiomeric forms, with S-enantiomer being mainly responsible for inhibition of prostaglandin synthesis and of inflammatory events (Brune et al., 1992; Williams, 1990; Adams et al., 1976; Mauleon et al., 1996; Otterness and Bliven, 1985; McCormack and Urquhart, 1995). This study reports that in contrast with current thinking (Mauleon et al., 1996), R-ketoprofen accounts for most of the analgesic activity of ketoprofen and has significant, although low, anti-inflammatory activity in vivo. The analgesic and anti-inflammatory effects of R-ketoprofen are not associated with enhancement of cytokine production either in vitro or in vivo, in contrast to what could be observed for the S isomer.

Carrageenan-induced edema is a standard prostaglandin-dependent model of inflammation, in which S-ketoprofen shows very good inhibitory action. Data reported here are indeed in agreement with previous reports indicating a major role in inhibition of paw inflammation for S isomers of 2-arylpropionic acid (Brune et al., 1991). The anti-inflammatory effect of these compounds usually correlates with their potency as peripheral prostaglandin inhibitors (Vane, 1971; Mauleon et al., 1996). In agreement with this hypothesis, the weak cyclooxygenase inhibitor R-ketoprofen could reduce paw swelling only at high doses. On the other hand, the main part of the analgesic activity of ketoprofen may be attributed to its R isomer. Within the past decade, conflicting results have been reported on the mode of action of NSAIDs in pain. Indeed, the analgesic effectiveness of many NSAIDs currently in use does not correlate with their potency as prostaglandin synthesis inhibitors (Brune et al., 1981; Lanz et al., 1986; McCormack and Brune, 1991). Thus, despite their different effectiveness on cyclooxygenase, salicylic acid shows analgesic effect in a concentration range comparable to that of aspirin (Graham et al., 1977) and protection against carrageenan-evoked hyperalgesia was reported for R-flurbiprofen (Brune et al., 1991). Recently, it has been proposed that the inflammatory cytokine TNF may play a crucial role in carrageenan-induced pain. The hyperalgesic action of TNF was apparently independent of prostaglandin production, further stressing the involvement of a cyclooxygenase-independent pain mechanism in carrageenan-evoked hyperalgesia (Cunha et al., 1991, 1992).

As very recently reported in a clinical study, one of the main side effects of ketoprofen, gastric damage, could be attributed to its S isomer (Jerussi et al., 1998). Although gastric damage by NSAIDs is generally associated to prostaglandin inhibition, several recent reports are consistent with the hypothesis that TNF-mediated leukocyte activation could play an important role in experimental NSAID-induced gastropathy (Santucci et al., 1994; Appleyard et al., 1996; Tsuboi et al., 1995). NSAIDs, including 2-arylpropionic acid derivatives, up-regulate TNF production by blocking the synthesis of PGE₂, which is a strong inhibitor of cytokine production (Jorres et al., 1997; Kunkel et al., 1988; Tannenbaum and Hamilton, 1989). This phenomenon is not restricted to experimental models, as it could also be observed in mononuclear cells from healthy volunteers after NSAID medication (Endres et al., 1996). Data reported here indicate that S-ketoprofen is in fact able to amplify TNF production, whereas the R-enantiomer is inactive. Because down-regulation of TNF production by PGE₂ has been studied in murine peritoneal macrophages (Lehmmann et al., 1988; Tannenbaum and Hamilton, 1989; Strassmann et al., 1994), the effect of ketoprofen enantiomers was first assessed in this experimental system. Interconversion from R- to S-ketoprofen in mouse peritoneal macrophages in vitro should be negligible because it usually occurs in mitochondria and microsomes of liver tissue (Menzel-Soglowek et al., 1992; Cox et al., 1985; Knihinicki et al., 1989; Sanins et al., 1990; Muller et al., 1990; Knadler and Hall, 1990). The cytokine-enhancing effect of S-ketoprofen paralleled its inhibitory activity on PGE₂ production. On the other hand, the R-enantiomer was unable to amplify cytokine production even at cyclooxygenase-blocking concentrations. The same difference between R- and S-ketoprofen on enhancement of TNF production and inhibition of PGE₂ synthesis was confirmed with guinea pig peritoneal macrophages in vitro and in vivo in guinea pigs. In addition to its cyclooxygenase inhibitory effect, the amplification of TNF production by S-ketoprofen could be thus correlated with its low gastric tolerability (Appleyard et al., 1996; Jerussi et al., 1998) and reduced analgesic effect (Cunha et al., 1991; 1992).

In summary, on investigating the effect of ketoprofen enantiomers on several inflammatory parameters, S-ketoprofen proved very efficient in inhibiting carrageenan-induced edema formation, but it had less potent analgesic effect than R-ketoprofen. The fact that S-ketoprofen, unlike the R isomer, could amplify LPS-induced cytokine production could in fact contribute to its reduced analgesic effect and could be correlated to the clinical evidence of its higher gastric toxicity. Taken together, these data suggest that beside the clear-cut anti-inflammatory effects of S-ketoprofen, the R-enantiomer also contributes to the overall activity of ketoprofen, because it is analgesic and does not amplify inflammatory cytokine production.