Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

PGE₂ is a major COX product at inflammatory sites where it acts as a potent relaxant of vascular smooth muscle cells, increasing blood flow and therefore potentiating edema formation by agents such as bradykinin and histamine (Williams and Peck, 1977). Apart from promoting the development of the characteristic inflammatory signs of vasodilatation and erythema (Solomon et al., 1968), PGE₂ also promotes inflammatory pain by sensitizing afferent nerve endings to the actions of bradykinin and histamine (Ferreira, 1972). Being the prostanoid most generally associated with inflammatory responses, the formation of PGE₂ at inflammatory sites is often taken as an indicator of local COX activity. The ability of nonsteroid anti-inflammatory drugs (NSAIDs) to reduce this formation is therefore often taken as an indicator of their ability to reduce the production of pro-inflammatory prostanoids. Recently, the production of pro-inflammatory prostanoids has been linked to the activities of inducible COX-2. Because of this, the local production of PGE₂ is often taken as an indicator of COX-2 activity and its inhibition as an index of the ability of NSAIDs or newer COX-2-selective agents to inhibit COX-2. For example, the rat carrageenan-induced paw edema and the rat carrageenan-air pouch are two of the most widely used models in which to characterize the selectivities of NSAIDs toward the inhibition of COX-2. In these models, edema formation, leukocyte infiltration, and prostaglandin levels (mainly PGE₂ but also 6-keto-PGF₁) in inflammatory exudates are measured as indicators of the inflammatory process that follows the injection of carrageenan. The assumption that the PGE₂ recovered from the air pouch or from the paw is solely derived from COX-2 (Seibert et al., 1994) and therefore can be taken as an index of COX-2 activity, however, has not been fully tested. In this study, we have examined the abilities of NSAIDs and a highly-selective COX-2 inhibitor, 5,5-dimethyl-3-(2-propoxy)-4-methanesulfonylphenyl)-2(5H)-furanone (DFP) (Leblanc et al., 1999) to inhibit PGE₂ production in vivo and in vitro in the rat. These data have been compared with that generated in human blood in vitro (Warner et al., 1999; this study). Some of this data has been presented to the British Pharmacological Society (Giuliano and Warner, 2001).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Materials

All compounds used were obtained from Sigma-Aldrich (Poole, UK) unless otherwise stated. Antibodies for immunoblotting and DFP (Leblanc et al., 1999) were a gift from Merck Frosst (Quebec, Canada). For the radioimmunoassays, antiserum to PGE₂ was obtained from Sigma-Aldrich; [³H]PGE₂ was purchased from Amersham Biosciences UK, Ltd. (Little Chalfont, UK).

Surgical Procedure in Rats Male Wistar rats were obtained from Tuck (Rayleigh, UK) and kept according to the guidelines set by the Home Office Code of Practice for the Housing and Care of Animals used in Scientific Procedures (1989). Rats (220-250 g) were anesthetized with thiobutabarbital sodium (Inactin; 120 mg kg¹, i.p.). Body temperature was maintained at 37°C by means of a homeothermic blanket connected to a rectal probe. The trachea was cannulated (1.67 × 2.42-mm tubing) to facilitate ventilation. The right carotid artery was cannulated (0.58 × 0.96-mm tubing) and connected to a pressure transducer for the monitoring of systemic blood pressure, which was displayed on a computer linked to a digital data acquisition system (PowerLab 8/s; ADInstruments, Hastings, UK). The jugular vein was also cannulated (0.40 × 0.80-mm tubing) to allow injection of drugs and/or infusion of saline as necessary. At the end of each experiment, animals were killed by an overdose of anesthetic.

In Vivo Experimental Design

Upon completion of the surgical procedure, animals in the control and bacterial lipopolysaccharide (LPS) groups were injected (t = 0 h) with saline (2 ml kg¹, i.p.) or Escherichia coli LPS (serotype 0127:B8; activity 3,000,000 endotoxin units mg¹; 6 mg kg¹, i.p.), respectively. Rats in the LPS group were injected i.p. at time t = 4 h with either vehicle (10% dimethyl sulfoxide in saline), diclofenac (3 mg kg¹), or the selective COX-2 inhibitor DFP (10 mg kg¹). Six hours after the LPS was injected, animals were challenged with arachidonic acid (3 mg kg¹), which was given as a bolus via the jugular vein. Blood samples (300 µl) were taken via the carotid artery cannula 1 min after administration of the arachidonic acid bolus. The samples were centrifuged at 12,000g for 3 min (4°C), and the plasma was removed, supplemented with heparin (15 IU ml) and stored at 20°C until further analysis.

Withdrawal of Blood

Rats. Male Wistar rats (250-280 g) were injected with a lethal dose of pentobarbitone sodium (Sagittal; 120 mg kg¹ i.p.). Upon occurrence of deep anesthesia, a laparotomy was performed, and blood from the abdominal aorta was collected into a tube containing 30 IU ml¹ heparin using a plastic syringe connected to an 18-gauge needle. Animals were then killed by thoracotomy.

Humans. Blood from healthy volunteers (25-55 years) who had not taken NSAIDs for at least 2 weeks was withdrawn from the antecubital vein using a 16-gauge butterfly needle and collected in a plastic tube containing heparin (20 IU ml¹, final).

Isolation of Rat Leukocytes

Rat peripheral leukocytes were isolated from heparinized whole blood by dextran sedimentation of erythrocytes. In detail, 0.8 parts of 3.5% dextran (mol.wt. 250,000) in saline were added to 1 part of blood and incubated for 45 min at 37°C. Following sedimentation of the red cells, the upper leukocyte-containing layer was removed and deprived of contaminating erythrocytes by hypotonic lysis. Cells were then washed by centrifugation in phosphate-buffered saline and resuspended in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, penicillin (50 IU ml¹), and streptomycin (50 µg ml¹), which was used as described below.

In Vitro Experimental Protocols

Protocol 1. Freshly taken rat blood was dispensed (100 µl/well) into two 96-well plates and treated with either vehicle (saline) or 1, 10, or 100 µg ml¹ LPS as appropriate. Wells receiving LPS were also treated with either vehicle (0.1% DMSO in culture medium), dexamethasone (1 µM, 1 h prior to LPS), aspirin (100 µM), or the COX-2 selective drug DFP (10 µM). Plates were then incubated in a humidified atmosphere of 5% CO₂/95% air at 37°C. Eighteen hours later, one plate received 50 µM Ca²⁺ ionophore (A23187), while the other was treated with vehicle (1% DMSO in culture medium). Fifteen minutes later, plates were centrifuged (1500g, 4°C, 5 min), and the plasma obtained was stored (40°C) until measurement of PGE₂ by a radioimmunoassay.

Protocol 2. Leukocytes were isolated from rat blood as described above. The cells obtained were resuspended at a density of 8 × 10⁶ ml¹ in culture medium plus 10% fetal bovine serum and plated into 96-well plates (100 µl/well). These were then treated as described for whole blood under Protocol 1.

Protocol 3. Rat platelet-rich plasma was obtained by centrifuging freshly collected heparinized (~30 IU ml¹) blood at 200g for 7 min. Rat platelet-rich plasma was aliquoted into 96-well plates (100 µl/well) and treated with either vehicle (0.1% DMSO in culture medium), aspirin (100 µM), or DFP (10 µM). Following incubation for 30 min (37°C, 5% CO₂/95% air), the Ca²⁺ ionophore A23187 (50 µM) was added, and rat platelet-rich plasma was incubated for further 15 min. The plates were then centrifuged (as above), and the supernatant was stored until measurement of PGE₂ by a radioimmunoassay. The treatment just described was also carried out in parallel on whole blood and washed leukocytes originating from the same batch of blood as that used to obtain platelet-rich plasma (PRP).

Protocol 4. Freshly taken human and rat blood was aliquoted in plastic tubes (100 µl/tube) and then treated with the Ca²⁺ ionophore A23187 (50 µM). Fifteen minutes later the tubes were centrifuged (1500g, 4°C, 5 min), and the plasma obtained was stored (40°C) until measurement of PGE₂ by a radioimmunoassay.

Western Blot Analysis

Rat leukocytes from a single animal (5 × 10⁷-7 × 10⁷) were divided into three aliquots. One aliquot (basal) was immediately processed to be later used for immunoblotting. The remaining two aliquots were seeded onto individual 35-mm Petri dishes and incubated (37°C, 5% CO₂/95% air), one in the absence and one in the presence of LPS (100 µg ml¹). Eighteen hours later cell extracts were prepared (lysis buffer: 50 mM tris-HCl, 10 mM EDTA, 1% (v/v) Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 0.05 mM pepstatin A, and 0.2 mM leupeptin in ddH₂O) and later used for gel electrophoresis and Western blot analysis. The protein concentrations of cell lysates were determined by the Bradford colorimetric assay (Bradford, 1976). Equal amounts of protein (20 µg) were loaded onto 10% SDS-polyacrylamide gels and subjected to electrophoresis for 1 h at 100 V. The proteins were then electro-transferred to nitrocellulose (Hybond-C; Amersham Biosciences UK, Ltd.) at 80 V for 1 h. Following electro-transfer, the blots were incubated overnight at 4°C in blocking solution (5% w/v dried low-fat milk and 0.1% v/v Tween 20 in phosphate-buffered saline) on an orbital shaker. The blots were then washed (3 × 5 min) with washing buffer (Tween 20 0.1% v/v in phosphate-buffered saline) before being probed (1 h at room temperature) with anti-COX-1 (1:3000) or anti-COX-2 antibody (1:5000) diluted in blocking solution. Following incubation with the primary antibody, the blots were washed (3 × 5 min) with blocking solution before being probed (1 h room temperature) with the horseradish peroxidase-conjugate secondary antibody (supplied as part of the Phototope Kit, see below) diluted 1:2000 in blocking solution. The blots were then developed using Phototope-HRP Western blot detection kit (New England Biolabs, Hitchin, UK). Images were captured on Hyperfilm (New England Biolabs) and acquired by a Macintosh computer connected to a densitometer. Densitometric analyses were by Molecular Analyst (Bio-Rad, Hemel Hempstead, UK).

Data Analysis

Data are expressed as mean ± S.E.M. of separate determinations as specified in individual figure legends. Unless otherwise stated, statistical analyses were performed by applying a one-way analysis of the variance followed by Dunnett's post-test. A P value smaller than 0.05 indicated a statistical difference.

For Figs. 3 and 5, PGE₂ formation was calculated as the difference between the PGE₂ measured after Ca²⁺ ionophore treatment and the PGE₂ accumulated in the same samples over 18 h. All graphs and analyses were by Prism 3.0 (GraphPad Software, San Diego, CA).

Terminology

In the writing of this article, we have used the word "accumulation" to signify the release of mediators in the absence of stimuli such as arachidonic acid or Ca²⁺ ionophore. Conversely, the expressions "stimulated release", "formation", or "acute formation" refer to the release of mediators following the application to the system of arachidonic acid or Ca²⁺ ionophore.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

In Vivo. Treatment of rats with LPS caused a marked increase in the levels of PGE₂ that followed injection of arachidonic acid (3 mg kg¹) (control, 4.5 ± 1.4 ng ml¹, n = 6; LPS-treated, 18.4 ± 3.7 ng ml¹, n = 7). Diclofenac given to LPS-treated rats inhibited the release of PGE₂ back to approximately control levels (6.1 ± 1.5 ng ml¹, n = 5, P < 0.05). The selective COX-2 inhibitor DFP produced a smaller insignificant reduction in PGE₂ formation in these LPS-treated rats (12.7 ± 1.6 ng ml¹, n = 5, P > 0.05).

In Vitro: Accumulation and Acute Formation of PGE₂ in LPS-Treated Blood (Protocol 1). Basal accumulation of PGE₂ in rat blood under control conditions was 0.8 ± 0.1 ng ml¹ over the 18-h incubation period. LPS caused concentration-dependent increases in the accumulation of PGE₂ (Fig. 1) that were inhibited by dexamethasone, aspirin, and DFP (P < 0.05; Fig. 1). Indeed, in blood incubated with 1 µg ml¹ LPS, aspirin inhibited PGE₂ accumulation to below-detection levels.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Accumulation of PGE2 in rat blood incubated with LPS for 18 h. LPS caused concentration-dependent accumulations of PGE2 above control levels (dashed line) in the presence of drug vehicle (unfilled bars). This increase was inhibited by dexamethasone (filled bars), aspirin (stippled bars), or DFP (hatched bars). , P < 0.05 (control versus treatment); #, P < 0.05 (drugs versus matched vehicle control). Data represent mean ± S.E.M. of six separate determinations.

1

View larger version (33K): [in this window] [in a new window]   Fig. 2.   The effects of vehicle (unfilled bars), dexamethasone (filled bars), aspirin (stippled), or DFP (hatched bars) on the production of PGE2 stimulated by A23187 in LPS-treated rat whole blood. The dashed line (control) represents the formation of PGE2 in the absence of LPS. , P < 0.05 (control versus treatment); #, P < 0.05 (drugs versus matched vehicle control). Data represent mean ± S.E.M. of eight separate determinations from four different experimental days.

In Vitro: Accumulation and Acute Formation of PGE₂ in LPS-Treated Leukocytes (Protocol 2). Incubation of rat washed leukocytes with LPS (1, 10, and 100 µg ml¹) caused a concentration-dependent accumulation of PGE₂ over the 18-h incubation period (0.7 ± 0.3, 2.9 ± 0.7, 6.3 ± 2.0 ng ml¹, respectively; n = 8 for all). In control samples and in blood treated with LPS and test drugs, however, PGE₂ was below detection levels.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   The effects of vehicle (unfilled bars), dexamethasone (filled bars), aspirin (stippled bars), or DFP (hatched bars) on the release of PGE2 by A23187 stimulation of LPS-treated rat washed leukocytes. The dashed line (control) represents the formation of PGE2 in the absence of LPS. , P < 0.05 (control versus treatment); #, P < 0.05 (drugs versus matched vehicle control). Data are expressed as mean ± S.E.M. of eight determinations from four different experiments.

In Vivo: Production of PGE₂ by Rat Whole-Blood, Platelet-Rich Plasma and Washed Leukocytes (Protocol 3). The production of PGE₂ by freshly isolated rat washed leukocytes following exposure to A23187 (50 µM, 15 min) was below detection levels. Conversely, following A23187 treatment, PGE₂ levels rose considerably and to similar extents (P > 0.05, t test) in rat whole blood and rat PRP (Fig. 4). This A23187-stimulated formation was strongly inhibited by aspirin (P < 0.05) but was unaffected by DFP (P > 0.05, vehicle versus drugs).

View larger version (23K): [in this window] [in a new window]   Fig. 4.   The effects of vehicle (unfilled bars), aspirin (stippled bars), or DFP (hatched bars) on the stimulated production of PGE2 by rat whole blood and PRP. , P < 0.01; , P < 0.05 (vehicle versus drugs). Data represent mean ± S.E.M. of eight determinations from four separate experiments.

In Vitro: Stimulated Release of PGE₂ in Rat and Human Whole Blood (Protocol 4). Under control conditions, the amount of PGE₂ produced by rat whole blood in response to the addition of A23187 (50 µM for 15 min) was approximately 20 times greater than that produced by human blood (rat, 24 ± 1.9 ng ml¹; human, 1.2 ± 0.16 ng ml¹; n = 9 for both).

In Vitro: Expression of COX-1 and COX-2 in Rat Leukocytes. Western blot analysis of cell extracts demonstrated that freshly isolated leukocytes expressed low levels of COX-1 (Fig. 5) but contained no detectable COX-2 (Fig. 6). Maintenance of the cells in culture conditions for 18 h resulted in a 3-fold increase in COX-1 expression that was not affected by the presence of LPS (Fig. 5). Conversely, COX-2 expression in isolated leukocytes was not affected by exposure to culture conditions for 18 h but was greatly increased by incubation with LPS (Fig. 6).

View larger version (39K): [in this window] [in a new window]   Fig. 5.   A, Western blot showing the expression of COX-1 in leukocytes isolated from rat blood. Lanes: 1, COX-1 standard (~25 ng); 2, COX-2 standard (~25 ng); 3, freshly isolated cells; 4, cells incubated for 18 h with LPS vehicle; 5, cells incubated for 18 h with LPS. B, densitometric analysis of displayed gel and two further experiments. , P < 0.01 (control versus treatment).

View larger version (31K): [in this window] [in a new window]   Fig. 6.   A, Western blot showing the expression of COX-2 in leukocytes isolated from rat blood. Lanes: 1, COX-1 standard (~25 ng); 2, COX-2 standard (~25 ng); 3, freshly isolated cells; 4, cells incubated for 18 h with LPS vehicle; 5, cells incubated for 18 h with LPS. B, densitometric analysis of displayed gel and two further experiments. , P < 0.01 (control versus treatment).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In this study, we show that the increase in the arachidonic acid-stimulated production of PGE₂ seen in LPS-treated rats is coupled to both COX-1 and COX-2. A similar relationship was also found in rat whole blood in vitro, although not in human whole blood in vitro (Warner et al., 1999). Given the important role played by PGE₂ in inflammation, these observations are consistent with the concept that COX-1 products contribute to inflammatory responses in rats and possibly other species, as reported by Wallace et al. (1998).

We have previously shown that LPS administered to the anesthetized rat induces both the up-regulation of COX-2 protein and an increase in the plasma levels of 6-keto-PGF₁ (Hamilton et al., 1999; Giuliano et al., 2001). In these previous experiments, challenge with arachidonic acid revealed that COX-2 induction was associated with a much greater increase in the synthetic capacity for 6-keto-PGF₁ than suggested by simple increases in plasma prostanoid levels. Similarly, in the present study, the stimulated release of PGE₂ was increased in LPS-treated rats compared with controls. Diclofenac given to LPS-treated rats was also found to inhibit the stimulated release of PGE₂. Surprisingly and in contrast to what was observed for 6-keto-PGF_1a (Hamilton et al., 1999; Giuliano et al., 2001), however, the selective COX-2 inhibitor DFP only produced a partial, insignificant reduction in PGE₂ production. This prompted us to investigate further because such an observation was consistent with the activities of both COX-1 and COX-2, underlying the elevated production of PGE₂.

Our in vitro data also support the contention that PGE₂ production in rat blood is driven by the activities of both COX-1 and COX-2. In particular, incubation of rat blood with LPS led to an increase in PGE₂ production that was sensitive to aspirin, dexamethasone, and the COX-2 selective inhibitor DFP. Importantly, the effectiveness of dexamethasone and DFP strongly suggested that the accumulation of PGE₂ observed in LPS-treated blood was brought about by COX-2 activity. However, in the same experiments, challenge of the system with the Ca²⁺ ionophore A23187 revealed a capacity for the formation of PGE₂ that was unaffected by LPS treatment and, most importantly, only partially sensitive to dexamethasone or DFP. Indeed, these findings imply that COX-1 activity largely underlies the acute formation of PGE₂ with or without the presence of COX-2 (i.e., with or without exposure to LPS). Further analyses showed that LPS caused a concentration-dependent accumulation of PGE₂ in incubates of rat washed leukocytes. Importantly, the production of PGE₂ by LPS-stimulated leukocytes and the effects of the inhibitors on this production closely resembled those observed in similarly treated rat whole blood. Therefore, the induction and activity of COX-2 observed in isolated leukocytes exposed to LPS may well account for the accumulation of PGE₂ measured in LPS-treated whole blood from rats.

Surprisingly, rat washed leukocytes appeared to be only partially capable of contributing the levels of PGE₂ measured in whole blood following Ca²⁺ ionophore treatment. In particular, the amounts of PGE₂ produced by control and LPS-treated leukocytes challenged with A23187 were about one-third of those observed in similarly treated whole blood. Most importantly and in contrast to what we observed in whole blood, dexamethasone inhibited the acute formation of PGE₂ in LPS-treated washed leukocytes. Although the effect of dexamethasone was not further investigated, one might speculate that in LPS-treated leukocytes dexamethasone may have curtailed PGE₂ production by down-regulating not only COX-2 but also COX-1 or, alternatively, by reducing the activity of phospholipase A₂. Notably, as shown by Western blot analyses, the expression of COX-1 protein was increased in leukocytes maintained in culture conditions for 18 h. Moreover, COX-1 expression has previously been reported to be sensitive to dexamethasone (Hamasaki et al., 1993; Jun et al., 1999).

Regardless of the mechanism involved, dexamethasone did not produce as great an inhibition of PGE₂ production in whole blood as it did in LPS-stimulated leukocytes. This indicated that a source other than the leukocytes was responsible for the stimulated formation of PGE₂ observed in whole blood. Indeed, a comparative analysis of rat platelet-rich plasma and washed leukocytes indicated that, in whole blood, platelets represent the main source of COX-1-derived PGE₂.

In summary, the experiments described here show that when exposed to LPS, rat whole blood produces increased levels of PGE₂ through induction of COX-2 in blood leukocytes. Although this model is not strictly one of inflammation, the acute application of LPS does induce the rapid expression of COX-2 and therefore provides an in vivo experimental system in which to study the interplays between COX-1 and COX-2. In addition, because in vitro assays employing LPS-treated whole blood as a source of COX-2 have been widely used to characterize the pharmacological activities of NSAIDs (Warner et al., 1999; Fitzgerald and Patrono, 2001), LPS treatment in vivo seems a further logical proving ground. Another caveat to keep in mind is that the rate of prostanoid formation following arachidonic acid challenge in vivo most likely is greatly in excess of that seen in inflammation. The application of excess substrate, however, does provide us once again with a system more aligned with those used to test NSAID activities in vitro (e.g., strong stimuli are applied to activate platelets and therefore drive COX-1 activity) (Warner et al., 1999; Fitzgerald and Patrono, 2001). At the same time, we must remember that in COX-2 overexpressing cells, for instance, prostanoid production may not be limited by arachidonic acid availability, and so our application of excess substrate could be biased toward COX-1 products. Taken together, of course, these considerations remind us that in this study we have measured COX-1 and COX-2 products in an acute in vivo system that does not directly model inflammatory disease.

Clearly, our data demonstrate that rat blood has the ability to produce copious amounts of PGE₂ via the actions of COX-1 enzyme constitutively present in platelets. This observation begs a simple question. Is platelet production of PGE₂ of any particular relevance? In comparison with the characterization of their function in blood clotting, the role of platelets as mediator and effector cells in inflammation has only recently been recognized (Klinger, 1997). Platelets participate in inflammatory events by releasing a considerable number of mediators and by interacting with leukocytes and endothelial cells. Furthermore, the contribution of platelets to inflammatory processes may not be solely restricted to interactions within the vasculature. In fact, several studies have shown that platelets can be found together with leukocytes in the exudates of numerous inflammatory diseases (Bazzoni et al., 1991). Indeed, of particular relevance to this discussion is the observation that platelets accumulate in the exudate provoked by the subdermal implantation of sponges in rats (Smith et al., 1976). Moreover, platelets also appear to accumulate in the carrageenan-induced paw edema in rats (Vincent et al., 1978). In these circumstances, one might expect, consistent with the results presented here, that platelet activation in vivo could result in a significant production of PGE₂ from COX-1. Indeed, the PGE₂ produced by platelets may be of extreme importance in the early stages of the inflammatory process.

Although our experiments suggest that COX-1 is a major source of "inflammatory" PGE₂ in the rat, they also indicate that this is not so in humans (Warner et al., 1999). Notably, as shown above, human blood can only produce a small fraction (1/20) of the levels of COX-1-derived PGE₂ produced by rat blood. Indeed, we and others have taken the production of PGE₂ in LPS-treated human as a signal of COX-2 activity and used it to characterize the activities of a range of NSAIDs (Warner et al., 1999; FitzGerald and Patrono, 2001), as has been done here in rat whole blood. At the same time, it is interesting to note that rats are often used in the characterization of NSAIDs and novel COX-2 inhibitors in inflammatory models (Chan et al., 1995; Riendeau et al., 1997). The data gathered in the present study support the idea that in these inflammatory models the contribution of COX-1-derived PGE₂, both as a confounding factor and as a mediator, may have been overlooked. This is in agreement with the experiments of Wallace et al. (1998, 1999) who reported that COX-2 inhibitors, such as SC-58125, DuP-697, or NS-398, need to be administered at nonselective COX-1-inhibiting doses to obtain a significant anti-inflammatory effect in the carrageenan air pouch or in the carrageenan paw edema in rats. From this, Wallace and his colleagues proposed that COX-1 may be a major contributor to the inflammatory process in the animal models used. Importantly, the results shown here, although looking at systemic levels of prostanoids and not local production at an inflammatory site, are consistent with this hypothesis in that the production of COX-1-derived PGE₂ by platelets may well underlie the involvement of COX-1 in inflammation. Therefore, it may not be surprising that in rats COX-2-selective inhibitors fail to produce the full anti-inflammatory effects observed with standard NSAIDs (Wallace et al., 1998, 1999). Most importantly, however, the differences found between human (this study and Warner et al., 1999) and rat blood suggest that extrapolations to humans of results from NSAID testing in the rat should only be made very cautiously. Indeed, the role of COX-1 in the production of "inflammatory" PGE₂ in the rat was not reflected in human samples, supporting the use of COX-2-selective inhibitors as human anti-inflammatories.